digital marketing research articles

Digital transformation and marketing: a systematic and thematic literature review

• Review Article
• Open access
• Published: 15 March 2023
• Volume 2023 , pages 207–288, ( 2023 )

Cite this article

You have full access to this open access article

• Marco Cioppi 1 ,
• Ilaria Curina   ORCID: orcid.org/0000-0001-7702-7664 1 ,
• Barbara Francioni 1 &
• Elisabetta Savelli 2  

10k Accesses

4 Citations

1 Altmetric

Explore all metrics

This article provides a systematic review of the extensive and fragmented literature focused on Digital Transformation (DT) and marketing by identifying the main themes and perspectives (i.e., employees, customers, and business processes) studied by previous research. By mapping the DT literature in the area of marketing, 117 articles, published between 2014 and 2020, have been identified. Through the adoption of a content analysis process, a multi-dimensional framework synthesizing the DT and marketing binomial has been provided. Results identify two thematical patterns: the macro-themes, related to the main digital technologies adopted within the marketing function, and the micro-themes, related to the effect/impact of these technologies on marketing processes and activities. Concerning the micro-themes, findings show how they have mainly studied from the customer and business processes’ perspectives, thus identifying an interesting research gap related to the analysis of the DT-marketing phenomenon from the employees’ standpoint. Based on these results, the paper derives a research agenda by also providing theoretical and managerial implications. Theoretically, it is the first systematic and thematic review focused on DT and marketing. In particular, it analyses this binomial from a broad and comprehensive perspective, thus offering a synergistic framework of the existing literature, which allows an inclusive vision and understanding about the phenomenon. At the managerial level, the paper could help organizations to enhance their awareness about marketing areas and processes that could better benefit from digitalization, thus driving the overall transition of firms towards DT.

Similar content being viewed by others

How digital technologies reshape marketing: evidence from a qualitative investigation

The Digital Marketing Toolkit: A Literature Review for the Identification of Digital Marketing Channels and Platforms

How do entrepreneurs perform digital marketing across the customer journey? A review and discussion of the main uses

Explore related subjects.

• Artificial Intelligence

Avoid common mistakes on your manuscript.

1 Introduction and background

Over the last decades, digital transformation (DT) has received growing attention in the business literature since it represents a prominent feature for organizations to be leaders of change and competitive in their domain (Kraus et al., 2022 ). At once, in light of the COVID-19 pandemic, the DT phenomenon has experienced an abrupt acceleration (Priyono et al., 2020 ), as firms and organizations are forced to redesign their strategies and operating models through a massive adoption of technologies in order to respond to the crisis-caused changes (Hai et al., 2021 ; Hanelt et al., 2021 ). Therefore, the necessity of analysing the DT topic has become ever more crucial in the last few years.

Conceptually, DT refers to all changes that digital technologies can bring in a firm’s business model, concerning products, processes, and organizational structures (Hess et al., 2016 ). Starting from this definition, it appears clear the pervasiveness of this phenomenon, which represents a real transition toward a new reality made of risks and challenges (Horvat and Szabo, 2019 ; Kraus et al., 2022 ; Vial, 2019 ). DT, indeed, can change every aspect of business, especially the marketing one (Caliskan et al., 2020 ).

Notably, the connection between DT and marketing has become ever more decisive in the last two years. The critical changes related to the COVID-19 crisis have particularly altered the firm and consumer relations, forcing companies to modify their marketing strategies through the massive exploitation of the digital technologies. In particular, marketing currently represents one of the main functions requiring to be adapted to the DT in order to protect firms’ competitiveness (Caliskan et al., 2020 ). By following this research stream, some authors have tried to synthetize the main impacts of DT on marketing practices (Shkurupskaya and Litovchenko, 2016 ; Sunday and Vera, 2018 ), including (i) The increasing spread of information and communication technology (ICT) in the marketing communication channels; (ii) The opportunity to adopt real-time communication with customers; (iii) The development of new relationships between producers and consumers; (iv) The increasing effectiveness of the marketing activities through the monitoring of real-time data. Meanwhile, other authors have specifically focused their attention on the main digital technologies able to offer significant benefits to the marketing function (Ardito et al., 2019 ; Cluley et al., 2019 ; Giannakis et al., 2019 ; Ungerman et al., 2018 ) by also categorizing them on the basis of the marketing mix (Caliskan et al., 2020 ).

Despite the DT-marketing topic has received growing attention, to date, no systematic review exists concerning the analysis of the DT phenomenon with specific application to the marketing processes and activities. Notably, several studies have tried to review the DT literature from very restricted research areas (Hanelt et al., 2021 ) different with respect to the broader one of marketing, such as B2B relationships (Hofacker et al., 2020 ), business model innovation (Favoretto et al., 2022 ; Li, 2020 ), accounting (Knudsen, 2020 ), multinational enterprises (George and Schillebeeckx, 2022 ), leadership (Carvalho et al., 2022 ; Henderikx and Stoffers, 2022 ), quality management (Dias et al., 2021 ; Thekkoote, 2022 ), production applications (D’Almeida et al., 2022 ), business management adaptability (Zhang et al., 2021 ), stakeholder management (Prebanić and Vukomanović, 2021 ), and sustainability (Gomez-Trujillo and Gonzalez-Perez, 2021 ). Faced with this context, some authors have tried to analyse and systematize the previous DT literature within broader research areas such as the business and management (Kraus et al., 2022 ) and the organizational change (Hanelt et al., 2021 ). However, despite these contributions, until now, no study has focused on reviewing the literature dedicated to the binomial DT-marketing.

Starting from these assumptions, the present study aims to provide a comprehensive review of the extant literature focused on DT in the marketing area by identifying the main themes and perspectives of analysis. More in detail, the paper addresses the following research questions: (i) What themes have been studied by previous research on DT in the field of marketing? (ii) What are the main perspectives adopted by the research on DT in the field of marketing?

To answer these research questions, the study has been organized in two phases: while in the first one the DT literature has been mapped by focusing on all studies addressing the digital transformation and marketing topics during the period 2014–2020, in the second phase a synergistic framework with the main macro and micro themes characterizing DT in the marketing area (concerning the digital technologies use and effects, respectively), along with the related analysed perspectives, has been provided.

By doing so, this study informs the academicians about the recent evolution of DT literature on marketing-related topics. Additionally, by proposing a synergistic framework of results, the paper provides a solid support for discussing and delineating future research directions. Finally, the main results of this review could help organizations to increase their awareness about marketing areas and processes that could better benefit from digitalization, thus driving the overall transition of firms towards DT.

The remainder of the paper is structured as follows. Section  2 presents the methodology and Sect.  3 outlines the descriptive and thematic results of the study. Section  4 provides theoretical and managerial implications and proposes future research directions based on the main gaps in existing literature. Finally, Sect.  5 concludes the study by also discussing the main limitations.

2 Methodology

This study adopts the systematic review method (Tranfield et al., 2003 ) to detect, classify, and interpret “all the available research relevant to a particular research question, or topic area or phenomenon of interest” (Kitchenham, 2004 ; p. 1). Structurally, the review process has been divided into three phases: (i) Data collection; (ii) Paper selection; (iii) Content analysis.

The identification of specific keywords and terms represents the first systematic review step (Tranfield et al., 2003 ). In our research, the following string has been adopted: [“Digital transformation” AND “marketing”], with the final aim of identifying all the contributions simultaneously focused on these two topics, regardless of the subject area (e.g., business, management, etc.) and research approach (e.g., qualitative vs . quantitative). The Scopus database has been employed as it represents the broader abstract and citation database of peer-review literature, and it also contains most of the publications from other databases (Guerrero et al., 2015 ).

All the proposed document typologies have been included in the analysis (i.e., articles, conference papers, conference reviews, literature reviews) by applying the above string on their title, abstract, and keywords (Table 1 ). As for the time frame, contributions published between 2014 and 2020 have been considered following the study of Vaska and Colleagues ( 2021 ), which reveals a growth in interest toward DT field, particularly from 2014.

A total number of 134 publications have been identified and further selected by considering only those studies effectively focused on the investigated topics. At the end of this process, 117 documents have been retained and subjected to content analysis to identify the main DT themes and perspectives in the marketing field (Fig.  1 ).

Main steps of the literature analysis

Notably, the content analysis allows the “systematic and theory-guided reduction of a large amount of text data from any type of communication down to its essence by classifying the material into unifying categories” (Hanelt et al., 2021 ; p. 1163). It is distinguished from other qualitative procedures, such as the thematic one, since it permits to build category systems in line with the research questions, thus providing both qualitative and quantitative insights (Mikelsone et al., 2019 ).

3 Results and discussion

In the following sub-paragraphs, the descriptive and thematic results of the literature review will be presented.

3.1 Descriptive results

Concerning the yearly research trend (Fig.  2 ), a growing interest in the digital transformation-marketing topic emerged during the time-period under review. Particularly, we went from only one contribution published in 2014 to three in 2017; starting from 2018, the attention increased with 13 published articles, while the most significant peaks have been reached between 2019 and 2020, characterized by the higher production of contributions (45 in 2019 and 50 in 2020).

Year distribution of contributions

Table 2 ranks the sources with the highest number of published contributions focused on the investigated topic. Ninety-three sources have published the 117 reviewed papers with the more relevant contribution from the Advances in Intelligent Systems and Computing (3,4%), followed by Industrial Marketing Management (3,4%), and IOP Conferences series: Materials Science and Engineering (3,4%), Communications in Computer and Information Science (2,6%), and Journal of Physics (2,6%).

Additional sources with only one published contribution are shown in Table 3 . Notably, fifty-seven sources are Journals, eighteen are conference proceedings, and two sources are book series. Concerning the Journals, those from a domain especially related to the business management, society, technology innovation, economics, and engineering have shown interest toward this specific issue. With respect to the conference proceedings, the main fields of study concern the smart trends, technology innovation management, computer science, and information systems. Finally, regarding the book series, they are specifically focused on the information and communication and tourism research streams.

The source’s distribution is informant about the main future publication opportunities in the area of DT and marketing. Equally relevant is the result related to the contributions’ ranking per citation since it allows to figure out the widespread and dissemination of the analysed research stream. Table 4 shows the top-ten contributions in terms of citations. Notably, the more cited contributions are very recent (published between 2019 and 2020) and mainly focused on the following topics: technological innovations as enablers for firms’ digitalization strategies (Ballestar et al., 2019 ; Gil-Gomez et al., 2020 ; Hausberg et al., 2019 ; Peter et al., 2020 ; Sestino et al., 2020 ; Ulas, 2019 ; Yigitcanlar et al., 2020 ) and business sustainability (Sivarajah et al., 2020 ), and the impact of the COVID-19 crisis on consumers’ (Kim, 2020 ) and firms’ digital behaviours (Almeida et al., 2020 ).

Finally, concerning the adopted methodologies, 93 (79,5%) contributions are based on qualitative methods, while the remaining 24 (20,5%) are quantitative in nature.

3.2 Thematic results

By employing the content analysis, it has been possible to extract the main DT themes and perspectives in the marketing fields. As for the DT themes, two main clusters have been identified:

Macro-themes related to the use of digital technologies within the marketing function;

Micro-themes related to the effects emerging from the use of digital technologies on marketing processes and activities.

3.2.1 Macro-themes related to the use of digital technologies

The identification of the most investigated digital technologies analysed in the marketing domain by the reviewed contributions represents the first result deriving from the content analysis. Appendix 1 displays the list of technologies along with their main conceptualizations. As shown in Table 5 , the majority of contributions (67,1%) have focused their attention on the analysis of specific digital tools. In particular, the social media channels (social media marketing) represent the most examined technology (being investigated by 9,4% of the selected studies), followed by Big Data (8,7%), mobile marketing (i.e., mobile technology and smart apps) (8,1%), Internet of Things (6,7%), Artificial Intelligence (6,7%), and Industry 4.0 (6,7%). The remaining technologies (i.e., Machine learning; Online collaborative/support platforms/systems; Virtual/Augmented Reality; Websites/SEO; Cloud infrastructures; Chatbots; Drones/Smart robots; Security Protection systems; 3D print) have experienced a reduced interest by the extant literature (less than 6% of the identified contributions). Finally, a not negligible percentage of studies (32,9%) has analysed the topic of digitalization without investigating specific technologies. Rather, they broadly referred to the “digitalization phenomenon” as an overall macro-theme investing the marketing area.

The sum of the identified macro-themes ( n  = 149) exceeds the number of papers analysed during the review process ( n  = 117) since some papers have simultaneously examined more than one macro-theme.

3.2.2 Micro-themes related to the effects emerging from the use of digital technologies

The second result achieved by the content analysis concerns the main effects (i.e., micro-themes) deriving from the adoption and exploitation of the already identified digital technologies (Par. 3.2.1 ) on the marketing function. The most examined effects fall within the following areas: customer relationship management, customer connectivity, and customer centricity (12,3%), human resources (10,3%), digital metrics (8,8%), customer experience/journey (8,3%), business process efficiency (8,3%), MarTech (7,8%), market knowledge (7,4%), communication policy (5,9%), and customer behaviour (5,4%). The remaining effects (i.e., product policy, sales processes; production; buying/consumption processes; value co-creation; supply chain; branding; customer service; etc.) received less attention, being investigated by less than 5% of the identified contributions (Table 6 ).

The sum of the identified micro-themes ( n  = 204) exceeds the number of papers analysed during the review process ( n  = 117) since some papers have simultaneously examined more than one micro-theme.

The content analysis allowed as to go deep into the study of each micro-theme by revealing both a detailed list of specific sub-themes (Table 7 ) and the main perspectives of analysis adopted in the reviewed manuscripts (Table 8 ).

Specifically, three main perspectives emerged from our study, namely employees, customers, and business. While the employee perspective focuses on the human resources and their coexistence with new technologies, the customer one is mainly related to the digital opportunities offered on the consumer side, especially concerning the overall shopping journey. Finally, the process-focused perspective is primarily concerned with the influence of digital technologies on the different business practices and procedures.

3.2.3 Macro-themes, micro-themes, and analysed perspectives: a combined overview

In this section, the macro-themes, micro-themes, and analysed perspectives will be combined with the final aim of building a comprehensive overview (Table 9 ).

By focusing on the first macro-theme (i.e., social media channels), no studies have specifically examined it from the employee perspective, thus identifying an interesting research gap. Conversely, research widely underlined the key-role of these tools from the business processes and customer perspectives. Concerning the first one, different contributions highlighted how social media support a multitude of business processes (e.g., segmentation, brand positioning, promotion, advertising, buying, after-sales), thus improving firms and marketing performance (Al-Azani and El-Alfy, 2020 ; Kazaishvili and Khmiadashvili, 2020 ; Lestari et al., 2019 ; Melović et al., 2020 ; Rebelli, 2019 ; Safiullin et al., 2020 ; Sivarajah et al., 2020 ; Ulas, 2019 ; Van Osch et al., 2019 ). At once, an equally relevant number of studies has also examined the social media impact from the customers’ viewpoint (Hahn, 2019 ; Kumar-Singh and Thirumoorthi, 2019 ; Rebelli, 2019 ; Yusmarni et al., 2020 ) by identifying the main advantages for them, such as their involvement and engagement in the value creation process and the access to personalized assistance services (Kazaishvili and Khmiadashvili, 2020 ; Sivarajah et al., 2020 ).

Big Data represent the second macro-theme extracted from the thematic literature review. These have been especially analysed from the business processes perspective, recognizing them as one of the most significant challenges and innovations of recent years within the DT framework. Almaslamani et al. ( 2020 ), for instance, explained how the Big Data adoption can lead firms to use intelligent market basket analysis, thus enhancing the relationship with customers. Similarly, the study of Miklosik and Evans ( 2020 ) analysed the impact of Big Data on the digital transformation of the marketing industry by examining the main challenges it faces from a data and information management viewpoint. At once, Sestino et al. ( 2020 ) provided interesting implications for marketers by underlining how the DT, enabled by Big Data, can positively influence many facets of business (e.g., collection of large-scale data allowing to identify emerging trends on consumer behaviour; creation of promotion campaigns with real-time data; creation of stronger bonds with consumers). By specifically focusing on the B2B market, the study of Sivarajah et al. ( 2020 ) demonstrated the Big Data capability to allow B2B firms to become profitable and remain sustainable through strategic operations and marketing-related business activities. Overall, the research offers interesting implications for all the stakeholders interested in understanding and exploiting the use of Big Data with the final aim of achieving business sustainability.

As for mobile marketing (mobile technology and smart apps), research has mainly examined it by focusing on the customer perspective. Indeed, mobile devices have deeply influenced customers’ behaviours and preferences toward online shopping (Sundaram et al., 2020 ) by also transforming them into an integral part of the value creation process. Meanwhile, mobile technology and smart apps have also been studied from the business processes viewpoint since they have become an excellent opportunity to analyse consumers in more meaningful manners, thus supporting the development of appropriate marketing strategies (Sundaram et al., 2020 ). Additionally, mobility, along with other digital technologies, is creating relevant opportunities for firms to transform themselves by impacting on their purchasing processes (Ulas, 2019 ) as well as on their distribution activities, since mobile apps represent omni-channel retail platforms allowing consumers to obtain products from different channels, such as e-commerce, modern markets, and traditional ones. In this way, the shopping experience streamlines and integrates itself across channels (Cahyadi, 2020 ). Conversely, even if the employee perspective has been less investigated, it represents an interesting field of study since the mobile technology is impacting, on a massive scale, the workplace (Attaran and Attaran, 2020 ). More in detail, it can raise employee engagement; increase productivity through the scheduling/automation of daily activities; enable real-time communications through different tools, such as group chats or one-to-one messaging. Moreover, the 5G advent could revolutionize the way employees work “in much the same way the Internet did in the 1980s” (Attaran and Attaran, 2020 ; p. 66). Notably, it can allow employees to (i) Fast download and upload files and documents; (ii) Quicker move data; (iii) Carry the office anywhere; (iv) Exploit resources such as real-time video interaction and smart conference/meetings rooms, thus maximizing the workplace productivity and efficiency, reducing travel time, and saving operational costs for remote employees; (v) Increase office collaboration; (vi) Synchronize and access to large amounts of data storage.

Another macro-theme widely analysed by the literature focused on the DT and marketing is Internet of Things, which represents one of the main megatrends related to the technological revolution (Hamidi et al., 2020 ). Extant research (e.g., Almeida et al., 2020 ; Chehri and Jeon, 2019 ) has particularly examined the main improvements provided by this technology in terms of business processes. Notably, Sestino et al. ( 2020 ) underlined how IoT can contribute to: (i) Design products/services based on consumers’ consumption experiences; (ii) Collect consumption data useful, for marketing managers, to identify new gaps, trends, or variables in understanding consumer behaviour; (iii) Identify consumers’ attitudes and choices on a large scale. At once, different studies (e.g., Almeida et al., 2020 ; Sestino et al., 2020 ) have also investigated the impact of IoT from the customer perspective by focusing on their ability to provide new types of services and high-quality products; as well as to improve the customer journey through more targeted promotions, announcements, and email marketing. Finally, even if the employee perspective represents the least investigated one, some authors (e.g., Almeida et al., 2020 ; Peter et al., 2020 ) identified several IoT advantages from this viewpoint, including the possibility of adopting mobile, flexible, team-oriented, and non-routine working methods, which allow the creation of digital workplaces; activating collaborative practices between all the staff’s levels; and communicating and disseminating corporate strategies, thus creating innovative workplaces.

Concerning the Artificial Intelligence (AI), it has been analysed from all the perspectives, especially the customer and business processes ones. Different studies investigated the advantages of the AI-based digital humans for customers, including the possibility to obtain better knowledge of their preferences and needs (Kumar-Singh and Thirumoorthi, 2019 ), to build an innovative and real-time relationship with the firms (Cherviakova and Cherviakova, 2018 ), to experience a completely new and interactive journey, and to receive personalized offers (Ianenko et al., 2019 ). From the processes perspective, AI significantly influences marketing processes and activities (Almeida et al., 2020 ; Ianenko et al., 2019 ; Sargut, 2019 ) through the analysis of the customers’ behaviours and the realization of more specific targeted profiles (Ianenko et al., 2019 ). AI also influences the distribution activities and, in particular, the automation of the ordering process of products and services (Cherviakova and Cherviakova, 2018 ). Moreover, by considering unexpected events, AI allows to recalculate new routes and to maintain constant contacts with clients and the logistics service providers. Literature (Cherviakova and Cherviakova, 2018 ) underlined the AI role in allowing the automatic placement of advertisements across channels, while Kumar-Singh and Thirumoorthi ( 2019 ) analysed the AI relevance also with respect to the buying/consumption process. Finally, it has been recognized the importance of AI with respect to both sales (Almeida et al., 2020 ) and after-sales processes, as it permits to better examine the customers’ opinions about products/services, and to identify their satisfaction level as well as the possible enhancements that could be applied to the firm’s offering. Concerning the employee perspective, AI–by representing a disruptive technology–has significantly influenced the labour relations model and, in particular, the knowledge sharing among employees (Almeida et al., 2020 ; Subramani, 2019 ; Ulas, 2019 ). Therefore, it becomes fundamental to enhance the employee training toward this digital tool, which is becoming more and more integrated into the workplace (Yigitcanlar et al., 2020 ).

By representing a multifaceted term, the Industry 4.0 has emerged as an additional macro-theme related to the DT-marketing binomial. Notably, research (e.g., Chehri and Jeong, 2019 , Del Giorgio and Mon, 2019 ; Hamidi et al., 2020 ) has mainly investigated this topic from the customer and business processes perspectives, especially by focusing on the main principles behind it, namely 5c (i.e., Cooperation, Conversation, Co-creation, Cognitivity, Connectivity). This technology has created the basis of the digital ecosystem, thus offering the key ability, for firms and customers, to exchange data in real-time (Nosalska and Mazurek, 2019 ). By specifically focusing on the business processes perspective, an interesting point of view has been provided by Naglič et al. ( 2020 ), who analysed the Industry 4.0 macro-theme in combination with the export market orientation/export performance micro-theme. The authors offered a framework on how companies can enhance their export performance through the knowledge related to the Industry 4.0. Overall, their study detected how firms that invest in digital technologies, by effectively embracing DT, are better prepared to compete internationally, thus achieving better export performance.

Also the Machine Learning (ML) macro-theme has been mainly analysed from the business processes perspective. In particular, some studies have tried to identify the main ML implications on DT in marketing (Miklosik and Evans, 2020 ) by investigating the advantages this technology can bring from this perspective (Kazaishvili and Khmiadashvili, 2020 ; Miklosik and Evans, 2020 ; Polyakov and Gordeeva, 2020 ; Sargut, 2019 ). Literature focused its attention on the social media analysis (e.g., sentiment analysis on social media); packaging; product and purchasing decision-making; and advertising (e.g., interactive ad placement and targeting ads). Given that ML is a subset of AI, the literature focused on ML usually underlined, from the employee and customer perspectives, advantages very similar to the AI-related ones. More in detail, from the customers’ perspective, ML can offer personalized shopping experiences thanks to its ability to deeply know their preferences and interests. Conversely, from the employees’ viewpoint, literature mainly highlighted the key impact of ML on knowledge building and sharing (Subramani, 2019 ).

Concerning the online collaborative/support platforms/systems macro-theme, it emerges how it has been equally analysed from the employee and business processes perspectives. From the employee perspective, Azeredo et al. ( 2020 ) provided a proposal for the realization of an online business consulting plan through the adoption of an online collaborative platform called LexDoBusiness. More in detail, the research aimed to analyse the acceptability of this platform, which offers several benefits, especially for what concerns the levels of cohesion and cooperation between the actors involved in the business plan. In their study, Bhatnagar and Grosse ( 2019 ) underlined the relevance of a digitalized agile workplace since it allows to make employees more productive and satisfied. Similarly, Minculete and Minculete ( 2019 ) emphasized the key role of education and training actions aimed at providing staff members with the required skills for the new technologies and systems adoption. By specifically focusing on the business processes perspective, Bruskin et al. ( 2017 ) examined the development of support systems for decision-making in terms of marketing by specifically focusing on the analysis of the business effects from the adoption of similar systems.

As regards the virtual and augmented reality, literature has mainly examined it from the customer and business processes perspectives. For what concerns the first viewpoint, the majority of studies have investigated the consumers’ propensity to interact with this tool (Voronkova, 2018 ). Additional researches have focused their attention on the new opportunities deriving from adopting virtual and augmented reality for personalized online shopping experiences (Kim, 2020 ). From the business processes perspective, the virtual/augmented reality has been particularly examined with respect to the communication and advertising procedures. Notably, extant research underlined how firms can adopt the virtual reality technology to promote products and services in innovative and visual ways (Voronkova, 2018 ).

For what concerns the last identified macro-themes (i.e., websites/SEO; cloud infrastructure; chatbots; drones/smart robots; security protection systems; 3D print), results have already revealed a minor attention dedicated to them by the extant research (Table 5 ). By focusing on the websites/SEO topic, the customer and business processes perspectives represent the most investigated viewpoints. Existing studies have particularly analysed the websites topic with respect to the customer relationship management/customer connectivity/centricity (Ballestar et al., 2019 ) and customer experience/journey (García et al., 2019 ) micro-themes. With regard to the business processes perspective, the reviewed contributions have especially deepened the micro-themes of branding, communication policy, and business process efficiency. Specifically, Natorina ( 2020 ) underlined the need to implement effective marketing strategies within the DT scenario by specifically focusing on the search engine optimization (SEO). Overall, the author highlighted how the SEO represents an integral component of a successful marketing strategy since it increases the organic traffic and conversion by also enhancing the firms’ attractiveness in the sight of the Internet users.

Concerning the cloud infrastructure, it has been especially analysed from the customer perspective (Ulas, 2019 ) by investigating its impact on consumers’ preferences and behaviours. At the same time, the cloud infrastructure has also increased the human resources capabilities (Ulas, 2019 ) and improved the business processes. Notably, Kumar-Singh and Thirumoorthi ( 2019 ) shown that cloud-based digital infrastructures allow firms to increase agility, maximize resources, and improve services by also reducing operational costs. The authors also underlined the importance to analyse the impact of this technology from the demand side in order to examine how it can impact on customer preferences and behaviours.

As for the chatbots, these have been analysed from the business processes perspective and, to a lesser extent, from the employee one. Hence, an interesting research gap emerges with respect to the customer viewpoint. In particular, concerning the business processes perspective, Damnjanovic ( 2019 ) proposed a case study analysing the international positioning and go-to-market strategy of a chatbot solution, namely Weaver, which can be defined as an AI-based firm platform allowing to facilitate and simplify the sales processes. In the same year, the study of Sargut ( 2019 ) offered an insight related to the SMEs awareness, readiness, and capability in facing the DT challenge. Almost all the interviewed SMEs have confirmed to be interested in the DT subject and ready to implement chatbots and/or voice-operated machines in their business activities and processes.

Even if results underlined scarce attention of the recent literature on the robotics macro-theme (with the few identified contributions focused on the employee and business processes perspective), with the advent of the COVID-19 and the consequent reduction of human contacts, this topic will probably obtain, in the future, greater emphasis. Notably, robots will be increasingly adopted not only in order to substitute human resources but also to interact with customers. Indeed, robots “are expected to be progressively more autonomous, flexible, and cooperative” (Almeida et al., 2020 , p. 102).

As for the last identified macro-themes (i.e., security protection systems and 3D print), while Li et al. ( 2020 ) emphasized the need to establish a new generation of security protection systems to increase the business processes efficiency, Ulas ( 2019 ) especially highlighted the key relevance of 3D printers in the process of new products development and design.

By considering the residual (but not irrelevant number of) contributions referring to the digitalization phenomenon as a broader macro-theme of analysis (i.e., digitalization phenomenon), it emerged an overall preference towards the adoption of a business processes and customer perspective. With regard to the former, two of the most investigated effects are the so-called “digital metrics” and “business process efficacy”. Indeed, the digitalization phenomenon has profoundly affected the analysis of the firms’ performance. Hence, the adoption of digital tools allows firms to precisely monitor and measure their social ROI (Return on Investment) in a totally new and disruptive way compared to the past. In particular, by measuring online reactions (e.g., customers’ views, likes, comments, shares), the digital metrics can contribute significantly to evaluating an ad campaign in real-time, thus permitting to modify it accordingly (e.g., Bughin et al., 2019 ). Moreover, a number of contributions focused on the business processes perspective has specifically analysed the role played by the digital tools in increasing the quality of the firms’ processes, thus elevating their levels of operational and organizational excellence (e.g., Kuimov et al., 2019 ). On the other hand, from the customer perspective, literature has mainly investigated the impact of the digitalization phenomenon on the customer journey (e.g., Taylor et al., 2020 ) and on the relationship management between firms and customers (e.g., Barann, 2018 ).

After the content analysis process has been concluded, Appendix 2 has been created, displaying the classification of the articles based on the following categorizations: (i) Author/s; (ii) Title; (iii) Source; (iv) Year of publication; (v) Analysed macro-theme; (vi) Analysed micro-theme with (vii) The respective analysis perspective (i.e., EP, CP, BPP).

4 Implications and future research agenda

4.1 general discussion.

Both the descriptive and thematic results of this study provide interesting insights into the analysis of the DT-marketing topic, while crafting new propositions for future research agenda.

Descriptive data highlight the growing focus of the literature on the digital transformation-marketing topic over the last few years, with the majority of contributions published between 2019 and 2020. Notably, only nine publications have been found in the four-year period 2014–2017, while thirteen publications were reviewed in 2018, forty-five in 2019, and fifty in 2020. The publication sources are highly fragmented, given that ninety-three sources have published the 117 reviewed papers. The more cited contributions—besides being published between 2019 and 2020—have especially focused on the impact of the digitalization phenomenon on (i) Customer relationship management (Ballestar et al., 2019 ; Gil-Gomez et al., 2020 ; Hausberg et al., 2019 ; Peter et al., 2020 ; Sivarajah et al., 2020 ), (ii) Its coexistence with the human resources (Almeida et al., 2020 ; Gil-Gomez et al., 2020 ; Ulas, 2019 ; Yigitcanlar et al., 2020 ), and (iii) The improvement of the business processes’ performance (Sestino et al., 2020 ) by specifically focusing on market knowledge (Hausberg et al., 2019 ), communication (Ballestar et al., 2019 ), product development (Ulas, 2019 ), and sales activities (Almeida et al., 2020 ). Moreover, the majority of contributions here analysed has employed qualitative methods. Overall, these data, while suggesting an increasing interest by the scientific community towards the DT-marketing phenomenon, depict the absence of sources systematically and continuously dealing with this field of study, a dominant focus on certain issues, and the need to improve the adoption of quantitative methods in future research, both to validate previous research findings and to make them more generalizable.

Concerning the research questions guiding this study and, in particular the analysed themes (RQ1), these can be grouped on a twofold level concerning (i) The study of digital technologies employed in the field of marketing ( macro-themes) , and (ii) The impact of such technologies on specific marketing activities ( micro-themes ). Overall, the literature analysis suggests an increasing pervasiveness of digital technologies in the marketing field. The use of such technologies, in fact, affects the consumer behaviour, as well as the way marketers work and marketing activities are managed and organized. In particular, it is worthy to note that DT involves the most operational marketing activities (e.g., Caliskan et al., 2020 ), such as sales (e.g., Almeida et al., 2020 ) and communication policies (e.g., Alassani and Göretz, 2019 ; Dasser, 2019 ), allowing a general increase in these processes’ quality. Meanwhile, DT also affects the analytic and strategic areas of marketing, improving the opportunities to reach new groups of consumers through the systematic use of digital technologies (such as Big Data) that allow a deeper segmentation of the market (e.g., Almaslamani et al., 2020 ). It supports the development of new branding strategies and the increasing visibility of brands, thanks to the use of online and social channels (e.g., Kazaishvili and Khmiadashvili, 2020 ; Melović et al., 2020 ). Moreover, DT impacts on companies’ innovativeness, helping the implementation of more effective and efficient innovative processes (Calle et al., 2020 ), and changes the overall relationships between firms and consumers by encouraging a customer-centric organizational culture (Cherviakova and Cherviakova, 2018 , Graf et al., 2019 ) and the customer participation in the value creation process (Hughes and Vafeas, 2019 ). According to Dasser ( 2019 ), DT also implies a deeper change of marketing by elevating its strategic role as a catalytic accelerator in the digital business transformation journey.

These studies are driven by different perspectives of analysis (RQ2). The majority of research considered in this review employed a business process perspective by examining how digital technologies impact on specific marketing processes, such as sales and communication management. Nevertheless, by focusing on the main investigated topics, findings reveal that the existing research has been principally guided by a customer perspective, i.e. the way in which digital technologies are transforming customers’ behaviour, experience, and relationship with companies, followed by the business processes perspective concerning the investigation of potential improvements occurring in the area of marketing analysis and control. The employees’ perspective emerges as the less relevant among the others, despite it includes a critical part of the literature focused on the relationship between DT and human resources management. More in detail, as it emerged from our dataset, the employees’ perspective mainly characterized the first publications, investigating how digital technologies are enhancing (and requiring) the development of new marketing and business skills dealing with DT (Kwon and Park, 2017 ; Van Belleghem, 2015 ). Over the time, the scientific attention has been moved increasingly towards the customer and business processes’ perspectives. Most of the contributions published in 2020, indeed, dealt with the analysis of the DT phenomenon from the consumer viewpoint, specifically investigating the management of the customer-firm relationship (e.g., Gil-Gomez et al., 2020 ; Sivarajah et al., 2020 ), and from the business processes’ viewpoint, especially analysing the key relevance of the digital tools in measuring the firms’ performance in the social sphere (e.g., Al-Azani and El-Alfy, 2020 ; Lin et al., 2020 ). Probably, this growing interest of the research derives from the advent and unleashing, during 2020, of the COVID-19 health crisis that has led companies to almost completely digitize the relationship with customers due to the limitations imposed by the anti-COVID-19 decrees.

All these findings provide several contributions both theoretically and practically.

4.2 Theoretical implications and research gaps

From a theoretical standpoint, this is the first study that offers a systematic and thematic review of the existing literature on DT and Marketing, while previous reviews, in the marketing field, have been very narrow in perspective. Hofacker et al. ( 2020 ), for example, examined the relevant literature on digital marketing and B2B relationships, while Miklosik and Evans ( 2020 ) focused on the impact of big data and machine learning on marketing activities. Our review, instead, addresses the DT-Marketing binomial from a wider and more comprehensive perspective, including all prior research dealing with DT in the marketing area. By doing so, this study outruns the scope of prior reviews that have been often limited to certain domains, and provides a comprehensive framework that offers a synergistic view of the existing literature, which allows a more inclusive vision and understanding about the phenomenon.

By doing so, this review also permits to highlight some relevant research gaps on which future studies might focus on.

From the combined overview between macro- and micro-themes, the main research gaps relate to the necessity of deepening the analysis of the impact of specific macro-themes from the employee (i.e., social media channels, big data, mobile marketing, Artificial Intelligence, Industry 4.0, Cloud infrastructure, Virtual/augmented reality, and websites), customer (i.e., Social media channels, Big Data, Industry 4.0; Internet of Things; Machine Learning; Websites; Chatbots), and business processes perspective (i.e., Mobile technology; Artificial Intelligence; Virtual/Augmented reality; Cloud infrastructure; Drones/Smart robots).

Besides that, the variety of analysed studies, while manifesting the pervasive use of digital technologies in the marketing field, reveals that the extant literature is quite fragmented and even sparse with regard to specific micro-themes. Some topics, like customer service, smart factories, consumer behaviour, have been investigated by few contributions, thus highlighting potential opportunities for further studies. In this respect, our review can be viewed as a solid basis for additional discussion and research within each perspective emerged from the analysis (see Fig.  3 ).

Areas of future research on DT and Marketing

More in detail, the findings reveal that the employees’ perspective is worthy of further attention, as it is the less investigated one. Although several contributions (n. 21) focused on DT and human resources by highlighting the need for enhanced skills in using technology (e.g., Dethine et al., 2020 ; Ulas, 2019 ), the development of new prominent job positions for the future (e.g. digital marketing manager; social media manager; big data/data analyst) (e.g., Di Gregorio et al., 2019 , Hafezieh and Pollock, 2018), and the critical role of training and educational actions enhancing the appropriate use of digital technologies in the marketing context (Yigitcanlar et al., 2020 ), other themes have been under-investigated. In particular, only two papers dealt with the subject of smart technologies by investigating how they can help cities to face the increasing urbanization (Visan and Ciurea, 2020 ), and their importance for establishing a predictive maintenance of production systems, which can increase the process quality (Chehri and Jeon, 2019 ). The application of smart technologies can also redefine the way people conduct business, bringing benefits in terms of productivity and employee well-being (Papagiannidis and Marikyan, 2020 ). Thus, there is scope for considering, in future research, how smart technologies are used to conduct marketing activities and how they are changing the way marketers work and organize their processes.

Under the customer perspective, several topics might deserve attention in future research. Most of the analysed contributions addressed the impact of DT on firms/customers relationships, highlighting the need for new forms of interaction and collaborations with customers due to changes in behaviour. Several scholars recognized the advantage of DT as it allows to establish innovative and real-time relationships with the market (e.g. Almaslamani et al., 2020 ), to engage customers in the value creation process (e.g. Saravanabhavan et al., 2020 ; Taylor et al., 2020 ), and to provide customers with more interactive and personalized experiences (e.g. Taylor et al., 2020 ; Venermo et al., 2020 ). However, our findings suggest that other topics, although relevant, are still at the begin of their investigation. Only three contributions focused on customer service (Lieberman, 2019 ; Lin et al., 2020 ; Safiullin et al., 2020 ), especially revealing the role of digital tools in the online customer service and the importance of electronic services for improving customer satisfaction (Lin et al., 2020 ). A recent study (Galvani and Bocconcelli, 2021 ) revealed that a new business model is emerging in the BtoB context characterized by an overall revolution towards the digital servitization strategy, which replaces the traditional product-centric paradigm. Hence, future research could investigate whether and how the digital servitization strategy is currently implemented in the BtoC context, which opportunities and benefits can offer—especially concerning the firm-customers’ relationship, and how marketing managers can act to face the imperative complexity linked to its adoption. Another theme receiving increasing—but still few—attention concerns the buying/consumption processes. Few scholars analysed the impact of digital tools on customers buying processes (Kim, 2020 ), the increasing use of e-commerce (Cahyadi, 2020 ), and structural changes occurring in consumption during COVID-19 pandemic (Kim, 2020 ). However, the identification of consumption patterns and trends has been always a central topic in the marketing literature, as proved by the wide number of literature reviews, even focused on specific areas such as electronic word of mouth (Huete-Alcocer, 2017 ), online consumption (Hwang and Jeong, 2016 ), or COVID-19 crisis (Cruz-Cárdenas et al., 2021 ). Therefore, continuing the research on DT and consumption/buying behaviour is desirable to properly adapt the marketing management with the aim of satisfying specific market needs and expectations, as well as realizing a stronger engagement of customers in the value creation process, which is getting more and more attention within the recent marketing and management literature (Fan and Luo, 2020 ). Besides, future studies on DT and consumption/buying behaviour might also employ modern research methods, such as neuromarketing. We found only one contribution based on the analysis of the use of advanced methods in the field of artificial neural networks (Polyakov and Gordeeva, 2020 ). However, neuromarketing could contribute to overcome several limitations associated with traditional data collection method (i.e. self-report data), while allowing to capture unconscious brain processes that relate to consumer decision-making (Sung et al., 2021 ).

Finally, an additional space for future research emerged from our review of publications is related to the business processes perspective. This area shows the greatest potential for exploration, given the richness of themes it includes. In this perspective, in fact, except for some activities related to marketing analysis and control, and operational policies—especially product and communication ones—the rest of the literature appears very fragmented and scarce. Notably, specific attention might be devoted to DT and export process management, as Naglič et al. ( 2020 ) found that firms which invest in DT are better prepared to compete internationally and achieve better export performance; branding strategies, as they have been recognized as critical for marketing competitiveness (Kazaishvili and Khmiadashvili, 2020 ), drivers/barriers and risks associated to DT implementation in the marketing areas; and sustainable/social opportunities and treats that digital technologies can bring with them, as they can differently affect the success of human-centric marketing programs in the digital environment (Agafonova et al., 2020 ). All these topics have been very little investigated by previous research, while deserving increasing attention given their relation with companies’ success and long-term competitiveness.

4.3 Practical implications

Regarding the practical contributions, our review offers a number of suggestions to marketing managers as it analyses the DT-Marketing binomial both internally (i.e. on the firm level) and externally (i.e. on the inter-firm level). This approach results from the recognition of different perspectives of analysis adopted by prior research, which combines contributions focused on the management of internal processes and marketing activities with studies investigating the DT phenomenon from a customer-based viewpoint. Consistent with our twofold approach of analysis, the practical implications deserving particular attention can be summarized into two main groups concerning (i) The changing role of marketing in the company resulting from the increasing use of digital technologies, and (ii) The changing relationships between firms (and marketing) and external stakeholders (especially consumers).

Literature suggests that DT could improve the strategic role of marketing within the firm, as it enhances the marketing capability to analyse the market scenario and to develop a more comprehensive understanding of the demand (Papagiannopoulos and Lopez, 2018 ), which, in turn, can support new products development that are better aligned with customers’ expectations (Kuimov et al., 2019 ). Overall, digital technologies can help companies to become data-driven subjects, where marketing covers a central position given its informative and intra-firm coordinating role. However, the full exploitation of such opportunities means change, at both cultural and structural levels. Our review, in particular, reveals that DT requires a cultural upgrading, to cope with DT and its effects on the business (e.g., Álvarez-Flores et al., 2018 ; Dethine et al., 2020 ), the enhancement of internal competences in the field of technology (Ulas, 2019 ), the development of new job positions (Di Gregorio et al., 2019 ), and the gradual adoption of new working habits and patterns (Minculete and Minculete, 2019 ). Of course, educational and training activities become prominent to support such changes, passing through the acquisition of new skills from the market labour, as well as through the enhancement and conversion of internal resources. Besides training programs, organized both internally and externally in collaboration with private and public institutions such as high schools and universities, companies could also provide ad hoc rewards to encourage the commitment and interest of marketing employees in digital innovation.

The second group of advices concerns the changing relationships between firms (and marketing) and external stakeholders (especially consumers). DT affects the customer behaviour and changes his ability to communicate with the company (e.g., Caliskan et al., 2020 ), to be engaged in the value creation process (e.g., Taylor et al., 2020 ), and to live personalized consumption experiences (e.g., Fokina and Barinov, 2019 ). All this implies a general re-thinking about the firm-customer relationship management. Consumers are becoming empowered subjects that no longer accept the role of passive receivers of marketing initiatives (Acar and Puntoni, 2016 ) and companies need to open to their customers, accepting their participation in the marketing decision-processes. Undoubtedly, the use of social-media platforms can be decisive to create engaging content and connect with customers, improving the interaction and the dialog with them, for example by responding to a specific comment or complaint (Acar and Puntoni, 2016 ). However, digital technologies can be also used to create more advanced tools that are able to strengthen the connection between brands and customers, such as crowdsourcing, co-creation, and/or brand communities. These platforms can be used successfully by firms to improve the dialog with customers and their involvement in several marketing processes, such as the selection of an advertising campaign and/or the creation of new product ideas.

Acknowledgements

This publication includes, among the authors, a researcher awarded with a fixed-term type A research contract on innovation topics as per art. 24, para. 3, of Italian Law no. 240 of 30 December 2010, co-financed by the European Union—NOP Research and Innovation 2014-2020 resources as per Italian MD no. 1062 of 10 August 2021.

Open access funding provided by Università degli Studi di Urbino Carlo Bo within the CRUI-CARE Agreement.

Author information

Authors and affiliations.

Department of Communication Sciences, Humanities and International Studies, University of Urbino Carlo Bo, Via Saffi 15, 61029, Urbino, Italy

Marco Cioppi, Ilaria Curina & Barbara Francioni

Department of Economics, Society, Politics, University of Urbino Carlo Bo, Via Saffi, 42, 61029, Urbino, Italy

Elisabetta Savelli

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Ilaria Curina .

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cioppi, M., Curina, I., Francioni, B. et al. Digital transformation and marketing: a systematic and thematic literature review. Ital. J. Mark. 2023 , 207–288 (2023). https://doi.org/10.1007/s43039-023-00067-2

Download citation

Received : 04 May 2022

Accepted : 19 February 2023

Published : 15 March 2023

Issue Date : June 2023

DOI : https://doi.org/10.1007/s43039-023-00067-2

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Digital transformation
• Systematic literature review
• Thematic literature review
• Synergistic framework
• Find a journal
• Publish with us
• Track your research

Smart. Open. Grounded. Inventive. Read our Ideas Made to Matter.

Which program is right for you?

Through intellectual rigor and experiential learning, this full-time, two-year MBA program develops leaders who make a difference in the world.

Earn your MBA and SM in engineering with this transformative two-year program.

A rigorous, hands-on program that prepares adaptive problem solvers for premier finance careers.

A 12-month program focused on applying the tools of modern data science, optimization and machine learning to solve real-world business problems.

Combine an international MBA with a deep dive into management science. A special opportunity for partner and affiliate schools only.

A doctoral program that produces outstanding scholars who are leading in their fields of research.

Bring a business perspective to your technical and quantitative expertise with a bachelor’s degree in management, business analytics, or finance.

Apply now and work for two to five years. We'll save you a seat in our MBA class when you're ready to come back to campus for your degree.

Executive Programs

The 20-month program teaches the science of management to mid-career leaders who want to move from success to significance.

A full-time MBA program for mid-career leaders eager to dedicate one year of discovery for a lifetime of impact.

A joint program for mid-career professionals that integrates engineering and systems thinking. Earn your master’s degree in engineering and management.

Non-degree programs for senior executives and high-potential managers.

A non-degree, customizable program for mid-career professionals.

Generative AI enables companies to execute with speed

4 ways for data officers to take the helm on AI initiatives

In election cycles, voters tend to believe news that confirms their biases

Credit: Anton Vierietin / Shutterstock

Ideas Made to Matter

Digital marketing trends for 2022

Beth Stackpole

Jul 13, 2022

The advent of digital tools has upended age-old processes in marketing and advertising. Digital marketing technology is now a requirement for identifying, attracting, and retaining customers in an omnichannel world.

A new e-book from the MIT Initiative on the Digital Economy highlights learnings from the 2022 MIT Chief Marketing Officer Summit held this spring. The topline message to marketing executives: Add data, analytics, and algorithms to better reach socially-linked modern consumers.

Here are MIT Sloan researchers’ top digital marketing trends for 2022:

Social consumers in broad digital and social media networks

Today’s consumers make brand decisions based on a very broad set of digitally connected networks, from Facebook to WhatsApp, and the mix is constantly in flux.

Since social consumers are influenced by what social network peers think about different products and services (a trend called “ social proof ”), marketers must employ granular analysis to really understand the role of social media in marketing, according to IDE director Sinan Aral.

Aral examined 71 different products in 25 categories purchased by 30 million people on WeChat and found significantly positive effects from inserting social proof into an ad, although the effectiveness varied. For example, Heineken had a 271% increase in the click-through rate, while Disney’s interactions rose by 21%. There were no brands for which social proof reduced the effectiveness of the ads, Aral said.

Video analytics on TikTok, YouTube, and other social media

TikTok influencers loom large, especially with Gen Z. The problem is whether or not those viral influencer videos actually translate beyond attention into sales.

Research shows that engagement and product appearance isn’t the crucial factor — it’s more about whether the product is complementary or well-synched to the video ad. And the effect is more pronounced for “product purchases that tend to be more impulsive, hedonic, and lower-priced,” according to research conducted by Harvard Business School assistant professor Jeremy Yang while he was a PhD student at MIT.

Measuring consumer engagement with machine learning

 Call it the “chip and dip” challenge: Marketers have long grappled with how to bundle goods, finding the right consumer products to combine for co-purchase from a huge assortment. With billions of options, this research is exacting and massive in scale, and data analysis can be daunting.

Researcher Madhav Kumar, a PhD candidate at MIT Sloan, developed a machine learning-based framework that churns through thousands of field scenarios to identify successful and less successful product pairs.

“The optimized bundling policy is expected to increase revenue by 35%,” he said.

Using machine learning to forecast outcomes

Most marketers are concerned about retention and revenue, but without good forecasts, decisions about effective marketing interventions can be arbitrary, said Dean Eckles,  social and digital experimentation research group lead at IDE. Instead, update customer targeting through use of AI and machine learning to forecast outcomes more quickly and accurately.

In collaboration with the Boston Globe, IDE researchers took a statistical machine learning approach to analyze the results of a discount offer on customer behavior after the first 90 days. The short-term surrogate prediction was just as accurate as a prediction made after 18 months.

“There’s a lot of value to applying statistical machine learning to predict long-term and hard-to-measure outcomes,” Eckles said.

Adding “good friction” to reduce AI bias

Digital marketers talk frequently about reducing customer “friction” points by using AI and automation to ease the customer experience. But many marketers don’t understand bias is a very real factor with AI, said  Renée Richardson Gosline,  lead for the Human/AI Interface Research Group at IDE. Instead of getting swept up in “frictionless fever,” marketers must think about when and where friction can actually play a positive role.

“Use friction to interrupt the automatic and potentially uncritical use of algorithms,” Gosline said. “Using AI in a way that’s human-centered as opposed to exploitative will be a true strategic advantage” for marketing.

Read the 2022 MIT CMO Summit Report

Related Articles

Closing the Gap Between Digital Marketing Spending and Performance

by Christine Moorman , Jana Soli and Dennis Cardoso

Summary .   

Marketers have used digital marketing to navigate through incredibly difficult business conditions, connecting with customers stuck at home during the pandemic, digitizing products and services, and driving revenues. Now, it’s time to build on those gains by redoubling their commitment to deepening data and digital mastery, building a culture of continuous learning and experimentation, and using insights to deliver personalized services to customers for higher ROI. Those willing to do so will outpace competitors, notching greater revenues and working more closely with the C-suite to drive business expansion.

Marketers know that digital marketing represents the future of their business. That’s why, according to the February 2022 edition of The CMO Survey , they’re happy to allocate 57% of their budgets to digital marketing activities and are planning to increase spending by another 16% in 2023.

Partner Center

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 28 September 2024

Low-carbon information quality dimensions and random forest algorithm evaluation model in digital marketing

• Weiji Gao 1 , 2 ,
• Zhihua Ding 1 ,
• Junyu Lu 3 &
• Yulong Wan 2  

Scientific Reports volume  14 , Article number:  22416 ( 2024 ) Cite this article

Metrics details

• Environmental economics
• Environmental impact
• Environmental social sciences
• Psychology and behaviour
• Socioeconomic scenarios
• Sustainability

The growing urgency for low-carbon lifestyles necessitates developing effective strategies to promote sustainable consumer choices. This study investigates key dimensions of information quality that shape consumer behavior within digital marketing to achieve this goal. Employing a mixed-methods approach that integrates grounded theory and machine learning, this study identifies three core dimensions of low-carbon information quality: matching quality, presentation quality, and interpretability quality. These dimensions underscore the importance of aligning information with consumer needs, ensuring clear and accurate presentation, and fostering transparency for trustworthiness. A Random Forest algorithm-based evaluation model is constructed to assess low-carbon information quality, demonstrating its effectiveness in identifying high-quality, sustainable content. This research provides a practical tool for digital marketers to enhance their strategies, raise consumer awareness of sustainable options, and ultimately contribute to the growth of the low-carbon consumption market.

Similar content being viewed by others

Examining consumers’ willingness to pay premium price for organic food

Interpreting random forest analysis of ecological models to move from prediction to explanation

Comparing content marketing strategies of digital brands using machine learning

Introduction.

The escalating challenges of global climate change have led to a growing demand for low-carbon products and services 1 . A 2023 World Economic Forum study revealed that over 65% of consumers actively seek low-carbon offerings 2 . This trend is further corroborated by the significant sales growth observed in the low-carbon product lines of companies like Unilever and Nike 3 . However, converting this interest into actual low-carbon consumption is challenging. Research indicates that consumer perceptions of information quality regarding low-carbon products, particularly in terms of accuracy, credibility, and alignment with their needs, significantly influence their purchasing decisions 4 . Therefore, to effectively guide consumer behavior toward low-carbon choices, businesses must prioritize the quality of low-carbon information disseminated through their digital marketing strategies 5 .

Digital marketing, a crucial promotional strategy for businesses, plays a vital role in promoting low-carbon consumption 6 . Companies are increasingly integrating low-carbon elements into these strategies 7 . For example, Patagonia successfully leverages its digital platform to promote environmental awareness and low-carbon products, achieving positive market responses 8 . However, despite growing attention to low-carbon initiatives, the adoption of low-carbon consumption practices has not increased proportionally 9 . This disparity highlights the need for businesses to effectively communicate low-carbon information through their digital marketing strategies 10 , optimizing information quality to maximize consumer engagement with low-carbon consumption 11 , 12 .

Previous research on low-carbon information has primarily focused on dissemination methods and consumer attitudes 4 . However, limited attention has been given to specific dimensions of low-carbon information quality and effective evaluation methods, particularly from a consumer-centric perspective 13 . This study addresses this gap by exploring: (1) key dimensions of low-carbon information quality influencing consumer behavior in digital marketing recommendations; and (2) whether a Random Forest algorithm-based model can effectively evaluate the quality of low-carbon information in these recommendations.

Drawing upon Information Quality Theory (IQT) 14 , this study employs a mixed-methods approach, integrating grounded theory and machine learning. While IQT provides a valuable framework for understanding information quality, it requires further contextualization to address the nuances of low-carbon information within digital marketing 15 . This research provides a nuanced understanding of low-carbon information quality, recognizing its potential to motivate low-carbon choices and evoke positive emotions associated with low-carbon lifestyles 16 . Through grounded theory analysis, a low-carbon information quality evaluation model incorporating multiple features is constructed, and a Random Forest algorithm is employed to train and validate this model. This research aims to provide digital marketing professionals with a practical tool for identifying and promoting high-quality low-carbon information, ultimately enhancing campaign effectiveness and maximizing consumer engagement with low-carbon products and services.

This research contributes to the existing literature by: (1) developing a novel, consumer-centric, low-carbon information quality evaluation model within the context of digital marketing, encompassing three core dimensions: matching quality, presentation quality, and interpretability quality; and (2) combining grounded theory and the Random Forest algorithm to construct a quantifiable model for assessing low-carbon information quality, analyzing the impact of features on model predictions, and providing practical guidance for enhancing information quality. This study equips digital marketers with an effective method for identifying and promoting high-quality low-carbon information, ultimately driving low-carbon consumption and contributing to a low-carbon economy.

Literature review

• Digital marketing

The concept of digital marketing can be traced back to the rise of internet commercialization in the 1990s 17 .Since the mid-2010s, the use of digital technologies for marketing purposes, known as digital marketing, has experienced rapid growth due to the widespread adoption of mobile internet and smart devices 18 . This growth has attracted significant attention across various industries 19 . Digital marketing, emerging alongside the rapid advancements in internet technology, has revolutionized industries practices by enhancing market responsiveness and fostering innovation. It empowers industries to expand their market power by reshaping firm boundaries and significantly improves customer engagement and brand loyalty, leading to increased sales growth and market share expansion 20 , 21 . These advancements offer companies a sustained competitive advantage 22 .

Technological innovation plays a pivotal role in this transformation. Blockchain technology, integrated into digital marketing management systems, has demonstrated significant improvements in sales efficiency and pre-sale consultation rates 23 . Artificial intelligence (AI), through automated recommendation systems and data analytics, optimizes marketing strategies, enabling targeted marketing and personalized services 24 . For instance, Netflix and Amazon leverage AI to recommend personalized movies and products to users, significantly enhancing user satisfaction and loyalty 20 . Big data analytics empowers businesses to understand customer needs and market trends better, facilitating more effective marketing decisions 25 . Furthermore, the integration of mobile marketing and augmented reality technologies has significantly influenced consumer purchasing decisions and shopping experiences 26 .

Despite these advancements, the adoption of digital marketing presents unique challenges and opportunities for businesses. Traditional retailers are leveraging digital innovation to reshape customer experiences and create new business models 27 However, small and medium-sized enterprises (SMEs) often face challenges in their digital marketing transformation, primarily due to limited funding and skill gaps 28 . Moreover, privacy and transparency concerns within digital marketing are increasingly gaining attention, requiring businesses to balance commercial interests with social responsibility 29 .

Consumer behavior occupies a central position within the realm of digital marketing 30 . Research indicates that digital marketing significantly influences consumer purchasing decisions and brand loyalty through personalized recommendations, social media interactions, and other strategies 31 , 32 . Furthermore, sustainable and green marketing are emerging as crucial trends in the future of digital marketing. Companies are adopting green marketing strategies to promote environmentally friendly consumption, enhance brand image, and demonstrate social responsibility.

Information quality theory

In the digital age, information quality has become a critical factor for organizational decision-making and operational efficiency across a wide range of businesses 33 . Information quality encompasses various dimensions, such as accessibility, accuracy, timeliness, and relevance, which collectively determine whether information effectively meets user needs 34 . High-quality information can significantly enhance marketing effectiveness and strengthen consumer trust 35 . A direct relationship exists between the accuracy and reliability of information content and its overall quality, emphasizing the importance of reliable information in influencing consumer decisions 36 . Effective information facilitates a deeper understanding of and response to information consumers’ needs, ultimately enhancing their consumption experience 37 . Moreover, timeliness plays a crucial role in influencing information quality 38 .

Assessing information quality requires establishing a multi-dimensional framework that comprehensively considers dimensions such as accuracy, completeness, and timeliness 34 . A combination of qualitative and quantitative methods, including questionnaires, expert reviews, and data analysis, can be employed to comprehensively evaluate information quality 39 . For instance, a mixed-methods approach, involving the definition of assessment objectives, identification of organization-specific needs, selection of relevant activities, and configuration of these activities to create an evaluation framework aligned with organizational requirements, can effectively assess information quality 40 .

The importance of information quality extends across various application domains, each presenting specific needs and challenges. Brand awareness plays a complex role in online knowledge services, impacting consumer repurchase intention and offering strategic insights for online service providers 41 . A model and methodological framework have been proposed to ensure information quality in web-based information systems, emphasizing its critical role in the accurate dissemination of web content 42 . Information quality strategies can effectively guide organizations navigating digital transformation, as demonstrated in the context of military organizations 43 . The impact of tracking information quality on customer loyalty has been investigated within the B2B logistics domain 44 .

High-quality information plays a crucial role in digital marketing, significantly enhancing the effectiveness of digital marketing campaigns, including brand awareness, user engagement, and conversion rates 22 , 45 . High-quality information directly influences user trust and purchase decisions, underlining its significance in digital marketing strategies. In the context of sustainable development, the quality of corporate carbon information disclosure is paramount. Information quality can act as an effective technological intermediary, enabling businesses to fully realize their social value and contribute to the low-carbon economy 29 , 46 .

Despite extensive research on information quality, exploring various dimensions, assessment methods, influencing factors, and applications, most studies focus on general principles and dimensions 33 . Limited attention has been given to the specific characteristics of low-carbon information in digital marketing. Information Quality Theory (IQT), while providing a valuable framework for understanding information quality, lacks the granularity to address the nuanced needs of users seeking information related to sustainable products and services 47 . Moreover, existing research on low-carbon information predominantly focuses on corporate disclosure practices and their impact on firm value, neglecting the crucial role of consumer-oriented information in promoting low-carbon consumption 46 . This study seeks to bridge the gap by developing a novel consumer-centric framework for evaluating the quality of low-carbon information specifically tailored to the digital marketing context.

Methods and materials

Research method.

This research employs a mixed-methods approach, integrating qualitative and quantitative research methods, to comprehensively explore the dimensions of low-carbon information quality and develop a corresponding evaluation model. This approach aims to provide both a theoretical foundation and practical tool for enhancing the effectiveness of low-carbon information dissemination.

Firstly, this research employs the qualitative research method of Grounded Theory (GT) to construct a dimensional model of low-carbon information quality within the context of digital marketing. Taking the perspective of low-carbon consumers, Grounded Theory is particularly suitable as it allows for the development of new theoretical concepts and models from raw data 48 . The research aims to gather subjective perspectives and experiences from various stakeholders through in-depth interviews with digital marketing practitioners, low-carbon experts, and general consumers, leading to a comprehensive understanding of the factors influencing low-carbon information quality. The GT analysis will follow a three-stage coding procedure:

Open Coding: The interview data will be analyzed line-by-line to identify and extract initial concepts and categories related to low-carbon information quality.

Axial Coding: Through methods such as cluster analysis, the concepts and categories derived from open coding will be integrated and c ategorized into higher-level themes and dimensions.

Selective Coding: From the results of axial coding, the core categories will be selected to construct the low-carbon information quality dimension model, explaining the relationships between different dimensions.

Secondly, based on the findings of the GT analysis, the research will utilize the Random Forest algorithm to construct a low-carbon information quality evaluation model and validate it using real-world data. The Random Forest algorithm is chosen for its suitability in handling multi-dimensional data, preventing overfitting, and identifying key features 49 . The application of the Random Forest algorithm will involve the following steps:

Feature Extraction: Based on the dimension model established through GT, relevant low-carbon information data will be acquired for the experimental study. Textual and multimedia features describing content quality will be extracted as input for the quantitative evaluation model.

Training Data Labeling: Experts and experimental subjects will annotate the data to construct a low-carbon information dataset with labeled samples. Dimensions from the theoretical model will be transformed into target variables for supervised learning.

Random Forest Model Training: The extracted features and labeled data will be used to train the Random Forest regression model. Hyperparameter tuning will be employed to optimize model performance.

Model Evaluation: The predictive performance of the trained Random Forest model will be assessed on a test dataset using evaluation metrics such as accuracy, recall, and F1-score. The results will shed light on the effectiveness of this model in evaluating low-carbon information quality.

By combining Grounded Theory and the Random Forest algorithm, this research aims to develop a low-carbon information quality evaluation system that is both theoretically sound and practically applicable. This approach provides a theoretical basis and technical support for enhancing the effectiveness of low-carbon information dissemination, while offering a novel mixed-methods approach to the research field.

All methods were conducted in accordance with relevant guidelines and regulations.

Ethical approval

The research design and interview protocol were approved by the Institutional Review Board (IRB) of China University of Mining and Technology. Participants were fully informed of the study procedures and provided written informed consent.

Consent to participate

Informed consent was obtained from all individual participants included in the study.

This study utilizes both qualitative and quantitative data. The following section outlines the rigorous adherence to relevant norms and standards throughout the data collection and handling processes, ensuring the reliability and validity of the research findings.

Qualitative data collection

This study primarily uses semi-structured interviews to explore the connotations and dimensions of low-carbon information quality 50 . To ensure participant cooperation and data reliability and representativeness, 30 individuals from three relevant groups were selected for interviews (Table  1 ):

Digital marketing professionals ( n  = 5): Individuals with practical experience and unique insights into low-carbon information marketing.

Experts and scholars in the low-carbon field ( n  = 5): Specialists from environmental science, energy policy, and sustainable development who can assess content quality from a professional perspective.

General consumers ( n  = 20): Randomly selected users engaged with low-carbon information, representing end consumers and reflecting the impact of content quality on them.

Initially, five respondents were randomly selected for preliminary interviews using open-ended questions (e.g., “What do you consider high-quality low-carbon information?”) to inform the design of the semi-structured interview guide. Based on feedback, the interview framework was refined, and in-depth interviews were conducted with 25 participants.

This study employed both face-to-face and remote in-depth interviews. For face-to-face interviews, questions were provided in advance, and specific times were scheduled to facilitate deeper understanding. Interviews were conducted in informal settings (e.g., cafes, offices). Before the formal interviews, researchers outlined relevant concepts of low-carbon information quality and types of digital marketing recommendations, using non-technical language to facilitate understanding. Each interview was kept to about 20 min 51 .

Quantitative data collection

To construct a robust quality assessment model for low-carbon information within digital marketing recommendations, we employed a comprehensive and systematic data collection approach, ensuring data diversity, representativeness, and timeliness (Smith, 2022). The final dataset comprised over 16,000 entries, encompassing text, images, and videos.

Data collection tools and sources

A custom web scraping tool was developed using Python’s Scrapy library (version 2.5.1) along with platform-specific APIs to gather a multifaceted dataset (Jones et al., 2023). The following data sources were utilized:

Official Media (33%): Data was scraped from the Xinhua News Agency website ( http://www.xinhuanet.com ), the People’s Daily Environmental Channel ( http://env.people.com.cn ), and specialized environmental news websites such as China Environmental News ( http://www.cenews.com.cn ) and Environmental Protection Magazine ( http://www.epc.org.cn ), using the Scrapy framework.

Social Media (46%): Data was collected from Weibo and WeChat public platforms using their respective APIs: Weibo API v2.0 and WeChat Public Platform API v3.0. This data included content from verified official Weibo accounts and WeChat official accounts of environmental organizations (e.g., @WWFChina, @Greenpeace), personal Weibo accounts and WeChat official accounts of environmental experts and key opinion leaders (KOLs), and user-generated content posted by environmental enthusiasts on both platforms.

Government Sources (21%): Data was primarily acquired by scraping publicly available information from official websites of government agencies, including the Ministry of Ecology and Environment of the People’s Republic of China ( http://www.mee.gov.cn ), the National Development and Reform Commission ( http://www.ndrc.gov.cn ), and local environmental protection bureaus at the provincial and municipal levels. The web scraper was developed in Python using the Beautiful Soup library for parsing web content.

Data collection period and frequency

Data collection spanned from January 1, 2023, to December 31, 2023, covering a full year to account for potential seasonal variations in low-carbon content. The following strategies were implemented to maintain data timeliness and comprehensiveness:

Daily Scheduled Tasks: Automated scraping of the latest information from news websites and social media platforms was performed daily at 2:00 AM using the Scrapy framework.

Weekly Updates: Data from government websites and policy documents was updated every Sunday.

Monthly Deep Dives: Archived pages were reviewed in depth on the 1st of each month to gather historical data for the past year and conduct trend analysis.

Sampling strategy and bias mitigation

To address potential biases and ensure data quality, a stratified random sampling technique was employed. Strata were defined based on source type (official media, social media, government sources) and content type (text, image, video). Within each stratum, samples were randomly selected to ensure representativeness 52 .

Specifically, we first obtained website traffic data for each source type using Alexa rankings ( https://www.alexa.com/ ). Websites were then ranked based on traffic volume and user activity. Stratified random sampling was performed proportionally to the website’s traffic share within its source type, ensuring that the collected data reflected the overall information dissemination patterns across different source types. For example, if Xinhua Net’s traffic accounted for 30% of the total traffic within official media websites, the data collected from Xinhua Net also constituted 30% of the total official media data. To mitigate potential geographical bias, data was weighted based on the population proportion of each province and municipality in China.

Technical challenges and solutions

Data collection encountered technical challenges including anti-scraping mechanisms, rate limiting, and dynamic content loading. These were addressed by implementing a random proxy mechanism with 100 proxy IPs, adhering to robots.txt guidelines, utilizing an exponential backoff algorithm for rate limiting, and employing Selenium WebDriver to handle dynamic content loading with a designated waiting time for complete page rendering 53 .

Ethical considerations and data privacy

Ethical considerations were prioritized throughout the data collection process. We strictly adhered to each platform’s terms of service and robots.txt guidelines. To protect user privacy, collected personally identifiable information (PII), such as user IDs, was anonymized using a one-way hashing algorithm, rendering it untraceable to individuals. For user-generated content, only publicly available data was collected, excluding any content marked as private or restricted. When full-text content was required, permission was obtained from the relevant platforms or users after collecting metadata and summary information 49 .

Data validation and quality assurance

A two-tier validation process was implemented to ensure the accuracy and completeness of the collected data. Firstly, automated data cleaning was performed using Python scripts to identify and address anomalies in data formatting and content. This included consistency checks for date formats, data encoding, and data types; duplicate detection using Simhash and pHash algorithms for text and images, respectively; and language detection using the langdetect library to ensure all content was in Simplified Chinese 54 . Secondly, three researchers independently reviewed a random 5% sample of the data, focusing on accuracy, relevance, and completeness. Detailed evaluation criteria and training were provided, and the researchers worked in pairs to cross-validate their assessments and reach a consensus 55 . Inter-rater reliability, calculated using Cohen’s kappa coefficient, was 0.85, exceeding the acceptable threshold of 0.8.

Data storage and management

The collected data was stored in a secure, encrypted MongoDB database (version 5.0) hosted on a private cloud server. Regular backups were performed to prevent data loss, and access to the database was restricted to authorized research team members only.

Limitations

Despite our comprehensive approach, we acknowledge several limitations: The reliance on publicly available data may not capture private discussions or sentiments about low-carbon topics. The one-year data collection period may not reflect long-term trends in low-carbon information dissemination. While we strived for representativeness, the online nature of our data sources may underrepresent segments of the population with limited internet access.

Analysis of low-carbon information quality dimensions

Building on the Information Quality Theory framework, this study employed a grounded theory qualitative research method to conduct three levels of coding on interview data. The aim is to establish a model of dimensions of low-carbon information quality that influence consumer behavior.

Open coding

Open coding is primarily used to identify categories and concepts related to the research question by breaking down, fragmenting, conceptualizing, and then recombining collected data to define categories and discover themes. In the open coding phase, 25 interview materials were first imported into Nvivo11 software, key information was extracted, irrelevant content was removed, and sentence-by-sentence coding was performed, resulting in 118 initial statements. After aggregating similar or identical concepts and removing initial concepts with a frequency of less than three, 19 initial concepts and eight low-carbon information quality categories were identified, as shown in Table  2 .

Axial coding

Open coding generated initial concepts and categories related to low-carbon information quality. However, these concepts and categories remain isolated and lack clear interrelationships. Axial coding addresses this issue by using cluster analysis to reveal the logical connections between these concepts. Through in-depth analysis of the eight categories and their interrelationships, three main categories were identified: Matching Quality, Presentation Quality, and Interpretability Quality. The specific meanings of these categories were elaborated to prepare for the next step of theoretical construction, as shown in Table  3 .

Selective coding

Selective Coding and Model Development: Selective coding involves systematically analyzing predefined categories to identify the most representative and comprehensive core categories. This process aims to integrate related concepts and categories into a theoretical framework, combining different levels of categories through a coherent narrative. The analysis indicates that Matching Quality and Presentation Quality are explicit quality requirements for low-carbon information in digital marketing recommendations, triggering low-carbon responses. Interpretability Quality, on the other hand, is an implicit condition and safeguard for low-carbon information quality. Based on these findings, a structural model was developed to examine the factors influencing the quality of low-carbon information in digital marketing (Fig.  1 ).

Model of factors influencing digital marketing-related low-carbon information quality.

Theoretical saturation testing

Theoretical saturation testing determines whether qualitative data has reached sufficient completeness. This is achieved by re-analyzing newly acquired data using open, axial, and selective coding. The absence of new core category relationships indicates that the grounded theory-derived categories and their hierarchical relationships have reached saturation.This study initially conducted interviews with 30 participants. Following data coding, no new concepts or relationships emerged. Employing the natural-occurring theoretical saturation testing method, four additional participants were interviewed, and the three-level grounded theory coding procedure was applied. The analysis revealed no new significant categories or concepts within the core categories, indicating data saturation 56 .

Through a three-stage coding process using grounded theory and theoretical saturation testing, we developed a model of low-carbon information quality in digital marketing, encompassing three core categories: Matching Quality, Presentation Quality, and Interpretability Quality. Theoretical saturation testing confirmed the distinctiveness of these categories, highlighting their representation of different aspects of low-carbon information quality and their varying influence mechanisms on user behavior.

Matching quality

This category emphasizes the alignment between low-carbon information and user needs, directly influencing user interest and attention. If information fails to match user needs, users are likely to ignore or reject it, even if the information itself is of high quality.

Presentation quality

This category emphasizes the manner in which low-carbon information is presented, directly impacting user comprehension and acceptance. If information is unclear or presented in a dull or uninteresting way, users may find it difficult to understand or be unwilling to engage with it, even if the content is valuable.

Interpretability quality

This category emphasizes the transparency and credibility of low-carbon information, directly influencing user trust and adoption. If the information source is unreliable or information disclosure lacks transparency, users may doubt the authenticity and reliability of the information, even if the content aligns with their needs.

By distinguishing these three core categories, we gain a deeper understanding of the various dimensions of low-carbon information quality and their influence mechanisms on user behavior. This understanding provides digital marketing platforms with more precise and effective guidance for developing low-carbon information dissemination strategies.

Construction of a random forest evaluation model for low-carbon information quality

Conceptual model design.

A conceptual model linking low-carbon information quality dimensions to relevant influencing factors is essential for effective assessment. Based on the three dimensions identified through qualitative research – Matching Quality (A1), Presentation Quality (A2), and Interpretability Quality (A3) – we propose the following model:

To comprehensively evaluate low-carbon information quality, we identified and quantified measurable features for each dimension (Table  3 ). The match dimension assesses the alignment between information and its target audience. The beneficiality dimension evaluates the potential benefits for the audience, while interpretability focuses on the clarity and ease of understanding. Based on these dimensions, we propose the following conceptual model, where Features 1 to n represent the various influencing factors:

Algorithm model training

Random forest algorithm is an ensemble learning algorithm based on decision trees, commonly used for classification and regression tasks. It operates by generating m subsets from the original dataset through bootstrap aggregating (bagging). Each subset is used to train a decision tree with d randomly selected features, resulting in m diverse trees. For a new sample, each tree generates a prediction, and the average of these predictions is taken as the final output. The mathematical expression for a random forest is as follows:

where X represents the input sample, Tree i represents the i -th decision tree, and N represents the total number of trees, the random forest algorithm offers several advantages for complex regression tasks like quality assessment. These include resistance to overfitting, effective handling of high-dimensional data, and the ability to provide feature importance scores. We formulated the low-carbon information quality assessment task as a machine learning regression problem, using the extracted multidimensional features as inputs and the final quality score as the output. A random forest algorithm was then employed to fit the proposed conceptual model 57 .

Evaluating the Conceptual Model involves feature engineering on raw low-carbon information data to extract relevant features corresponding to the quality dimensions. These features are then standardized and compiled into a feature matrix. This matrix, along with manually annotated content quality scores, forms the training set for a random forest regression model. The resulting model serves as the final discriminant model. To assess the quality of new low-carbon information, features are extracted and input into the trained model to predict its quality score.

Data preprocessing

Constructing a robust quality assessment model for low-carbon information necessitates a substantial amount of training data. The initial data, collected through web scraping and API calls, may contain quality issues such as duplicates, irrelevant content, and formatting errors. To address these issues, a three-step data preprocessing pipeline was implemented: data cleaning, feature engineering, and data encoding 58 .

Data cleaning

Automated data cleaning was performed using Python scripts to ensure data quality and consistency. This process involved standardizing date formats, data encoding, and data types, with inconsistent entries being either corrected or removed. Duplicate entries were identified and removed using the Simhash algorithm for text data and the pHash algorithm for image data. Additionally, the langdetect library was employed to identify and remove any content not in Simplified Chinese 55 .

Feature engineering

Based on the low-carbon information quality dimensions of Matching Quality, Presentation Quality, and Interpretability Quality, we conducted comprehensive feature engineering, extracting multi-dimensional features from the cleaned data. Open-source libraries such as Scikit-learn, NLTK, and Keras were employed for feature extraction from text, images, and video data 59 . Specifically, we constructed features according to the following sub-dimensions:

Behavioral Matching: We analyzed users’ low-carbon preferences based on their historical browsing data, likes, and favorites. We then calculated the correlation coefficients (e.g., cosine similarity) between the keywords of the current information and the user’s interests.

Scenario Matching: We assessed whether the information matched the user’s current needs based on the information’s publication time, the user’s geographical location (if available), and the user’s potential current context (e.g., work, leisure, travel). This involved using rule-based matching or machine learning models for classification.

Interest Matching: Combining “Behavioral Matching” and “Low-carbon Behavior Relevance,” we analyzed whether the information aligned with the user’s environmental values and beliefs and whether it could potentially stimulate their purchasing desire.

Content Quality: We evaluated the accuracy and completeness of the information content. Accuracy was assessed by comparing the information with existing low-carbon knowledge bases or through manual review, and a 0–1 variable was used to represent the outcome. Completeness was assessed by checking whether the information contained essential low-carbon elements, such as the product’s raw materials, production process, usage methods, and recycling methods, using a checklist or information entropy.

Format Quality: We used readability metrics, such as the Flesch-Kincaid readability ease score, to assess the readability and ease of understanding of the text. Text length, image size, and video duration were also used to evaluate conciseness. Furthermore, we assessed the vividness and appeal of the content based on text sentiment scores, image attractiveness, and video engagement. For example, we employed image recognition techniques to identify objects, scenes, and emotions in images and used speech recognition and natural language processing techniques to analyze the sentiment and theme of video content.

Public Transparency: We assessed the reliability of the information source based on the authority and credibility of the website, considering factors such as domain name, registration information, and institutional background. We also evaluated the transparency of information based on whether the information source, data source, and algorithms used were publicly disclosed and traceable. For example, we checked whether the source of information was clearly labeled, whether the data processing procedure was transparent, and whether interpretable machine learning models were used.

Privacy Assurance: We evaluated the level of privacy protection based on whether the information contained sensitive information and whether it complied with privacy regulations. This involved using regular expressions or machine learning models to identify sensitive information.

Data encoding

After feature engineering, we encoded the extracted features to prepare the data for training the Random Forest model 60 . This involved:

Standardization : All features were scaled to a range between 0 and 1 to ensure that features with different scales did not disproportionately influence the model.

One-Hot Encoding : Categorical features, such as information source type (official media, social media, government) and content type (text, image, video), were converted into numerical representations using one-hot encoding.

Label Encoding : Ordinal features, such as the level of user engagement (e.g., low, medium, high), were converted into numerical values using label encoding.

Model training

The feature matrix from feature engineering was combined with manually labeled quality scores to form a supervised learning dataset. To prevent overfitting, k-fold cross-validation (k-fold CV) was employed to divide the dataset into training and testing sets. Scikit-learn’s Random Forest Regressor was utilized for model training 61 . Hyperparameter optimization using Randomized Search CV was performed to enhance performance and generalization ability. Tuning the number of decision trees (n_estimators) with values below 100 (10, 30, 50, 80) revealed n_estimators = 30 as the optimal choice (Fig.  2 ).

Performance analysis of the random forest algorithm with varying numbers of decision trees.

After hyperparameter tuning, the final configuration for model evaluation was determined as follows: n_estimators = 30 , max_depth = 12 , min_samples_split = 4 , min_samples_leaf = 2 , max_features=’log2’ , and bootstrap = True. This configuration yielded a random forest regression model with excellent performance on the training set.

By using a tuned and trained random forest model, predictions are made on out-of-bag (OOB) data (i.e., samples not used during training). Then, each feature’s values are randomly shuffled, and predictions are made again. The importance score for each feature is obtained by comparing the accuracy differences between the two sets of predictions 62 . A higher score indicates a greater impact of that feature on the model’s prediction results, playing a more critical role in evaluating low-carbon information quality. The heatmap of feature importance scores after model training is shown in Fig.  3 .

Model evaluation

To evaluate the performance of the random forest model, we applied it to a completely new set of low-carbon information content data and compared its key performance metrics with those of several other commonly used regression methods 62 . As shown in Table  4 , the random forest model achieved the lowest Root Mean Squared Error (RMSE) of 0.12 and the highest R-squared (R²) of 0.95, significantly outperforming the other methods.

We cleaned 200 data points of low-carbon information recommended by digital marketing channels through experiments and validated by experts. The data were then divided into training and test sets with a 7:3 ratio 63 . Analysis of the residual distribution of the trained random forest model on the test set revealed an approximate normal distribution, with a mean near 0, kurtosis of 0.75, and skewness of -0.27 (Fig.  4 ). This suggests minimal overall model bias. However, identified outliers in the residual distribution indicate that the model may struggle with certain data points, leading to suboptimal prediction performance for some samples and highlighting areas for further improvement.

Residual analysis of the random forest model.

To evaluate the predictive performance of the random forest model, we visualized its results using 20 independent datasets 64 . The model demonstrated a 93% prediction accuracy for samples with quality scores ranging from 2 to 4 (The maximum score is 5). However, underfitting was observed for samples with scores above 4 and below 2. Despite these limitations, the model achieved a 93% accuracy in identifying low-carbon information content that aligns with established quality standards, demonstrating its ability to effectively distinguish high-quality content (Fig.  5 ).

Prediction performance of the random forest model.

This study aimed to identify the key dimensions of low-carbon information quality that influence consumer behavior in digital marketing recommendations (RQ1) and to develop a Random Forest model for evaluating and enhancing the quality of such information (RQ2).

Compared with existing research, this study contributes to a more nuanced understanding of low-carbon information quality in several key ways. Firstly, this research introduces a novel consumer-centric framework for evaluating low-carbon information quality, specifically tailored to the digital marketing context. Secondly, this study identifies three core dimensions of low-carbon information quality: Matching Quality, Presentation Quality, and Interpretability Quality, which extends the traditional Information Quality Theory (IQT) framework by incorporating consumer-centric considerations. Finally, this research develops a quantifiable model, based on grounded theory and the Random Forest algorithm, for assessing low-carbon information quality, providing practical guidance for digital marketers to enhance their strategies and promote sustainable consumption.

Regarding RQ1, our findings, derived from both qualitative and quantitative analyses, revealed three core dimensions of low-carbon information quality: Matching Quality, Presentation Quality, and Interpretability Quality.

Matching Quality: This dimension emphasizes the alignment between low-carbon information and user needs, directly influencing user interest and attention. Our qualitative analysis revealed that users are more receptive to information that aligns with their behavioral habits, interests, and usage scenarios. For example, P15 and P17 expressed preferences for information tailored to their individual low-carbon interests and past browsing behavior. This observation is further supported by our Random Forest model, which identified “correlation with low-carbon behavior” and “low-carbon interest alignment” as two of the most influential features in predicting low-carbon information quality. This finding highlights the importance of leveraging user data to personalize low-carbon information and ensure its relevance to individual consumers. By providing information that resonates with users’ existing knowledge, values, and behavior patterns, digital marketers can capture attention, stimulate interest, and ultimately increase the likelihood of engagement with low-carbon content.

Presentation Quality: This dimension highlights the importance of conveying low-carbon information in a clear, engaging, and easily understandable manner. Our qualitative analysis indicated that users value accuracy and completeness in information presentation. For instance, P05 emphasized the importance of accurate product descriptions that clearly state low-carbon characteristics, while P09 highlighted the need for comprehensive information, including certifications and data, to support low-carbon claims. Furthermore, users emphasized the need for vivid and concise presentation styles. P07 expressed a preference for dynamic information that brings low-carbon features to life, while P12 valued concise messaging that focuses on key points. These preferences align with the findings from our Random Forest model, which identified “content accuracy,” “information completeness,” and “content vividness” as significant predictors of low-carbon information quality. Therefore, digital marketing platforms should prioritize both the factual accuracy of information and its presentation style to ensure user comprehension and engagement.

Interpretability Quality: This dimension emphasizes the transparency and credibility of low-carbon information, which directly influences user trust and adoption. Our qualitative analysis revealed that users are more likely to trust information from reliable sources that are transparently presented and protect user privacy. For instance, P02 highlighted that platform endorsements enhance user trust in recommendations, while P04 emphasized the importance of privacy protection in building user confidence. While not directly incorporated as features in our quantitative analysis, interpretability can be considered a foundational aspect of information quality. Users are less likely to engage with information they perceive as untrustworthy or opaque, regardless of its relevance or presentation. Therefore, platforms should prioritize building user trust by establishing robust information review mechanisms, transparently disclosing information sources, and strictly adhering to data security and privacy regulations.

Addressing RQ2, our findings indicate that the Random Forest model is an effective tool for evaluating low-carbon information quality. The model achieved a 93% accuracy rate in identifying high-quality low-carbon information, confirming the significant influence of the three identified dimensions. However, the model exhibited limitations in identifying information with extremely high or low quality scores, indicating potential underfitting in these cases. This suggests that while the model shows promise, there is room for improvement in its ability to handle extreme cases. Future research can address this by collecting a larger, more diverse dataset encompassing a wider range of quality levels and exploring alternative machine learning algorithms, such as deep learning models, which may be better suited to capturing complex relationships and nuanced features in the data.

The key dimensions of low-carbon information quality identified in this study, particularly Matching Quality and Interpretability Quality, hold significant implications for broader sustainability initiatives. Aligning information with consumer needs and preferences (Matching Quality) and ensuring transparency and trustworthiness (Interpretability Quality) are crucial for promoting sustainable choices in various contexts. For instance, in facilitating the adoption of renewable energy technologies, providing tailored information that caters to consumers’ specific needs and knowledge levels, while ensuring the credibility of information sources, can significantly enhance consumer confidence and encourage a transition towards cleaner energy sources. Similarly, in the realm of green public procurement and government-led initiatives for low-carbon development, ensuring the accuracy, relevance, and timeliness of low-carbon information, presented in a clear and accessible manner, is paramount for guiding consumer and corporate decision-making towards sustainable practices 65 , 66 , 67 .

This study contributes to a more comprehensive understanding of low-carbon information quality by combining empirical data with a consumer-centric perspective. Our proposed model provides a nuanced framework for evaluating and enhancing the quality of low-carbon information disseminated through digital marketing channels, ultimately contributing to the promotion of sustainable consumption.

Research conclusions, limitations, and prospects

This study aimed to explore the key dimensions of low-carbon information quality influencing digital marketing recommendations and develop a Random Forest model to evaluate and enhance the quality of such information, ultimately promoting sustainable consumer choices. By integrating information quality theory and digital marketing theory, this study proposes a novel low-carbon information quality model encompassing three core dimensions: Matching Quality, Presentation Quality, and Interpretability Quality. We employed grounded theory and conducted in-depth interviews ( n  = 30) with digital marketing practitioners, low-carbon experts, and general consumers. The collected data was then used to construct and validate a Random Forest evaluation model. Our findings confirmed that the three dimensions—Matching Quality, Presentation Quality, and Interpretability Quality—significantly influenced the perceived quality of low-carbon information. Additionally, the constructed model effectively identified high-quality low-carbon information.

Theoretical contributions

This study makes several significant theoretical contributions to the existing literature on information quality and low-carbon consumption. Firstly, we extend the application of Information Quality Theory (IQT) to the domain of digital marketing, demonstrating its relevance in this increasingly important context for promoting sustainable consumption. However, traditional IQT frameworks primarily focus on general information quality criteria and lack the granularity to address the unique characteristics of low-carbon information and the heterogeneity of user needs in this domain. This research addresses this gap by introducing “Matching Quality” as a new dimension of low-carbon information quality. This dimension, defined as the alignment between low-carbon information and user’s low-carbon behaviors and preferences, is crucial for effectively influencing consumer choices toward sustainability. Unlike previous studies that primarily focus on corporate information disclosure, we develop a consumer-oriented low-carbon information quality evaluation framework. This framework adopts a consumer-centric perspective and recognizes the importance of understanding and addressing the information needs of individual consumers, which is vital for promoting sustainable consumption.

To enhance low-carbon information quality for effective digital marketing strategies, low-carbon product enterprises should prioritize the development of accurate user profiles to enable personalized low-carbon recommendations. Furthermore, clarity, conciseness, and engaging presentation styles are crucial for effective information dissemination. Additionally, establishing robust information review mechanisms and transparently disclosing information sources are essential for fostering user trust.

Limitations and future research directions

This study acknowledges limitations regarding potential model underfitting when identifying information with extreme quality scores, possibly due to an imbalanced training dataset. Future research should prioritize collecting larger, more diverse datasets encompassing various low-carbon information sources, content types, and quality levels to enhance model robustness and generalizability. Additionally, refining evaluation metrics, exploring alternative machine learning algorithms (e.g., deep learning), incorporating granular user behavior data (e.g., clicks, shares, comments), and examining the influence mechanisms of high-quality low-carbon information on consumer behavior are recommended avenues for further investigation.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Agnew, M. D., Pettifor, H. & Wilson, C. Lifestyle, an integrative concept: Cross-disciplinary insights for low-carbon research. Wiley Interdiscip. Rev. Energy Environ. 12 (6). https://doi.org/10.1002/wene.490 (2023).

Díaz-Padilla, V. T., Travar, I., Acosta-Rubio, Z. & Parra-López, E. Tourism competitiveness versus sustainability: Impact on the world economic forum model using the. Rasch Methodol. Sustain. 15 (18), 13700. https://doi.org/10.3390/su151813700 (2023).

Article   Google Scholar  

Jain, P., Chou, M. C., Fan, F. & Santoso, M. P. Embedding sustainability in the consumer goods innovation cycle and enabling tools to measure progress and capabilities. Sustainability 13 (12), 6662. https://doi.org/10.3390/su13126662 (2021).

Wei, J., Zhang, L., Yang, R. & Song, M. A new perspective to promote sustainable low-carbon consumption: The influence of informational incentive and social influence. J. Environ. Manag. 327 , 116848. https://doi.org/10.1016/j.jenvman.2022.116848 (2023).

Zhang, J., Lyu, Y., Li, Y. & Geng, Y. Digital economy: an innovation driving factor for low-carbon development. Environ. Impact Assess. Rev. 96 , 106821. https://doi.org/10.1016/j.eiar.2022.106821 (2022).

Denga, E. M., Vajjhala, N. R. & Rakshit, S. The role of digital marketing in achieving sustainable competitive advantage. Digit. Transform. Int. Strateg. Organ. 44–60. https://doi.org/10.3390/su132413679 (2022).

Lewis, J. I. & Nemet, G. F. Assessing learning in low carbon technologies: Toward a more comprehensive approach. Wiley Interdiscip. Rev. Clim. Change 12 (5), e730. https://doi.org/10.1002/wcc.730 (2021).

Miguel, A. & Miranda, S. The role of digital platforms in promoting pro-sustainable behavior and conscious consumption by brands. Ecocycles 9 (2), 37–48. https://doi.org/10.19040/ecocycles.v9i2.298 (2023).

Yang, S., Jahanger, A. & Hossain, M. R. Does China’s low-carbon city pilot intervention limit electricity consumption? An analysis of industrial energy efficiency using time-varying DID model. Energy Econ. 121 , 106636. https://doi.org/10.1016/j.eneco.2023.106636 (2023).

Castro-Santa, J., Drews, S. & Bergh, J. Nudging low-carbon consumption through advertising and social norms. J. Behav. Exp. Econ. 102 , 101956. https://doi.org/10.1016/j.socec.2022.101956 (2023).

Wang, T., Shen, B., Springer, C. H. & Hou, J. What prevents us from taking low-carbon actions? A comprehensive review of influencing factors affecting low-carbon behaviors. Energy Res. Social. Sci. 71 , 101844. https://doi.org/10.1016/j.erss.2020.101844 (2021).

Wu, Z., Duan, C., Cui, Y. & Qin, R. Consumers’ attitudes toward low-carbon consumption based on a computational model: Evidence from China. Technol. Forecast. Soc. Chang. 186 , 122119. https://doi.org/10.1016/j.techfore.2022.122119 (2023).

Zhang, L. et al. A data-driven approach to objective evaluation of urban low carbon development performance. J. Clean. Prod. 368 , 133238. https://doi.org/10.1016/j.jclepro.2022.133238 (2022).

Wang, R. Y. & Strong, D. M. Beyond accuracy: What data quality means to data consumers. J. Manag. Inform. Syst. 12 (4), 5–33. https://doi.org/10.1080/07421222.1996.11518099 (1996).

Del Rio, D. D. F., Sovacool, B. K. & Griffiths, S. Culture, energy and climate sustainability, and smart home technologies: A mixed methods comparison of four countries. Energy Clim. Change 2 , 100035. https://doi.org/10.1016/j.egycc.2021.100035 (2021).

Cheng, X., Wu, F., Li, W., Yang, J. & Long, R. What maintains low-carbon consumption behaviors: Evidence from China. Renew. Sustain. Energy Rev. 189 , 114050. https://doi.org/10.1016/j.rser.2023.114050 (2024).

Peterson, R. A., Balasubramanian, S. & Bronnenberg, B. J. Exploring the implications of the internet for consumer marketing. J. Acad. Mark. Sci. 25 , 329–346. https://doi.org/10.1177/0092070397254005 (1997).

Sağkaya Güngör, A. & Ozansoy Çadırcı, T. Understanding digital consumer: A review, synthesis, and future research agenda. Int. J. Consum. Stud. 46 (5), 1829–1858. https://doi.org/10.1111/ijcs.12809 (2022).

Hofacker, C., Golgeci, I., Pillai, K. G. & Gligor, D. M. Digital marketing and business-to-business relationships: A close look at the interface and a roadmap for the future. Eur. J. Market. 54 (6), 1161–1179. https://doi.org/10.1108/EJM-04-2020-0247 (2020).

Xu, Y. & Li, C. M. Digital transformation, firm boundaries, and market power: Evidence from china’s listed companies. Systems 11 (9), 479. https://doi.org/10.3390/systems11090479 (2023).

Sanbella, L., Van Versie, I. & Audiah, S. Online marketing strategy optimization to increase sales and e-commerce development: An integrated approach in the digital age. Startupreneur Bus. Digit. 3 (1), 54–66. https://doi.org/10.33050/sabda.v3i1.492 (2024).

Liu, Y., Suo, X. K., Du, X. H., Wu, H. Q. & Lin, H. Corporate digital innovation and stock price crash risk. Finance Res. Lett. 66 , Article 105690. https://doi.org/10.1016/j.frl.2024.105690 (2024).

Yang, R. Q. & Jiang, H. C. Digital marketing management control system based on blockchain under the internet background. Soft. Comput. https://doi.org/10.1007/s00500-023-08352-0 (2023).

Article   PubMed   PubMed Central   Google Scholar  

Chintalapati, S. & Pandey, S. K. Artificial intelligence in marketing: A systematic literature review. Int. J. Market Res. 64 (1), 38–68. https://doi.org/10.1177/14707853211018428 (2022).

Varzaru, A. A. Assessing digital transformation acceptance in public organizations’ marketing. Sustainability 15 (1), Article 265. https://doi.org/10.3390/su15010265 (2023).

Sahli, A. & Lichy, J. The role of augmented reality in the customer shopping experience. Int. J. Organ. Anal. https://doi.org/10.1108/IJOA-02-2024-4300 (2024).

Jocevski, M. Blurring the lines between physical and digital spaces: Business model innovation in retailing. Calif. Manag. Rev. 63 (1), 99–117. https://doi.org/10.1177/000812562095 (2020).

Bradač Hojnik, B. & Huđek, I. Small and medium-sized enterprises in the digital age: Understanding characteristics and essential demands. Information 14 (11), 606. https://doi.org/10.3390/info14110606 (2023).

Peng, H., Bumailikaimu, S. & Feng, T. The power of market: venture capital and enterprise digital transformation. North Am. J. Econ. Finance 74 , Article 102218. https://doi.org/10.1016/j.najef.2024.102218 (2024).

Dahiya, R. & Gayatri A research paper on digital marketing communication and consumer buying decision process: An empirical study in the Indian passenger car market. J. Global Mark. 31 (2), 73–95. https://doi.org/10.1080/08911762.2017.1365991 (2018).

Kim, J., Kang, S. & Lee, K. H. Evolution of digital marketing communication: Bibliometric analysis and network visualization from key articles. J. Bus. Res. 130 , 552–563. https://doi.org/10.1016/j.jbusres.2019.09.043 (2021).

Wang, H., Wu, D. L. & Zeng, Y. M. Digital economy, market segmentation and carbon emission performance. Environ. Dev. Sustain. https://doi.org/10.1007/s10668-023-03465-w (2023).

Surbakti, F. P. S., Wang, W., Indulska, M. & Sadiq, S. Factors influencing effective use of big data: A research framework. Inf. Manag. 57 (1), 103146. https://doi.org/10.1016/j.im.2019.02.001 (2020).

Bovee, M., Srivastava, R. P. & Mak, B. A conceptual framework and belief-function approach to assessing overall information quality. Int. J. Intell. Syst. 18 (1), 51–74. https://doi.org/10.1002/int.10074 (2003).

Rane, N. L., Achari, A. & Choudhary, S. P. Enhancing customer loyalty through quality of service: Effective strategies to improve customer satisfaction, experience, relationship, and engagement. Int. Res. J. Modernization Eng. Technol. Sci. 5 (5), 427–452. https://doi.org/10.56726/IRJMETS38104 (2023).

Lăzăroiu, G., Neguriţă, O., Grecu, I., Grecu, G. & Mitran, P. C. Consumers’ decision-making process on social commerce platforms: Online trust, perceived risk, and purchase intentions. Front. Psychol. 11 , 890. https://doi.org/10.3389/fpsyg.2020.00890 (2020).

Godoy, M. P., Rusu, C., Hatibovic, F., Granollers, T. & Ugalde, J. Addressing information consumer experience through a user-centered information management system in a Chilean university. Sustainability 15 (22), Article 15998. https://doi.org/10.3390/su152215998 (2023).

Kim, H. & Niehm, L. S. The impact of website quality on information quality, value, and loyalty intentions in apparel retailing. J. Interact. Mark. 23 (3), 221–233 (2009). https://www.taylorfrancis.com/chapters/edit/10.4324/9780429293276-3

Venkatesh, V., Brown, S. A. & Bala, H. Bridging the qualitative-quantitative divide: Guidelines for conducting mixed methods research in information systems. MIS Q. 21–54. https://doi.org/10.25300/MISQ/2013/37.1.02 (2013).

Al-Fraihat, D., Joy, M. & Sinclair, J. Evaluating E-learning systems success: An empirical study. Comput. Hum. Behav. 102 , 67–86. https://doi.org/10.1016/j.chb.2019.08.004 (2020).

Du, H. L. S., Xu, J. H., Tang, H. & Jiang, R. X. Repurchase intention in online knowledge service: The brand awareness perspective. J. Comput. Inform. Syst. 62 (1), 174–185. https://doi.org/10.1080/08874417.2020.1759159 (2022).

Naim, A. & Alahmari, F. Reference model of e-learning and quality to establish interoperability in higher education systems. Int. J. Emerg. Technol. Learn. 15 (2), 15–28. https://doi.org/10.3991/ijet.v15i02.11605 (2020).

Struijk, M., Angelopoulos, S., Ou, C. X. & Davison, R. M. Navigating digital transformation through an information quality strategy: Evidence from a military organization. Inform. Syst. J. 33 (4), 912–952. https://doi.org/10.1111/isj.12430 (2023).

Rai, A., Tang, X., Yin, Z. & Du, S. Gaining customer loyalty with tracking information quality in B2B logistics. J. Manag. Inform. Syst. 39 (2), 307–335. https://doi.org/10.1080/07421222.2022.2063552 (2022).

Alterkait, M. A. & Alduaij, M. Y. Impact of information quality on satisfaction with e-learning platforms: Moderating role of instructor and learner quality. SAGE OPEN 14 (1), Article 21582440241233400. https://doi.org/10.1177/21582440241233400 (2024).

Cao, Q., Zhou, Y., Du, H., Ren, M. & Zhen, W. Carbon information disclosure quality, greenwashing behavior, and enterprise value. Front. Psychol. 13 , 892415. https://doi.org/10.3389/fpsyg.2022.892415 (2022).

Zha, D., Zhang, C., Jiang, P. & Wang, F. What makes energy consumption behavior visible? Conceptualization, scale development and validation of customized information feedback. J. Bus. Res. 182 , 114761. https://doi.org/10.1016/j.jbusres.2024.114761 (2024).

Mills, J., Bonner, A. & Francis, K. The development of constructivist grounded theory. Int. J. Qual. Methods 5 (1), 25–35. https://doi.org/10.1177/160940690600500103 (2006).

Lin, W., Wu, Z., Lin, L., Wen, A. & Li, J. An ensemble random forest algorithm for insurance big data analysis. IEEE Access 5 , 16568–16575. https://doi.org/10.1109/ACCESS.2017.2738069 (2017).

Gustavsson, E. & Elander, I. Behaving clean without having to think green? Local eco-technological and dialogue-based, low-carbon projects in Sweden. J. Urban Technol. 24 (1), 93–116. https://doi.org/10.1080/10630732.2016.1175806 (2017).

Sturges, J. E. & Hanrahan, K. J. Comparing telephone and face-to-face qualitative interviewing: A research note. Qualitative Res. 4 (1), 107–118. https://doi.org/10.1177/1468794104041110 (2004).

Hurwitz, L. B. et al. Content analysis across new media platforms: Methodological considerations for capturing media-rich data. New. Media Soc. 20 (2), 532–548. https://doi.org/10.1177/1461444816663927 (2018).

Roumeliotis, K. I. & Tselikas, N. D. A machine learning python-based search engine optimization audit software. In Informatics 10 (3), 68. https://doi.org/10.3390/informatics10030068 (2023).

Pamulaparty, L., Rao, C. G. & Rao, M. S. Critical review of various near-duplicate detection methods in web crawl and their prospective application in drug discovery. Int. J. BioMed. Eng. Technol. 25 (2–4), 212–226. https://doi.org/10.1504/IJBET.2017.087723 (2017).

Sun, J. Improving Quality of Programming and Software Through Knowledge Graph Construction and Application . Doctoral dissertation (The Australian National University, 2023). https://www.proquest.com/openview/5b5e3c1260a05ccb1313dace42759240

Peters, H. C. A-methodological saturation: A grounded theory analysis. Couns. Psychol. 51 (7), 933–969. https://doi.org/10.1177/00110000231179706 (2023).

Biau, G. & Scornet, E. A random forest guided tour. Test 25 , 197–227. https://doi.org/10.1007/s11749-016-0481-7 (2016).

Article   MathSciNet   Google Scholar  

Gupta, S., Aga, D., Pruden, A., Zhang, L. & Vikesland, P. Data analytics for environmental science and engineering research. Environ. Sci. Technol. 55 (16), 10895–10907. https://doi.org/10.1021/acs.est.1c01026 (2021).

Article   ADS   CAS   PubMed   Google Scholar  

Jantunen, E., Campos, J., Sharma, P. & McKay, M. Open source analytics solutions for maintenance. In 5th International Conference on Control, Decision and Information Technologies (CoDIT) , vol. 8, 688–693. (IEEE, 2018). https://doi.org/10.1109/CoDIT.2018.8394819 .

Yin, Y., Alqahtani, Y., Feng, J. H., Chakraborty, J. & McGuire, M. P. Classification of eye tracking data in visual information processing tasks using convolutional neural networks and feature engineering. SN Comput. Sci. 2 , 1–26. https://doi.org/10.1007/s42979-020-00444-0 (2021).

Kadiyala, A. & Kumar Applications of python to evaluate the performance of bagging methods. Environ. Prog. Sustain. Energy 37 (5), 1555–1559. https://doi.org/10.1002/ep.13018 (2018).

Article   CAS   Google Scholar  

Farnaaz, N. & Jabbar, M. A. Random forest modeling for network intrusion detection system. Procedia Comput. Sci. 89 , 213–217. https://doi.org/10.1016/j.procs.2016.06.047 (2016).

Yin, H. et al. A deep learning-based data-driven approach for predicting mining water inrush from coal seam floor using micro-seismic monitoring data. IEEE Trans. Geosci. Remote Sens. https://doi.org/10.1109/TGRS.2023.3300012 (2023).

Zermane, A., Tohir, M. Z. M., Zermane, H., Baharudin, M. R. & Yusoff, H. M. Predicting fatal fall from heights accidents using random forest classification machine learning model. Saf. Sci. 159 , 106023. https://doi.org/10.1016/j.ssci.2022.106023 (2023).

Yin, H. T., Wen, J. & Chang, C. P. Going green with artificial intelligence: The path of technological change towards the renewable energy transition. Oecon. Copernic. 14 (4), 1059–1095. https://doi.org/10.24136/oc.2023.032 (2023).

Lăzăroiu, G. et al. Environmentally responsible behavior and sustainability policy adoption in green public procurement. Sustainability 12 (5), 2110. https://doi.org/10.3390/su12052110 (2020).

Bai, T. et al. Paths to low-carbon development in China: The role of government environmental target constraints. Oecon. Copernic. 14 (4), 1139–1173. https://doi.org/10.24136/oc.2023.034 (2023).

Download references

Author information

Authors and affiliations.

School of Economics and Management, China University of Mining and Technology, Xuzhou, 221116, China

Weiji Gao & Zhihua Ding

Jiangsu Vocational College of Electronic and Information, School of Business, Huai’an, 223001, China

Weiji Gao & Yulong Wan

Shanghai National Accounting Institute, Shanghai, 21000, China

You can also search for this author in PubMed   Google Scholar

Contributions

All authors contributed to the study meaningfully, providing feedback on the interpretation of the results and on the writing of the paper.

Corresponding author

Correspondence to Weiji Gao .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .

Reprints and permissions

About this article

Cite this article.

Gao, W., Ding, Z., Lu, J. et al. Low-carbon information quality dimensions and random forest algorithm evaluation model in digital marketing. Sci Rep 14 , 22416 (2024). https://doi.org/10.1038/s41598-024-72910-1

Download citation

Received : 03 May 2024

Accepted : 11 September 2024

Published : 28 September 2024

DOI : https://doi.org/10.1038/s41598-024-72910-1

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Information quality evaluation
• Random forest

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: Anthropocene newsletter — what matters in anthropocene research, free to your inbox weekly.

• Browse All Articles
• Newsletter Sign-Up

Marketing →

• 17 Sep 2024
• Cold Call Podcast

Fawn Weaver’s Entrepreneurial Journey as an Outsider in the Spirits Industry

In 2017 Fawn Weaver launched a premium American whiskey brand, Uncle Nearest. It became the fastest growing and most awarded whiskey brand in America, despite the challenges Weaver faced as a Black woman and outsider to the spirits industry, which is capital-intensive, highly regulated, competitive, and male-dominated. In October 2023, Weaver announced plans to expand into cognac with the goal of building the next major alcoholic beverages conglomerate. But the company was still heavily reliant on capital. How could Weaver convince new investors that her plans for cognac would yield success? Harvard Business School senior lecturer Hise Gibson discusses Weaver’s leadership style, growth strategies, and her use of storytelling to connect customers with her brand in the case, "Uncle Nearest: Creating a Legacy."

• 20 Aug 2024

Angel City Football Club: A New Business Model for Women’s Sports

Angel City Football Club (ACFC) was founded in 2020 by venture capitalist Kara Nortman, entrepreneur Julie Uhrman, and actor and activist Natalie Portman. As outsiders to professional sports, the all-female founding team had rewritten the playbook for how to build a sports franchise by applying lessons from the tech and entertainment industries. Unlike typical sports franchises that built their teams and track records over many years before extending their brand beyond a local base, ACFC had inverted the model, generating both global and local interest in the club during its first three years. The club’s early success was reflected in its market valuation of $250 million as of its sale in July 2024 — the highest in the National Women’s Soccer League. Equally important, ACFC had started to bend the curve toward greater pay equity in women’s sports — the club’s ultimate goal. But the founders knew there was much more to do to capitalize on the club’s momentum. As they developed ACFC’s first three-year strategic plan in 2024, they weighed the most effective ways to build value for the franchise. Was it better to allocate the incremental budget to investments in digital brand building or to investments in the on-field product? Senior Lecturer Jeffrey Rayport is joined by case co-author Nicole Keller and club co-founder Kara Nortman to discuss the case, “Angel City Football Club: Scoring a New Model.”

Why Competing With Tech Giants Requires Finding Your Own Edge

In the new book Smart Rivals, Feng Zhu and Bonnie Yining Cao show business leaders how to create competitive advantages by uncovering their hidden strengths and leveraging their individual capabilities.

• 06 Aug 2024

How EdTech Firm Coursera Is Incorporating GenAI into Its Products and Services

In early 2023, Jeff Maggioncalda, CEO of Coursera, started developing the EdTech firm’s strategy for incorporating GenAI into their offerings. By early 2024, the firm had made significant progress in bringing four key capabilities to market, but GenAI was evolving quickly and Coursera needed to continuously improve its offerings. While the firm had been an early mover, competitors were adapting fast. Was Coursera taking full advantage of the opportunities presented by the technology? What more could it do to remain competitive?

• 09 Jul 2024

Non-Fungible Tokens (NFTs) and Brand Building

Non-fungible tokens (NFTs), which allow individuals to own their digital assets and move them from place to place, are changing the interaction between consumers and digital goods, brands, and platforms. Professor Scott Duke Kominers and tech entrepreneur Steve Kaczynski discuss the case, “Bored Ape Yacht Club: Navigating the NFT World,” and the related book they co-authored, The Everything Token: How NFTs and Web3 Will Transform The Way We Buy, Sell, And Create. They focus on the rise and popularity of the Bored Ape Yacht Club NFTs and the new model of brand building created by owning those tokens.

• 07 May 2024

Lessons in Business Innovation from Legendary Restaurant elBulli

Ferran Adrià, chef at legendary Barcelona-based restaurant elBulli, was facing two related decisions. First, he and his team must continue to develop new and different dishes for elBulli to guarantee a continuous stream of innovation, the cornerstone of the restaurant's success. But they also need to focus on growing the restaurant’s business. Can the team balance both objectives? Professor Michael I. Norton discusses the connections between creativity, emotions, rituals, and innovation – and how they can be applied to other domains – in the case, “elBulli: The Taste of Innovation,” and his new book, The Ritual Effect.

• 29 Feb 2024

Beyond Goals: David Beckham's Playbook for Mobilizing Star Talent

Reach soccer's pinnacle. Become a global brand. Buy a team. Sign Lionel Messi. David Beckham makes success look as easy as his epic free kicks. But leveraging world-class talent takes discipline and deft decision-making, as case studies by Anita Elberse reveal. What could other businesses learn from his ascent?

• 17 Jan 2024

Psychological Pricing Tactics to Fight the Inflation Blues

Inflation has slowed from the epic rates of 2021 and 2022, but many consumers still feel pinched. What will it take to encourage them to spend? Thoughtful pricing strategies that empower customers as they make purchasing decisions, says research by Elie Ofek.

• 05 Dec 2023

What Founders Get Wrong about Sales and Marketing

Which sales candidate is a startup’s ideal first hire? What marketing channels are best to invest in? How aggressively should an executive team align sales with customer success? Senior Lecturer Mark Roberge discusses how early-stage founders, sales leaders, and marketing executives can address these challenges as they grow their ventures in the case, “Entrepreneurial Sales and Marketing Vignettes.”

Tommy Hilfiger’s Adaptive Clothing Line: Making Fashion Inclusive

In 2017, Tommy Hilfiger launched its adaptive fashion line to provide fashion apparel that aims to make dressing easier. By 2020, it was still a relatively unknown line in the U.S. and the Tommy Hilfiger team was continuing to learn more about how to serve these new customers. Should the team make adaptive clothing available beyond the U.S., or is a global expansion premature? Assistant Professor Elizabeth Keenan discusses the opportunities and challenges that accompanied the introduction of a new product line that effectively serves an entirely new customer while simultaneously starting a movement to provide fashion for all in the case, “Tommy Hilfiger Adaptive: Fashion for All.”

• Research & Ideas

Are Virtual Tours Still Worth It in Real Estate? Evidence from 75,000 Home Sales

Many real estate listings still feature videos and interactive tools that simulate the experience of walking through properties. But do they help homes sell faster? Research by Isamar Troncoso probes the post-pandemic value of virtual home tours.

• 17 Oct 2023

With Subscription Fatigue Setting In, Companies Need to Think Hard About Fees

Subscriptions are available for everything from dental floss to dog toys, but are consumers tiring of monthly fees? Elie Ofek says that subscription revenue can provide stability, but companies need to tread carefully or risk alienating customers.

• 29 Aug 2023

As Social Networks Get More Competitive, Which Ones Will Survive?

In early 2023, TikTok reached close to 1 billion users globally, placing it fourth behind the leading social networks: Facebook, YouTube, and Instagram. Meanwhile, competition in the market for videos had intensified. Can all four networks continue to attract audiences and creators? Felix Oberholzer-Gee discusses competition and imitation among social networks in his case “Hey, Insta & YouTube, Are You Watching TikTok?”

• 26 Jun 2023

Want to Leave a Lasting Impression on Customers? Don't Forget the (Proverbial) Fireworks

Some of the most successful customer experiences end with a bang. Julian De Freitas provides three tips to help businesses invest in the kind of memorable moments that will keep customers coming back.

• 31 May 2023

With Predictive Analytics, Companies Can Tap the Ultimate Opportunity: Customers’ Routines

Armed with more data than ever, many companies know what key customers need. But how many know exactly when they need it? An analysis of 2,000 ridesharing commuters by Eva Ascarza and colleagues shows what's possible for companies that can anticipate a customer's routine.

• 30 May 2023

Can AI Predict Whether Shoppers Would Pick Crest Over Colgate?

Is it the end of customer surveys? Definitely not, but research by Ayelet Israeli sheds light on the potential for generative AI to improve market research. But first, businesses will need to learn to harness the technology.

• 24 Apr 2023

What Does It Take to Build as Much Buzz as Booze? Inside the Epic Challenge of Cannabis-Infused Drinks

The market for cannabis products has exploded as more states legalize marijuana. But the path to success is rife with complexity as a case study about the beverage company Cann by Ayelet Israeli illustrates.

• 07 Apr 2023

When Celebrity ‘Crypto-Influencers’ Rake in Cash, Investors Lose Big

Kim Kardashian, Lindsay Lohan, and other entertainers have been accused of promoting crypto products on social media without disclosing conflicts. Research by Joseph Pacelli shows what can happen to eager investors who follow them.

• 10 Feb 2023

COVID-19 Lessons: Social Media Can Nudge More People to Get Vaccinated

Social networks have been criticized for spreading COVID-19 misinformation, but the platforms have also helped public health agencies spread the word on vaccines, says research by Michael Luca and colleagues. What does this mean for the next pandemic?

• 02 Feb 2023

Why We Still Need Twitter: How Social Media Holds Companies Accountable

Remember the viral video of the United passenger being removed from a plane? An analysis of Twitter activity and corporate misconduct by Jonas Heese and Joseph Pacelli reveals the power of social media to uncover questionable situations at companies.