Introduction

During the COVID-19 pandemic, consumers have increased their online shopping volume. In Brazil, in the first half of 2020, 40% more consumers were registered on digital platforms compared to the same period in 2019, totalizing 41 million new users (Nielsen; Elo 2020). However, many Brazilian companies have not been able to keep up with these transformations and around 600,000 have closed their doors since the beginning of the pandemic, from 2019 to 2020 (IBGE 2021). The country also has very low rates of innovation (Global Innovation Index 2022), which confirms the difficulty for companies to survive in the digital environment.

Without being able to adopt an effective digital strategy, Brazilian retailers lose competitiveness, as a business can no longer exist without a great presence on the internet (Dykha et al. 2021). According to the Resource-Advantage (R-A) Theory, companies are constantly fighting for comparative advantages. In today´s dynamic and hypercompetitive economy, they need to renew their strategies in the market. Thus, obtaining and using reliable information, which is heterogeneous, become increasingly important (Hunt and Madhavaram 2020; Varadarajan 2020). The problem is that most organizations are not managing to deal with it effectively, and the literature is still not clear on how to best do this. Managers seem to lack clarity on the different elements that need to be considered in their efforts to digitally transform their business (Salume et al. 2021).

In fact, the response from most professionals has been to use a variety of tactical tools to address specific needs. This tactical approach, without connection to a broader marketing strategy, may be interesting for a quick prompt decision, but it does not provide alignment or strategic learning for more effective decisions in terms of performance (Kovala et al. 2017). Also, as the complexity of operations has increased, specialists in subsections of the business have emerged, such as SEO or social media specialists. This fragmentation further contributes to the formation of silos, which makes it difficult to integrate information into the business to formulate a broader strategy. Thus, if companies do not adopt a strategic mindset, they risk losing competitive advantages. Ultimately, marketing should be seen as an overall process, not a separate set of functions (Joensuu-Salo et al. 2018).

As a consequence, although many investments in Big Data were made, they were gradually diminished because organizations were failing to understand its value (Ghasemaghaei and Calic 2019). On the other hand, some studies show that companies that use data analytics outperform those that do not (Järvinen 2016; LaValle et al. 2011). Furthermore, the acquisition of technology helps the company to use knowledge to increase performance (Caputo et al. 2019). Thus, according to Wang et al. (2020), it is not the lack of data or interest that inhibits companies, but knowing strategically when and how to use them.

As the use of technological tools by firms and consumers increases, so does the amount of data. Likewise, a central question in marketing today is: how to use all the information available? (De Luca et al. 2021). Unfortunately, the literature also still does not provide a clear understanding of how analytics tools can be used to guide decisions and improve the company's strategy. In reality, scholars tried to answer questions related to “how” and “what” in the scope of marketing analytics, offering definitions, techniques, benefits, applications, and evaluations of their impact. However, there is still no consensus on which processes are most effective and how marketing analytics can be integrated into a marketing strategy to produce better decisions. Thus, there is still a need for greater knowledge about the organizational mechanisms and capabilities that can help the company use this resource more effectively (Herhausen et al. 2020; Vollrath and Villegas 2022).

In this way, this work seeks to go further in understanding how the use of information can lead to better marketing decisions for the company, made strategically, and placing the absorptive capacity at the center of the process, as a mediator. In other words, the main objective of this study is to analyze the mediating effect of absorptive capacity usage (AC) in the relationship between the use of digital marketing analytics (DMA) and effective marketing decisions (EMD) in Brazilian retail.

The AC explains how organizations learn by absorbing external knowledge and integrating it with the company's internal knowledge (Cohen and Levinthal 1990). In this study, the information obtained externally is based on digital marketing data, related to consumer behavior, competitors, stakeholders, the market, and the organization itself, which we called digital marketing analytics (Cao et al. 2015; Kovala et al. 2017).

Nonetheless, even if the knowledge generated by the data is absorbed by the organization, there is no guarantee that assertive decisions will be made (Sharma et al. 2014). Thus, this research sought to analyze not only the use of data in decision-making but also how they are converted into value for the organization, through what we conceptualize as effective marketing decisions, which are those that are in accordance with the company's marketing objectives (Cao et al. 2015).

With this, we seek to understand how the direct impact of DMA on CA would improve EMD and also how the mediating effect of CA impacts the relationship between DMA and EMD. For this understanding, we carried out a survey with 144 Brazilian retail companies and then analyzed it using PLS-SEM. As a result, this study contributes to organizations and the marketing literature by analyzing how organizations strategically use digital marketing data, in which the entire company can benefit from the learning gained from digital marketing information, with effective decisions that generate organizational value.

This is important to bring more clarity about the benefits of using digital information for retail companies, as this sector is often characterized as highly competitive and dynamic in terms of technological advances and innovation. In addition, it can bring insights to improve the development of Brazilian companies and place them at the same level of development as their foreign competitors (Salume et al. 2021).

The study also contributes to the literature in the comprehension of the role of a dynamic capability in the current competitive and global scenario, when dealing with digital marketing information. This is important because, in addition to being a new science that lacks greater understanding, marketing has lost space for theoretical developments in the technology sector (Kovala et al. 2017).

It is clear that digital technologies and the consequent massive production of data have already modified the organizational and marketing environment. Whether this presents a challenge or an opportunity depends on how the organization strategically approaches the issue (Quinton et al. 2018). The best use of digital information can bring competitive advantages by producing insights on how to solve consumers' problems and by connecting them with what they are looking for (Caliskan et al. 2021).

Literature review

Resource-advantage (R-A) theory

The essence of the R-A Theory is competition, which consists of a continuous and unbalancing process in the constant struggle between firms for comparative advantages, in which the proper use of resources will yield a position of competitive advantage in the market (Hunt 2015; Varadarajan 2015). According to R-A Theory, as companies cannot all be superior at the same time, they must focus on constantly renewing their competencies to anticipate market needs and produce the foreseen market offers, understanding the dynamics of the environment. So, success is achieved as the company manages to escape perfect competition (Hunt and Madhavaram 2020).

Due to the dynamic nature of the environment, an important market-based resource is consumer information. In the current digital market, companies can generate insights about consumers, integrate them with their expertise, and transform them into customized solutions, whether in terms of product, price, communication, etc. (Hunt 2013; Varadarajan 2015). However, both the possession of such information and the use that the company makes of them are heterogeneous. This causes firms to constantly change their positions in the market, depending on their momentum, the combination of resources and the skills to manage it (Hunt and Madhavaram 2020; Varadarajan 2020).

Also, there is a lot of information available, so just accessing it is not enough. The company needs to develop skills to select and generate value from the information obtained (Herhausen et al. 2020). Thus, in this paper, heterogeneity in customer information is used as an explanation for part of the heterogeneity in firms’ competitive advantage. That is, all other things being equal, by effectively leveraging its resource advantage in customer information resources to implement strategies that deliver superior value to customers, a company can achieve and sustain positional competitive advantage in the marketplace and, in turn, superior financial performance (Hunt 2015; Varadarajan 2018).

Digital marketing analytics

The term "Big Data" became widely used in 2010, in reference to the huge, unstructured, complex dataset that require advanced and unique technologies to be stored, managed, analyzed, and visualized (Xu et al. 2016; Davenport and Harris 2014). Due to these characteristics and the difficulty in dealing with the transformation of data into useful information by organizations, other terms have also emerged: Business Intelligence (BI), Business Analytics (BA), Digital Analytics (DA), Web Analytics (WA), Data Mining, etc.

Data are presented in different forms and can be obtained from various sources (Sharma et al. 2014). Among other things, they serves for formulating well-informed business decisions. Therefore, it can be said that analytics is a knowledge generator for organizations and, if used well, could bring positive performance impacts for firms (Järvinen 2016). Table 1 provides a summary of the main concepts related to data analytics.

Table 1 Summary of terms related to data analysis
Full size table

In this paper, we use the digital marketing analytics (DMA) concept, which refers to the generation of marketing knowledge, especially regarding consumer preferences in the digital environment, to improve marketing activities. In this process, analysts must first focus on getting to know the consumer. Then the data need to be qualified and targeed to the customer with the right offers. After that, the data must record the result of the transaction (Cao et al. 2015).

Thus, the use of marketing analytics presupposes: data aggregation (data extraction from different digital marketing sources and, if necessary, transformation into a readable format for storage), analysis (reports that makes it possible to visualize and understand the data) and data interpretation (data processing to generate business insights) (Wang and Byrd 2017).

Marketers must understand that analytics is an ongoing process throughout the customer lifecycle. More than tracking events, DMA focuses on a predictive approach to future behaviors (LaValle et al.; 2011). Data science is still a recent phenomenon, and there is a lack of greater understanding of its application in organizations (George et al.; 2014). Most firms are still not managing to take full advantage of it and end up just outsourcing this activity, hoping that everything will be resolved (Quinn et al. 2016).

Absorptive capacity usage

Cohen and Levinthal (1990) established the seminal definition of AC, conceptualizing it as the way an organization acquires and transforms external knowledge and integrates it with its internal one, to support business-related decisions. Thus, knowledge is seen as a dynamic process that enables transformations with a positive impact on organizations (Kovala et al. 2017). For the authors, AC had three dimensions: identification of external knowledge, assimilation of it into internal knowledge and exploitation of business opportunities. This definition was then widely studied and little modified in the literature (e.g., Arbussà and Coenders 2007).

Zahra and George (2002) offered a reconceptualization of AC, attributing one more dimension to the construct. For them, four capabilities influence the firm's ability to create and implement knowledge. (i) acquisition: the company's ability to find external knowledge that is crucial for its operations; (ii) assimilation: the processes and routines that allow the company to analyze and interpret the acquired knowledge; (iii) transformation: the company's ability to develop practices that integrate and combine the acquired knowledge, and the organization's internal knowledge; and (iv) exploitation: the routines that allow the company to refine, leverage or produce certain competencies as a consequence of the acquisition and transformation of knowledge.

According to a bibliometric analysis of AC carried out by Apriliyanti and Alon (2017), in the field of strategic marketing, there is a lack of research that relates it to the processes of creation and integration of new knowledge in the organization, coming from different sources, like digital ones. Furthermore, Abbady et al. (2019) already reported that there is a relationship between dynamic capabilities and the effectiveness of decisions, but the authors reinforced the need to understand what these capabilities are and suggested research developments that consider AC in this scenario.

Effective marketing decisions

In the past, it was possible to get the data, analyze it, and create a model that could be implemented. Today, a continuous approach to analysis and implementation is needed, as everything changes all the time (Davenport and Harris 2014).

In this sense, decision-making based on information gathering to be effective must enable managers to promptly achieve their objectives. In this work, we analyzed marketing decisions, that is, those related to knowledge acquisition inside and outside the organization to improve products and processes involving consumers, stakeholders and the market (Kovala et al. 2017; Dean and Sharfman 1996). Cao et al. (2015) define effective marketing decisions as those that are made in real-time, responsive to change, feature a deep understanding of consumer preferences, and are more effective than those made by their competitors.

Following this logic, having trustful information that reflects the exact scenario of the decision is essential for decision-makers to have a more accurate perception of the situational conditions that they are subject to (Ji and Dimitratos 2013; Abbady et al. 2019).

Decisions constitute the core of marketing studies, and the quality of managers’ judgments is the most determining factor for a successful strategy (Wierenga 2011). In this sense, there is a lack of studies that place managerial decision-makers at the center (Jocumsen 2004; Peter et al. 2020), and most research does not consider the interaction between decision-makers and data analysts. Therefore, deepening the understanding of the decision-making process is relevant for the best use of data and to support decisions that produce better results (Kowalczyk 2017).

Hypotheses development

Digital marketing analytics and effective marketing decisions

According to the Resource-Advantage Theory, in today's dynamic and hypercompetitive environment, companies must constantly renew their strategies. In this sense, information is one of the most important resources, and those who can use it more assertively and quickly can gain competitive advantages (Hunt and Madhavaram 2020; Varadarajan 2020).

Many companies have already been using data for improved marketing decision-making. This can result in both internal benefits, such as increased business efficiency, and external benefits, such as the creation of new products or services, ultimately leading to a competitive advantage (Cao et al. 2015). For example, some manufacturers are using DMA to offer preventive product repair, before failures or customer operational disruption (Brown et al. 2011).

One of the biggest benefits is that decisions based on the use of digital marketing analytics can be controlled: companies can use experiments to test hypotheses and analyze results before actions are taken. This process helps distinguish correlation from causation, reduces the variability of results, and generates more assertive decisionss according to the organization's goals (Brown et al. 2011). LaValle et al. (2011) found that top-performing companies make decisions based on data analytics more than twice as often as the lowest performers. Furthermore, the use of data for effective marketing decision-making is believed to be a cumulative process in which the prior knowledge contributes to the integration of new knowledge (Apriliyanti and Alon 2017).

Organizations that utilize data are not only able to improve internal business efficiency, but can also gain insights to create new products and achieve a better performance (Cao et al. 2015). Therefore, our first hypothesis is based on the premise that data-driven companies are more assertive in their marketing decisions and less reliant on executive personal insights:

H1

The use of digital marketing analytics has a positive impact on effective marketing decisions.

Digital marketing analytics and absorptive capacity usage

The benefits created by DMA depend on understanding how to translate the knowledge generated into a competitive advantage. In this sense, the use of absorptive capacity can play an important role in value creation (Wang and Byrd 2017). Moreover, acquiring knowledge means broadening and deepening the understanding of various decision circumstances, such as the customer, competitors and the market, anticipating future changes, and ensuring that the worst alternatives are discarded (Fink et al. 2017).

Reliable information and learning from the use of data are necessary for future projections and, according to the AC literature, firms must share the acquired knowledge and integrate it with other internal processes (Day 1994). For Işik et al. (2013), this capability is linked closely to the successful use of DMA. Organizations are still using data from various sources, formats, and mostly through systems that do not communicate with each other. A dynamic strategic approach is crucial to managing DMA performance and ensuring reliable results.

Therefore, it is believed that the more reliable information an organization possesses, the better its ability to acquire, assimilate, transform, and exploit such information. Thus, our second hypothesis attests that:

H2

There is a positive impact in the use of digital marketing analytics on absorptive capacity usage.

Absorptive capacity usage and effective marketing decisions

Following the R-A Theory, companies that can use the information to their advantage, are more likely to stand out if their decisions are more assertive and faster than their competitors (Varadarajan 2020). The process is simple: if the firm can learn more about what its consumers think, its competitors, and the market in general, it will have a better chance to adapt and innovate (Cao et al. 2015).

This is important because marketing is part of a constantly changing context in which the  complexity of the market has only increased. Issues like increasing global competition, changes in consumer behavior, the growing concern for environmental and social issues are making decision makers’ work more difficult. As a result, they need more reliable and real-time information to support their decisions (Shukla 2008). In this sense, we present our third hypothesis:

H3

There is a positive influence of absorptive capacity usage on effective marketing decisions.

Absorptive capacity usage mediation: the use of digital marketing analytics and effective marketing decisions

It is already known that data quantity is not the answer to more effective marketing decisions. Firms should focus on how data are employed and on their quality (Järvinen 2016). Thus, the learning process must include the managers’ ability to ask the right questions at the right time and absorb the answers in processes with other team members to act decisively into a proactive strategic mindset (Day 1994). Structured knowledge assimilation and transformation processes can convert tacit knowledge into organizational knowledge (Apriliyanti and Alon, 2017).

The constant use of digital data generates a set of capabilities in the firm that enables it to transform marketing data into insights and marketing value (Wang and Byrd 2017). Therefore, our fourth argument is that a high level of AC retains knowledge, integrating it into the existing one, present in the organization's routines. This integration provided by the AC enhances and transforms the use of data into effective marketing decisions, leading companies to respond more quickly and assertively to environmental challenges.

H4

The relationship between the use of digital marketing analytics toward effective marketing decisions is mediated by absorptive capacity usage.

Research design

To verify the relationships proposed by the hypotheses above, we conducted a survey. Thus, we designed a questionnaire that used the Flatten et al. (2011) scale to measure the four AC dimensions, which is a first- order construct, meaning that it is a reflective second-order construct, and its dimensions are reflective as well. To measure the use of DMA, we adapted the three-dimensional scale developed by Wang and Byrd (2017) that considers the use, compilation and analysis of consumer data, such as those coming from social media, websites, etc. It is also represented in the model as a second-order reflective-formative type construct. Finally, to measure EMD, we adapted a 3-dimensional scale used by Abbady et al. (2019), which has a reflective first-order structure that considers speed, achievement of organizational goals and customer increased knowledge. The constructs were measured with a 7-point Likert scale.

Sample

The sample consisted of Brazilian retail companies that use digital information. To ensure that respondents were eligible to participate in the survey, two exclusionary questions were asked: (i) whether the respondent was responsible for the company's marketing; (ii) if the company was used to collect consumer data online, such as through websites or social media. Approximately 940 contacts were made and we obtained 281 responses, which is equivalent to a 29.89% response rate.

We excluded from the sample firms that did not identify themselves as retailers and/or that had only one employee. We also remove outliers based on the Mahalanobis Distance criterion, which measures the position of each variable compared to the set of variables. Observations that deviate from the majority are considered atypical and may bias the study (Hair et al. 2016). Therefore, we ended up with 144 responses. Table 2 indicates some characteristics of the interviewed companies. As most firms in Brazil are micro and small (SEBRAE 2014), the sample seems to reflect the country's reality.

Table 2 Sample characteristics
Full size table

Validity and reliability

The questionnaire was analyzed by experts and academics. A pre-test was conducted with an initial sample of 30 respondents, just to assess the quality of the instrument. These techniques ensured that there was no common method bias, which was confirmed through Harman's Single Factor test, which indicates whether the relationship between the constructs has an alternative explanation due to any problem with the measurement instrument. It points out whether there is a predominance of a single factor to explain all the variables (Hair et al. 2016).

Then, the internal consistency analysis was performed through Crombach's Alpha and Composite Reliability. As indicated in Table 3, all constructs had values above 0.7 for both tests and, therefore, the scales could be considered reliable (Hair et al. 2016).

Table 3 Summary of statistical procedures
Full size table

Convergent validity and discriminant validity were also assessed. The former was verified with outer loadings above 0.7 and AVE values greater than 0.50. For the latter, we used the heterotrait-monotrait ratio of correlations (HTMT), which is the most conservative criterion for recognizing whether a construct is, in fact, unique. Therefore, the value of HTMT must be less than 0.90, as was the case of this work, which did not present problems of discriminant validity (Henseler et al. 2015).

Finally, before testing the hypotheses, collinearity was analyzed to check whether the independent variables were highly correlated. For this, the VIF and TOL values were analyzed and both were within the recommended limits: TOL greater than 0.20 and VIF less than 5 (Ho 2006).

Results

The hypothesis test was performed through structural equation modeling based on Partial Least Square (PLS) estimation model. This technique is most indicated when the data are not normal, as is the case of this study (Hair et al. 2014). As recommended by Hair et al. (2016), 5,000 subsamples were used through the bootstrapping technique to find the t-value. For these, values above 1.96 with 95% confidence were accepted, allowing the null hypothesis to be rejected. Figure 1 show the results. All paths have t-values greater than 1.96 and even greater than 2.57, at a significance level of 1% for the first 3 proposed paths and, therefore, hypotheses 1, 2, and 3 were accepted.

Fig. 1
figure 1

Source The authors (2023)

Hypothesis test—significance of paths and structural model.

Full size image

The first hypothesis, regarding the influence of DMA´s use on EMD, not only showed a positive and significant impact among the variables but also presented a strong effect. Among all the aspects that contribute  the most to DMA´s use, we identified as most relevant: (a) the identification of ideas or business trends, and the formulation of reports; and, (b) to recognize opportunities that contribute to the improvement of marketing activities, including consumer records for future analysis. These two aspects are even more relevant than social media data analysis and presenting the data in a visual way, such as through dashboards.

H2, in addition to being significant (t = 10.101) with a 99% confidence interval, pointed to a strong effect, indicated by a factor loading of 0.633. This means that the greater the grouping, analysis, and interpretation of marketing data from the digital environment, the greater the company's ability to acquire, assimilate, transform, and exploit knowledge.

H3 also presents a significant effect (t = 3.163), indicating that the AC positively influences EMD. This was the weakest effect found (33.4%), and the confidence interval was the only one that registered a value different from 0.000 (p < 0.002), but still within the parameters sought (p ≤ 0.05). Of the observable variables, the one that contributed most to the EDM construct was decision speed. This finding is especially relevant in a period of crisis, caused by the coronavirus pandemic, where most companies had to adapt their marketing strategies.

To verify the mediating effect of AC, Baron and Kenny's (1986) criteria were used, which consider a series of structural equation models. The first regression was performed between the independent variable (DMA) and the dependent variable (EDM), without the insertion of the mediating variable (total effect). The results indicated that there is significance in the relationship at a 99% confidence interval (t = 7.817; p = 0.000) and the regression coefficient was 0.549. Next, another regression was made with the independent variable DMA, and the mediating variable AC (path a). The results were also significant for a 99% confidence interval (t = 10.685/p = 0.000), and the regression coefficient was 0.668.

Finally, a third regression was performed including the two independent variables and the dependent variable. The coefficients of the direct path between DMA and EDM were analyzed considering the mediating effect and also the coefficient of AC concerning the dependent variable EDM (path b). The results indicated a regression coefficient of 0.233 for the direct path (DMA and CA) and 0.473 for the relationship between CA and DEM. Both relationships were also significant for a 95% confidence interval (t = 2.714/p = 0.007; t = 5.512/p = , 000—, respectively).

After performing the tests, the total value of the path 'a' and 'b' (indirect effect) was calculated. For this, the values ​​were multiplied (0.669 × 0.473 = 0.315). This value represents how much of the relationship between DMA and EDM is mediated by the CA. Thus, it is understood that, after the insertion of the mediating effect, the regression coefficient dropped from 0.549 to 0.315. This value indicates that there is a mediation, but since it is different from zero, such mediation is partial rather than complete. From that, it is also possible to calculate the proportion of the mediation, or Variance Accounted For (VAF), dividing the direct effect by the total effect (0.232/0.549 = 0.422). The mediated effect is given by 1–0.422, which is equivalent to 57%. This indicates that the CA mediation explains approximately 57% of the relationship between DMA and EDM. According to Hair et al. (2000), the minimum acceptable for a partial mediation is 20%.

In other words, it is more likely that the use of marketing data will result in effective marketing decisions if the organization knows how to acquire, assimilate, transform and exploit the knowledge that comes from the digital environment. Thus, almost 60% of the DMA effect on EMD is not given directly, but through the development of absorptive capacity. However, because it is partial, although the effect is potent, it is not a necessary condition for a more effective marketing decision.

The results of the hypothesis test are presented in Table 4:

Table 4 Hypothesis tests
Full size table

Subsequently, the R2 was calculated. It can be concluded that 40% of the company's absorptive capacity usage is attributed to the use of digital marketing data, and 44% of EMD is explained by the two predecessors: DMA and CA. In any case, the size of the effect of AC on EDM (f2 = 17%) was greater than the effect between DMA and EDM (f2 = 12%). This means that, while alone, DMA explains 12% of EDMs, CA explains 17%.

The search for information from the company's operating sector (acquisition), the collaboration between areas (assimilation), the connection of external knowledge with the internal one for the generation of new ideas (transformation), and the revision and adaptation of technologies that facilitate the processing of new information acquired (exploration) are among the observable variables that contributed most to each dimension and make up the factors that can most influence the development of AC.

Discussion

Based on the R-A theory, this study places information as an important resource for obtaining competitive advantage and aimed to explore the effects of using AC as a way of obtaining value from such information, understanding which organizational mechanisms can facilitate the learning process, to generate more effective marketing decisions, since previous studies have failed to explore the use of digital marketing data strategically.

More specifically, the results indicate that the AC plays a critical role in effective decisions in the marketing area. EMD is facilitated by the incorporation and the internal dissemination of external information, particularly related to consumers, the market, or the organization's network. The stronger AC effect on EDM indicates that it is more relevant for a company to know how to search, disseminate and translate information than just collect and analyze data in an isolated marketing department. This finding corroborates Cao et al. (2015), LaValle et al. (2011) and Wang and Byrd (2017), who emphasize the importance of learning from data. Besides, gathering information and disseminating it throughout the organization, our results show that it is also important to ensure that this process becomes an organizational routine.

Aside from that, H1 confirms Brown et al. (2011) and Elbanna and Child’s (2007) insights that the simple acquisition of reliable information (not based on inferences) is sufficient enough for the company to make more assertive decisions. This result places the use of marketing data at the center of the strategic process, generating a need for more interaction between managers and data analysts, especially in larger companies, that have more functional and divisional areas.

The positive influence of DMA on AC indicates that the more reliable information the company can obtain, the greater its ability to learn. This reinforces the cumulative characteristic of the AC, concerning the benefits of data processing and interpretation (Cohen and Levinthal 1990). Hence, if DMA tools indicate, for example, the lack of a product in stock, or a sudden increase in sales of an item due to the impact of some external influence, such as the coronavirus pandemic, they can quickly analyze the situation, and incorporate knowledge to obtain an immediate response, such as searching for new suppliers or the acquisition of similar products. Thus, when learning is incorporated into the routines, the next time information is absorbed, the company will probably be able to come up with more assertive actions (Cao et al. 2015).

The third hypothesis, when relating AC to EDM, highlights the use of new external information in the company's strategic process, to understand and make effective decisions in a dynamic scenario, where fast decisions are necessary. Therefore, firms that are unable to learn and decide quickly, especially in a complex and changing environment, will be at disadvantage, in accordance with Day (1994).

Finally, the fourth hypothesis, about the mediating effect of absorptive capacity usage, has a major contribution to the understanding of how the use of analytics can generate value (Ciampi et al. 2021; Wang et al. 2020). It is more interesting and effective for companies to invest in learning from data than in just obtaining and aggregating them. Therefore, more important than obtaining knowledge is knowing how to learn.

Theoretical and practical contributions

Although there is a wealth of marketing literature available that covers some aspects of digital marketing analytics, including basic definitions (Davenport and Harris 2014; Hauser 2007) and specific applications (Dykha et al. 2021), there has been relatively little written about the integration of analytics with marketing strategy (Suasana et al. 2022). However, interest in marketing analytics has grown rapidly since the advent of the internet and the exponential growth of available data, prompting marketers to grapple with the question of how to effectively use all these information (Vollrath and Villegas 2022).

Furthermore, as posited by Kovala et al. (2017), the vast majority of studies do not incorporate DMA into marketing strategy, only in some isolated tactical actions, and much has been studied about data analysis in the IT discipline. Therefore, this study advances the understanding of how digital marketing data enables marketing managers of retailing companies to make decisions that meet their objectives, through the absorption of new knowledge provided by DMA. Thus, a major contribution of this study is to place marketing data analysis as a strategic managerial activity. The findings of this research also corroborate with the literature and reinforce the need to improve data investigation, so that it can be transformed into decisions that add value to the organization (Cao et al. 2015; Kovala, 2017; Sharma et al. 2014).

Previous studies have already pointed out that many companies were failing to manage the large volume of information obtained, and the benefit of using DMA was unclear (Ghasemaghaei and Calic 2019; Järvinen and Karjaluoto 2015). In this sense, this research results indicated that the use of digital marketing data can indeed generate benefits for companies by elevating marketing decisions. According to our research, this effect is intensified through AC.

In fact, our study also provides a secondary contribution to the absorptive capacity literature. Wang et al. (2020) pointed out that using data with other analytical capabilities is the best path to more satisfied consumers. However, the authors did not specifically determine which capabilities these would be. Similarly, Abbady et al. (2019) report that dynamic capabilities contribute to effective decisions based on data analysis. But the authors also did not indicate what dynamic capabilities. The bibliometric analysis carried out by Apriliyanti and Alon (2017) also indicated that in the field of strategic marketing, there is a lack of research that relates it to the processes of creation and integration of new knowledge in the organization, from different sources, such as digital ones. This work contributes to these issues by placing the absorptive capacity as a mediator of the relationship between DMA and EMD.

This study also corroborates with others (e.g., Brown et al. 2011 and Kane et al. 2015), who report that the increase in technology is not enough to improve decisions. It is necessary to ensure that the organization knows how to learn. In particular, in line with Wang's et al. (2019) findings, our study confirms the importance of analyzing and better understanding data to make them more fruitful. For this, a central aspect is that the company incorporates data interpretation tools, such as visual dashboards to produce essential information and knowledge, historical reports, executive summaries, and time series comparisons, so that data can be viewed and customized in various formats to enable informed decisions.

The practices that can most help managers in this process are as follows: encouraging employees to seek sources of information in the company's sector of activity (acquisition of information), encouraging collaboration between areas to solve problems (assimilation of knowledge), connecting external knowledge to internal knowledge and adapting existing technologies to facilitate the learning process (knowledge transformation), and the review and adaptation of technologies that facilitate the processing of new acquired information (exploration).

Regarding the collaboration between areas, the importance of disseminating knowledge in the organization is an aspect of attention for organizations that see outsourcing as a solution to deal with it (Quinn et al. 2016). This is because information should not be restricted to just a few people or areas in the company, especially since effective decision-making involves multiple individuals (Sharma et al. 2014). Thus, if knowledge transfer processes are not implemented, it is possible that the company ends up wasting resources by hiring a third party. In the sample of this work, most companies had one or more than one advertising agency.

For Kowalczyk, (2017), the main reason that the data potential is untapped is that it does not reach the decision-makers. This study indicates the need for systems that integrate the different areas of the organization, so that they do not end up in organizational silos (Sharma et al. 2014). Managers reported seeing improvements in the quality of decisions that used data analytics strategies. But in practice, there is still room for better data interpretation, report formulation, and statistics to help behavior prediction.

Thus, to be effective, the implementation of DMA requires organizational change. The integration of the knowledge generated by DMA, which must be part of the organization's routines, is important for the success of strategies that encompass marketing data. Retailing, for its proximity to consumers and due to the acceleration of digital strategies in the past few years, can benefit from our findings. Activities such as social media monitoring or web analytics enable more clarity about which products are in short supply, which consumer or market trends are being registered, competitor behavior, etc.

In short, corroborating with Kovala et al. (2017), the ability to make information from data result in effective marketing decisions depends on the company being able to overcome some organizational obstacles, which could prevent the integration of this new marketing knowledge. In this case, it is necessary to assess whether the company's structures, processes, and culture are aligned with the use of this new knowledge.

Managers need to understand that just reacting to consumer desires or market changes is not the best strategy. Most companies reported having taken some action to survive the coronavirus crisis, for example. However, just surviving shouldn't be the answer. It is necessary to be ahead, looking at changes in the environment and seeking to take advantage of them, developing new ways of thinking to reach new positions in the market. Not recording learnings (such as those from the pandemic) for use in future could be a mistake. Our study reinforces the need for quick decision-making in the pursuit of competitive advantages.

Limitations and recommendations for future research

This study has some points of attention. In the first place, the field of study was Brazilian retail stores of various industries and sizes.  When implementing solutions using DMA, different industries may have varying needs. Furthermore, although it is a quantitative analysis, it does not represent the voice of all companies. However, it does guide a behavioral organizational tendency (Shukla 2008).

2020 also presented an adverse scenario, where most companies had to adapt. More than half of the respondents reported that they changed their sales and marketing strategy, with the adaptation of sales to the online environment, so the context may have also influenced the responses. According to Dean and Sharfman (1996), the environmental factor influences decisions, and unfortunately, this could not be controlled. A third limitation is that, while decision-making is a good starting point, it is not possible to be certain of the implementation and success of a decision (Sharma et al. 2014).

Finally, although no bias was evidenced in the statistical procedures, we cannot be sure that the problem did not occur, because the responses were collected from only one professional in each organization (Podsakoff et al. 2003).

This research has highlighted the importance of AC as a way in which companies can better use data potential. However, what practices can be most effective? The use of dashboards or the formulation of an information bank? Future research may seek to better understand these nuances. This might include qualitative research, which can provide clues about the best tools for knowledge absorption in organizations since the options are quite diverse.

Besides that, as stated before, research about other factors that facilitate, mediate, or moderate the proposed path may be welcomed. For example, human resources (the analytical skills of decision-makers in aggregating and interpreting data), specific organizational factors (culture, other dynamic capabilities, separation by size or sector of firms), environmental factors (market turbulence or competitive intensity), etc. Furthermore, although there are studies that relate EMD to performance, future studies benefit from understating better how data are related to this variable.

Since the research was conducted in a scenario of many changes due to the coronavirus, it may be interesting to apply the research in another period to assess digital analytics evolvement. In this regard, a survey with a longitudinal design could show how companies' behavior changes over time (Shukla 2008). For example: were companies that had to reinvent themselves during the coronavirus able to continue innovating by using the potential offered by data?

Conclusion

This paper aimed to verify the influence of the use of DMA on EMD mediated by AC, to increase the understanding of how data usage can be absorbed by the organization to generate better strategic decisions. In this sense, we identified a partial mediation in retailing with active digital marketing practices. Our findings indicate that, although this is not a necessary condition, companies that develop absorptive capacity can improve their marketing decision-making, and suggest new insights for managers and scholars, placing the marketing discipline at the center of the organizational strategy and highlighting the importance of learning and integrating information from the digital environment. Thus, through greater efforts to develop AC, the investments made in marketing data can be justified, and the company can gain long-term competitive advantages.