Connecting Digital Channels to Consumers’ Purchase Decision-Making Process in Online Stores

: This research establishes the relationship between the digital channels that organizations use to communicate with their audience and the stages of the consumer buying decision process in online stores. Researchers have not treated this relationship in much detail and little-known empirical research has focused on exploring relationships between the two subjects. Establishing this relationship is of crucial importance for organizations and consumers, as it ensures organizations use the digital channels that consumers want. A literature review of digital channels and consumer behavior models was performed, which allowed us to deﬁne which are the digital channels and to identify the different models of consumer behavior appropriate for the digital age. A quantitative methodology was used, supported on a questionnaire that allowed us to ﬁnd out which digital channels are the most appropriate for each stage of the buying decision process. The results show that consumers use more than one digital channel at each stage of the buying decision process and for each stage, a set of digital channels is identiﬁable that is most preferred. In light of the above, those who are responsible for deﬁning the digital marketing strategy know what types of content they should produce for each digital channel, allowing them to guarantee efﬁciency in the use of resources while ensuring that consumers get what they want.


Introduction
Technology has not only provided new digital channels that allow organizations to communicate differently with consumers; technology has changed consumer behavior itself.Consumers today have more information, can access that information instantly, and have more purchasing options.One consequence of the above is that a transformation has taken place whereby the power held by organizations, by possession of information, has shifted to consumers, who now enjoy information without limitations.Besides the impact of technology, it should also be noted that communication channels had undergone several prominent transformations, but none more so than the introduction of digital channels.Recent years have seen a continuous introduction of new digital channels that are used for more efficient communication with consumers.Although they have some superficial similarities, the differences between traditional and digital channels are pronounced and the reality is nevertheless striking, showing an evident decline in traditional communication channels to the detriment of the latter which show a notable preference and a sharp growth.
Truth be told, digital channels and consumer behavior have generated a lot of interest lately.The most important questions about digital channels and consumer behavior are simple.Can we identify consumer preference for digital channels and the stages in the consumer buying decision process, and if so, how is this correspondence drawn?There are relatively few historical studies that relate communication channels and the stages of the consumer buying decision process.But research into the relationship between digital channels and consumer behavior has no history and it has been surprisingly neglected by academics, as the search in the main bibliographic databases results in zero value.Such a fact reveals the opportunity to delve deeper into why there is no research on the topic and gives rise to the research question, "what is the relationship between consumers' preference for digital channels and the stages of the consumer buying decision process"?
The structure of this study is organized as follows.First, the literature review of consumer behavior models, digital channels and digital marketing frameworks is carried out.This is followed by a critical analysis of the literature review, and then a section describing the methodology and research hypotheses.The last part consists in the presentation and discussion of the findings.

Review of the Literature on Consumer Behavior
Consumer behavior has over time been one of the hot topics in marketing, and such importance is reflected in the explicit recognition of the existence of the school of marketing thought of consumer behavior.Seth et al. [1] proposed the classification of marketing thinking into twelve schools, highlighting the difficulty of this process.One of these schools is the school of marketing thought of consumer behavior, in which the authors state that it focused on the consumer, opening a new dimension in which the economic component was complemented with other criteria from areas such as psychology and sociology.Prominent names such as Ernest Dichter, John Howard, George Katona, Engel, Nicosia, and others stand out.Shaw and Jones [2], in their review of the evolution of marketing thought associated with the schools and their authors, also agree on the existence of the consumer behavior school of marketing thought.The school answered questions, such as: consumers, why do they buy?People, how do they think, feel, and act?How is persuasion exercised?The key concepts and theories focused on: Subconscious motivations; Rational and emotional motives; Personality; Attitudes; Culture.Also, Arndt [3] 1985) in his logical empiricist paradigm includes the consumer behavior school of thought.
Friedman [4] supported that conceptual models of consumer choice behavior do not become obsolete or outdated, offering in his work five of the primary models that are still unavoidable references: (1) Andreasen Model (1965) [5]; (2) Nicosia Model (1966) [6]; (3) Howard -Sheth Model (1969) [7,8]; (4) Engel Kollat Blackwell Model (1986) [9]; (5) Bettman Model (1979) [10].The author emphasizes the importance of these models in conceiving a framework that connects the variables in the decision-making process but points out their fundamental weakness as limited domains of coverage.Wolny and Charoensuksai [11] point out that existing consumer decision-making models were developed in the pre-internet era and have remained mostly unquestioned in the digital marketing discourse.Even if the classical models are outdated, their usefulness is undeniable since they serve as the foundation for the most current ones, emphasizing the more prevalent five-step model that served as the concept for Edelman and Singer's model [12].Some authors imply a larger breadth of consumer behavior models by integrating references such as AIDA (Attention, Interest, Desire, Action) [13], the hierarchy of effects [14], and the hierarchy of sequence [15].With the rise of online and digital, the large literature on consumer behavior is expanding and new concepts, such as customer journey, customer experience, and customer engagement, are being developed.
Cheung et al. [16] showed that online consumer behavior has been investigated under the aegis of several theories, in which TRA (theory of reasoned action), TAM (theory acceptance model), and TPB (theory of planned behavior) have been predominant, relegating to a secondary position the Expectation Confirmation Theory [17] and the Diffusion of Innovation Theory [18], which were corroborated by Hwang and Jeong [19], adding also the Social Exchange Theory, the Attribution Theory, and the Balance Theory.There are different theories in the literature about what factors determine consumers' intention and behavior toward technology adoption and acceptance, and a great deal of previous research has focused on these associations.Research on this subject has a long history, beginning with the work of Fishbein and Azjen [20][21][22], followed by several models, the last one with confirmed recognition being designated as UTAUT by Venkatesh et al. [23].Other significant contributions stand out [20][21][22][23][24][25][26][27][28][29][30][31] besides the models, theories and frameworks described previously, such as the Theory of Interpersonal Behavior (TIB) by Triandis [32] Triandis (1977), the MM (motivational model) by Davis et al. [33], IDT (Innovation Diffusion Theory) by Rogers [18], SCT (Social Cognitive Theory) by Bandura and Cervone [34], the Model of PC Utilization by Thompson [35] and the Uses and Gratification Theory (U&G) developed by several researchers.
Problem identification (or recognition): the perceived difference between the situation that would be ideal and the current (actual situation).

2.
Information search: search for information in the external environment that is relevant to potentially solve the problem or activate the knowledge residing in memory.

3.
Evaluation of alternatives: evaluation of alternatives, which compete with each other in solving the problem, weighing convictions about the benefits that result in knowledge to make a choice.

4.
Purchase decision: purchase of the chosen alternative.

5.
Post-purchase behavior: new evaluation of the chosen alternative weighing the performance after using it.
It should be noted however that the stages of the buying decision process are always framed in three major phases: pre-purchase, purchase, and post-purchase.It is also important to know that even though the consumer behavior school of marketing thought produced its most important references between the 1950s and 1970s [1], the study of consumer behavior and decision-making goes back further.Contrary to what is being studied, research on consumer decision-making has a long history.As early as 1908, Lewis [43] suggested AIDA (Attention, Interest, Desire, Action), followed in 1924 by Townsend [44] with the stages awareness, opinion, consideration, preference, and purchase.Simon [45], in 1959, proposed the sequence intelligence, design, and choice.Three of the classic and most recognized models are Nicosia [6], Howard and Sheth [7], and Engel Kollat Blackwell Model [9].For Nicosia [6] the decision-making process includes the stages attitude, information search, act of purchase and feedback, while for Howard and Sheth [7] it includes attention, brand understanding, attitude, intention and purchase, whereas Blackwell et al. [9] proposed problem recognition, search, evaluation of alternatives, choice, outcome, and dissonance or satisfaction.Other similar references are still relevant in the bibliography [46][47][48][49].
The innovative and seminal work of Lecinski [50] pioneered a new approach to examining the decision-making process of consumers and provided a relevant insight with the introduction of the zero moment of the truth-ZMOT.Also, innovative and disruptive was the work of Adams et al. [51], which introduced the concept of micro-moments.It is necessary here to clarify exactly what is meant by zero moment of truth and by micromoments.ZMOT can be defined as the moment in the decision process when the consumer searches for product information prior to purchase.Micro-moments are a different and more complex concept.This shows a demand to be explicit about what exactly is meant by micro-moments.Micro-moments are intent-driven decision-making and preferenceformation moments that take place in the course of the customer experience, and there are four types of micro-moments [51]:

•
I want-to-know: Exploration or investigation, but not yet in the purchasing phase, and seeking helpful information and maybe inspiration.

•
I want-to-go: People are looking for a local company or are thinking about purchasing a product from a local retailer.

•
I want-to-do: These are "how to" moments when consumers need aid getting things done or attempting something new and can occur before or after the acquisition.A well-known study that is often cited in consumer behavior research is that of Edelman and Singer [52], who found that consumer behavior changed with the advent of web technologies.The authors maintain that currently, the decision-making process takes place in five stages-consider, evaluate, buy, enjoy, advocate, bond-and with the possibility that a loyalty loop may occur.The difference is that consumers compress the consideration phase and shorten or remove the evaluation phase altogether.
Any review of consumer behavior is incomplete without references to the works of Lemon and Verhoef [53], in which the authors describe the concept of customer experience, and the works of Verhoef et al. [54], where the authors present the customer engagement construct.Customer experience (CX) is defined as customers' perception of the organization, which is formed by weighting interactions across all touchpoints, people, and technologies."The customer experience construct is holistic in nature and involves cognitive, affective, emotional, social, and physical responses of the customer toward the retailer" [55].The logical implication of claiming that customer experience is based on perception and all touch points is that customer experience is defined by a diversity of contextual observation of the moment, i.e., temporal, and relies on the trip that is taken [56].It should also be noted that decision support models should also be mentioned [57,58].
Finally and in conclusion, it is worth pointing out that there is also a body of relevant and recent references that advocate several sequences for the consumer decision-making process [40,[59][60][61][62][63].

Review of the Literature on Digital Channels
Retailers have shifted their services mostly to digital channels, developing new business models [64].Digital channels have a global reach capability, which makes marketing campaigns reach many customers in a fast and effective way, something that was not possible with traditional channels [65].While the existence of multiple channels is an opportunity, it also translates into a challenge.Consumers browse multiple channels and absorb different types of content from multiple devices.It is therefore important for retailers to thoroughly understand consumer behavior and which channels consumers use the most and thus tailor them to their marketing strategies [64].An appropriate choice of channels can help develop engagement between brands and customers and enable companies to achieve greater financial returns [65].
Digital channels, according to Wagner et al. [66] Wagner et al. (2020), are "digital shopping formats that businesses use to offer online shopping opportunities to consumers".Companies seek to make the experience of browsing a digital channel an emotional experience rather than just a traditional sales process.More than just selling, digital channels seek to interact with the customer throughout the buying process, evoke emotions and states of mind common to both businesses and consumers to establish a long-term relationship [67,68], and seek for customers to become loyal to brands [68].
A personalized marketing campaign generates traffic on digital channels [67,68], helps in selling products and developing long-term relationships with customers.Having personalized recommendations based on the consumer's characteristics and purchase history generates more positive results on parameters such as sales, revenue, and the average order price.Consumers feel that their minds have been read, which generates positive feelings associated with their experiences and enhances the relationship between the company and the consumer [67].Online product ratings and reviews are crucial for brand image and company reputation.Consumers, before seeking to learn more about certain products, tend to read other customers' comments and reviews about the company [65].To achieve maximum potential, digital channels should focus not only on existing customers but also on seeking to attract new customers.Behera et al. [67] consider that companies should take measures that attract and retain new customers, as well as encourage the adoption of behaviors that increase the conversion rate.
With the emergence of mobile devices, digital channels have strengthened the interaction between companies and consumers.Digital media are unlimited and, in most cases, free of charge.They inform about the company's news, describe the services provided, and allow permanent contact between companies and consumers [69].Some authors have been proposing different definitions and typologies for digital channels.The following paragraphs describe the different considerations.Straker and Wrigley [68] divide digital channels into four typologies: (1) functional, in which website and email are included; (2) social, in which social media are included; (3) community, in which forums and blogs are included and (4) corporate, in which online advertising is included.Hallikainen et al. [64] consider the following typologies: (1) functional typology, which contains the website, email, search engines, and online chat; (2) social typology, which encompasses social media, such as social networks and photo and video content sharing pages and (3) the community typology, such as blogs and forums.The classifications made by this author are quite like those of Straker and Wrigley [68], however, online advertisements are not considered in their definitions.Labanauskaité [70] on the other hand, identify the website, email, social media, and search engines through the search engine optimization tool as the channels used by companies to communicate with customers.
Anderl et al. [71] proposed a taxonomy of digital channels in which the first variable characterizes who initiates the contact: the company or the customer.In the case of the customer, a division is made with a new variable, which characterizes if the channel is associated with the brand or is generic.The channels identified are website, brand search, generic search, price comparison, display advertising, retargeting, affiliates, and email.The authors also contribute with a review on the topic.It is also worth mentioning the relevant work of Straker et al. [72] who proposed a taxonomy of digital channels having found 34 digital touchpoints and 4 typologies of digital channels (functional, social, community, and commercial).Also, Bagga and Bhatt [73] showed that digital channels influence consumer behavior, noting, however, that the authors do not establish a taxonomy of them.In another source, Jayaram et al. [74] propose a model in which digital channels are determinants for the interaction with consumers and the execution of marketing activities, not formalizing explicitly a taxonomy of them.As for Key [75], digital channels are email marketing, social media, and SEM (SEO and pay-per-click).
More relevant and practical are the contributions of the works by Duarte [76] Duarte (2018) and Ribeiro [77], which propose that digital channels are classified into website, SEO, digital advertising, email marketing, and social media.The authors established the relationships between the stages of the decision process and the digital channels, indicating which channels are more appropriate for each stage of consumer behavior.
Based on the analyses of different authors, this study will look at what are considered the five major existing digital channels: (1) the website; (2) search engines; (3) email; (4) social media_ and (5) online advertising.

Website
Currently, websites function as a channel for companies to communicate with customers.They enable the provision of links to other company channels, such as social media or other platforms [78], and function as a means of authentication for companies [65].Websites affect consumers' perceptions, help them build their opinions about the company, and influence brand image.Companies that have a good image and reputation arouse consumers' trust in the products or services launched by the company [79].These are platforms from which consumers can purchase or view recommendations about certain products, marketed by a given company.Users can also be registered on these portals and receive recommendations based on their purchase history; however, it is not necessary to register on a page to perform the operations previously mentioned [67].
The evolution of mobile devices has changed some aspects related to the layout of websites.The latter continue to be the most used channels by consumers to research about a particular product or service [80], however, the development of mobile devices [67] has changed the perceptions of this channel by customers, if they browse from a mobile device such as a smartphone.The adaptation and consequent creation of websites, adjusted to these devices are attributes valued by buyers [66].
Several authors have been studying and identifying various features that help improve a website's performance and develop bonds with consumers [79].Providing an accessible structure is key to a better browsing experience [78,[81][82][83].One of the main factors is related to the response time of websites over the period of browsing by consumers.A matter of seconds could be enough to leave consumers frustrated with their shopping experiences.Just one second of difference in wait time could directly impact about 20% when it comes to consumer conversion rates [84].
The quality of information and the use of techniques that make it easier for businesses to interpret content are important to consumers.Sellers are experts regarding products that interest customers, making use of this high level of knowledge [85] to offer useful advice and information in an accurate, credible, truthful, and unbiased manner [82,83].Shoppers may present questions throughout a visit to a website, which causes them to ask questions to support services [82].Quality of service can help consumers reduce their uncertainty and make more informed purchasing decisions [78,80,81,85].Recommending products in real time or providing additional information regarding the best-selling products may influence consumers' purchase decisions on products they were not previously interested in [81].If websites fail to clarify the questions raised by customers, they may leave the page and seek an alternative from a competing company [78,83].

Search Engines
Search engines are channels used by consumers in the search process, to (1) search for information about a product or service, (2) know more about a given company and its suppliers, (3) compare prices, and (4) compare product features.After consumers enter an expression, hyperlinked result lists are presented, which redirect the consumer to the companies' pages [70,86].
A set of techniques, called search engine optimization, is used to try to increase their ranking, generate traffic on their websites, and make consumers visit their pages.Search engine marketing seeks to introduce relevant content into search listings, improve their results, and attempts to link the information that consumers want to find with the content presented by companies, helping the latter to spread the word about their content [86,87].
Unequivocally, the most widely used technique is search engine optimization-SEO [86,88,89].SEO is a structured approach, which seeks to improve the position of a website's pages in organic search results, following the input of certain key expressions by consumers [86,89].When implemented well, it generates an increment of visitors to the website pages.This method targets prospective customers, as consumers searching on search engines, already show interest in the product [86].Sheffield [90] identified nine areas for content improvement by companies: (1) optimization for mobile devices; (2) accessible interpretation, sometimes accompanying the texts with videos and illustrative images; (3) creativity; (4) associating the text with the audience's intention; (5) credibility; (6) content without grammatical errors, which can influence consumer confidence; (7) organization; (8) clickability in a certain part of the text, which guides the user to the source of the content and (9) existence of permanent links.

E-mail Marketing
Email marketing is one of the most important digital channels [65].The effects of sending an email are associated with face-to-face communication [91] and can be defined as "sending of commercial and non-commercial messages to a detailed list of receivers respectively e-mail addresses " [92] (p.342), who belong to a set of target segments [70].It also is one of the most efficient strategies in brand building, consumer relationship development, customer acquisition, and sales promotions [92].It confers several advantages to companies, namely: (1) low cost of audience reach; (2) direct communication with consumers, which encourages action by the latter; (3) less time for marketing campaign development; (4) possibility of a message with personalized content; (5) possibility to test different email approaches and (6) integration with other digital channels [86].This tool also helps to create databases of customers and group them by profiles according to their preferences and buying behaviors [65].
Consumers' intention to open emails can be influenced by many factors.Sharma and Kaur [93] identify the importance of the existing relationship between company and customer, as well as the perceived value associated with the content that is intended for the latter as some of these factors.Factors such as the subject matter of the message and the engagement of consumers with other users on social media can aid the dissemination of the message.In turn, feelings also play an important role in the decision.If recipients associate the message with positive feelings, they may forward it to friends and family, helping to spread its content; conversely, a message that generates negative feelings may cause consumers not to forward it.Customer perceptions can be further influenced by factors such as age, gender, income level, place of residence, and the culture of a given country [93,94].

Social Media
The proliferation of social media has caused companies to strengthen their focus on these channels [81,95], redefining the way companies reach, communicate and interact with consumers [95].Social media also allows consumers to share their experiences regarding products or services.The sharing of content in real-time facilitates the dissemination of information and develops social contact between people.The use of social media is an important part of marketing strategies [96,97].Social media marketing constitutes an efficient communication technique that helps strengthen brand performance and disseminate its communication faster [91,95,97].Companies' incorporation of these channels maximizes interaction, offers accurate information about products, makes personalized product recommendations based on the customer's profile, and suggests on-trend products [97], influencing consumers' purchasing decisions [95].
Social media encompass social networking platforms such as Facebook, Youtube, Instagram, Snapchat, and Twitter, blogs, forums such as Reddit and Tumblr, and company websites [98].
These media have added several advantages to companies, namely reaching large audiences at a low cost.Important to this is the role of electronic word-of-mouth (e-WOM) [99,100].The e-WOM allows consumers to talk about products or services, evaluate them and share their buying experiences, acting as an important source of information.e-WOM is one of the most prevalent factors generating traffic on digital channels.The associated results are sometimes superior to those of paid marketing campaigns.However, consumers are not always comfortable having their content shared with others.There are privacy-related concerns that can negatively affect user interactions on these media [98].To counteract these negative feelings, organizations seek to understand customers' views on products, services, or campaigns; try to improve how the brand is perceived by encouraging user participation on these media; and strive to provide better shopping experiences for their customers [98].Consumers seek these platforms for obtaining services tailored to their needs, getting information from businesses, sharing experiences, and recommending products [94].Providing a positive experience helps increase the perceived value and quality of the brand for customers, which can lead to trigger purchasing behaviors and subsequently, make the consumer loyal to the brand [95,101].
Constant communication by companies is valued by consumers.The perceived usefulness and hedonic value of publications influence consumers' intentions to share the message with friends [79].The provision of promotional offers and content such as trial periods, discounts, and offers are also important in reinforcing the image that customers have of the brand [102].Social media marketing activities and social identification positively influence user perceptions and elicit customer satisfaction, which helps influence customer intention to continue to be participative in these media, as well as improve purchase intention [96].

Digital Advertising
Digital advertising includes all messages issued by an entity that unambiguously identifies who is issuing them or the offer they contain.Digital advertising can be done on search engines, social media, or websites and its cost generally follows three alternatives which are cost per click, cost per thousand views, or cost per acquisition.
In the last two decades, the concept of advertising has evolved.Today, online ads are one of the most frequent communication methods [103], and it is possible to see different types of ads displayed on various pages [104].Retailers are looking to generate traffic on their websites through ads.However, potential consumers are not the only focus.Brand customers are also faced with information displayed to them by the action of companies, whose main goal is to influence the consumer's purchase decision [104].The ads are displayed in different forms, namely image, video, or text [105], in spaces suitable for this purpose.Ads assume various types, such as banners, display ads, pop-up ads, sponsored hyperlinks, sending emails with campaigns to consumers, and sponsored ads [104,106].
Yoldar and Özcan [107] categorize ads into relevant or non-relevant, which are further subdivided into two other categories: selected and non-selected.The relevance of an ad relates to whether the consumer views the ad (relevant) or does not view the ad (not relevant).The division is thus made by the number of views that a particular ad has had; if an ad is recommended to a certain target segment, it is considered selected, while if it is not directed to anyone, it is considered not selected.
The competitiveness and the limited time to display ads limit the efficiency of the messages conveyed, so companies should seek to maximize this efficiency and convert the ad into revenue [108].Each time a page is visited, it contains several ads, which generate a set of sensations in consumers [105].The frequency of exposure to an ad requires consideration.Increased exposure to an ad causes the click-through rate to decrease.Many of the customers, by not showing initial interest, continue to not express interest in the product in question and may feel bored over continued exposure [109].
The efficiency of digital advertising is irrefutable and is currently dominated by the two big behemoths which are Alphabet and Meta.

Review of the Literature on Digital Marketing Frameworks
It remains to be reviewed the digital marketing frameworks that are described hereafter.The works of Duarte [76] Ribeiro [77], and Kannan and Li [110] will be reviewed.
The framework of Kannan and Li [110] highlights the impact that digital technologies have on consumer behavior.According to the author, these technologies have reduced the existing information asymmetries between the company and the customer.Nowadays, anyone can access a company's website to get more information about a product.The growth of digital channels has improved communication between the company and consumers, and channels such as social media, search engines, or email have helped improve the value proposition by companies, causing the latter to win the desired customers and increase the value associated with the consumer experience.The spread of social media has transformed the concept of markets, making them bilateral.This channel allowed customers to expose their opinions in public spaces, which may belong to the company, as in the case of the website, or not, as in the case of blogs or other social networks.The comments made by consumers can now be seen by anyone, anywhere in the world.As a result, companies are now considering customer reviews of products or services and including their recommendations in new product lines.Search engines have made it possible for consumers to acquire free information about products and services as well as brands that fit their search criteria.By entering a certain keyword, the consumer gets a set of search results, consisting of different websites and paid advertisements.Companies thus seek to develop the informational content of their websites, to match as many keywords as possible, and to arouse consumer attention towards their content.

•
Coverage-the percentage of the target audience reached by a message.It allows us to understand which channels are used to contact the public and understand if the message reaches its intended destination; • Frequency-the number of times the consumer, on average, is reached by a message, in a given time interval.It allows us to understand whether consumers are aware of the frequency of contact that digital channels impose; • Acquisition-activities and methods developed by the company, whose purpose is to acquire traffic in its digital channels; • Persuasion-seeks to change attitudes, create taste and preference, convince the potential customer about a particular product or service, and lead him to the act of buying.This construct allows us to evaluate whether the message transmitted in the channels arouses the consumers' interest; • Conversion-achievement of an action by the consumer; • Loyalty-degree of effectiveness or contribution of a company's digital channel in developing and maintaining a long-term relationship with customers.Consumers know, like, and trust the brand, which is reflected in repeated and regular buying behavior.

Literature Review Critique
Three important arguments emerge from the studies discussed so far.First, all the studies reviewed here support the argument that the research of the consumer buying decision process continues to be an area of marketing with recent publications and research has been accompanying the challenges that continually emerge with the advances in technology and the web.The transformation of classical consumer behavior models to online is proof of this statement.
Moving on to the subject of digital channels, we find an antithesis of the previous context.Restricting the scope of the taxonomy of digital channels, one finds that the number of references and studies is limited and their focus somewhat ambiguous.As for the characterization of digital channels, the scenario is more promising, with more and more incisive bibliography.
As for digital marketing models, only two studies were identified that address the relationship between the stages of consumer behavior in online purchase decision-making and the digital channels used by companies.The evidence reviewed here seems to suggest a pertinent opportunity to develop this subject and contribute with new perspectives to the field of knowledge.Consequently, given the absence of an established relationship between the stages of the purchase decision making process in consumer behavior and digital channels, this research will provide this knowledge by identifying which digital channel(s) are most appropriate, according to consumers' perceptions, for each stage of the purchase process.

Materials and Methods
There are several tools available to establish the suitability of digital channels and the stages of consumer behavior in the purchase decision-making process.For this purpose, a questionnaire was developed and followed by data analysis.A quantitative approach was therefore used.Previously and arising from the literature review the research hypotheses are proposed.

Formulation of the Research Hypotheses
This study considers digital channels to be five and includes three models of consumer behavior: (1) Edelman and Singer's model [52]; (2) Micro-moments [51] and (3) consumer conversion process [76,77].The formulation of the research hypotheses is shown in Table 1.Social media are the most suitable digital channels for responding to the micro-moments: (a) "I want to know", (b) "I want to do", (c) "I want to go" and (d) "I want to buy".[75,76,85,90,[94][95][96][97]

Questionnaire
The design of the questionnaires was initially thought of and based on the ideas underlying Servqual [114,115] and Servperf [116][117][118].The aim was to measure the expectations that consumers would have about each channel, assuming what an excellent company would do, and then to gauge the perceptions that the consumer would have about a company.The discussion between advocates of the two alternatives is well known [117,[119][120][121][122] but considering the better long-term results of Servperf and the implication of a questionnaire half the size of the other, the latter has been chosen.Consequently, only perceptions will be measured.
The questionnaire is divided into two sections: (1) questions regarding users' perceptions of a retailer's digital channels and (2) questions regarding respondents' sociodemographic characteristics.respondents.The questions consider previous studies [76,77], as well as the statements of several authors, indicated in the literature review.The questions presented are all closed questions.Likert scales were used in a horizontal 5-point format with response anchors (e.g., strongly disagree/strongly agree), and in which respondents respond with their degree of agreement with the statement presented.
Table 2 contains the list of questions to be asked to measure each construct and stage and their respective scale (also included were socio-demographic questions that allow us to characterize the sample, which due to their relative irrelevance are not in the table).

Stage/Construct
Question Scale

Consideration Coverage
When I receive or am confronted with digital channel X regarding Retailer's technology products, I view its contents.

1-Never 5-Always Consideration Acquisition
When I want to get more information about a technology product, I take into consideration Retailer's digital channel X.

1-Never 5-Always Evaluation
Retailer's digital channel X allows me to evaluate a technological product.
1-Strongly Disagree 5-Strongly Agree Evaluation I want to know I am receptive to considering Retailer's Digital Channel X information when I want to know more about a technology product.

Buy I want to buy
When I want to make an online purchase, Retailer's digital channel X allows me to do so.

1-Strongly disagree
5-Strongly agree Enjoy I want to do it When I want to know how a technological product is used, I consult Retailer's digital channel X.
1-Never 5-Always I am receptive to considering Retailer's Digital Channel X information when I want to know how a particular technology product is used.

Advocate
I use Retailer's digital channel X to make recommendations, suggestions or comments.

1-Never 5-Always
Bond Loyalty I check Retailer's digital channel X in order to maintain a connection with the company.

1-Never 5-Always I Want to Know Coverage
When I want to know something about a certain technological product, I consult Retailer's digital channel X.

Conversion
Retailer's digital channel X presents content on which I can exercise a certain action.

1-Strongly disagree 5-Strongly agree Acquisition
When I receive information from a digital channel X regarding a Retailer's technological product, I consult other channels of the company, namely its website.

1-Never 5-Always
Recommendation I use Retailer's digital channel X to make recommendations, suggestions or comments.

Pre-Test, Sample, and Sampling
Before the administration of the questionnaire, a pre-test was conducted, in which 6 people were surveyed, to validate the questions, correct gaps and errors, and adjust the vocabulary to facilitate understanding, as well as avoid different interpretations by the respondents.
Only when the samples analyzed are representative of the theoretical population under investigation from which they were generated is the statistical inference procedure legitimate.The target population of this study includes adult individuals who are users of the company's digital channels.In this particular study, the target population was limited to the entire Portuguese population that uses the company's digital channels.No other limitations were imposed on the definition of the population and sample.
The sample has members with heterogeneous characteristics, to approximate, within the inherent limitations, the results obtained with the behavior of the entire population.The sample that constitutes this study is a non-probability convenience sample, using the list of subscribers of a higher education institution and the respective social networks.These alternatives ensured the heterogeneity of the responses.A total of 268 people participated, but after the initial validation, which asked if the respondent "had ever accessed the retailer's website, the number of responses was reduced to 234.
Having defined that consumers' perceptions would be assessed, this implies that consumers can express their opinion about perceptions of digital channels only if they are users of those channels.For this purpose, one of the largest retailers of technology products in the Portuguese market was selected as the study company.The data was collected through an online questionnaire distributed to consumers since the objective of the study is their perception of the use of the company's digital channels when searching for information and interacting with it.The questionnaire was online and open for responses in April and May.

Data Analysis
The analysis of the results was done using two different tests: (1) the parametric t-Student test for paired samples and (2) the non-parametric Wilcoxon Signed-Rank Test.The t-Student test compares two populations from which two paired samples of the same people were chosen, based on a unifying criterion of sample elements concerning a quantitative dependent variable.It is assumed that the dependent variable has a normal distribution in both populations, and variance homoscedasticity is not required [123].If the sample does not follow a normal distribution, a non-parametric test can be applied, but for large sample sizes (n > 30) the Central Limit Theorem is applied.For samples in which n is less than 30, as is the case for the tests comparing E-mail and Social Media channels, a non-parametric test is justified.The sign test and the Wilcoxon test are the non-parametric alternatives to the t-test for two paired samples [124].
Since the objective of the investigation is to identify which digital channel is most suitable for each stage, the two tests indicated above allow this objective to be achieved.The two tests compare the average of each channel in the different stages of the purchase decision-making process for each respondent's answer.The channel that has the highest statistical average and if the statistical test is significant, then that channel is preferred by the respondents.
The t-Student test gives the following hypotheses for the two-sided test: H 0 : µ1 = µ2 vs. H 1 : µ1 = µ2, where µ1 and µ2 represent the population means.For the Wilcoxon test one has: H0: E(X) = E(Y) and H1: E(X) E(Y), where "E" denotes the expected value [124].
To evaluate the preference that the digital channel has for a stage in the consumer's purchase decision-making process, a set of calculations were performed, as described below.It should first be noted that the preference of each stage of each model for each digital channel was rated by the consumers.Note that we have three consumer behavior modelsthe first with six stages, the second with four stages, and the third with six stages-and that we have five digital channels.For each stage, the preferred averages of each digital channel were calculated, and a table was constructed containing the list of digital channels ordered by descending value.To find out if the digital channels had different preference values the t-Student or Wilcoxon tests were performed (note that there are channels that are not used by all respondents and therefore have a number less than or equal to 30).Table 3 contains the different stages of the models described previously, containing the digital channels listed in descending order of preference.Note that for a stage a digital channel can have more than one instance, and when this happens, in the second instance it appears in square brackets.This means that all the other channels that are in square brackets constitute a new level of comparison.

Results
Using the results of consumers' perceptions of digital channels for those stages where a preferred channel was obtained, the associated research hypothesis was classified as validated; for stages with more than one preferred digital channel, the hypotheses were classified as partially validated; for channels that are in a lower position, the respective hypotheses were classified as not validated.The Table 4 contains the validation of the research hypotheses.

Discussion and Conclusions
The framework interconnects the digital channels with (1) the consumer buying decision process, (2) the micro-moments and (3) the conversion process.The questionnaire design was inspired by the Servperf model, focusing on the quantification of respondent perceptions concerning the different channels of retailers of electronic products, to understand which ones are more appropriate to the different stages of the purchase decision process.
The results obtained are not entirely conclusive, and in most of the stages of the models considered in this study, there is more than one preferred channel.Given the information contained in Table 4 only the research hypotheses H2.3 (c) and H5.3(b) are not rejected.This means that search engines are considered the most suitable for persuasion, while digital advertising is the most accepted for frequency.All other research hypotheses were not confirmed.It is, however, possible to derive additional information from the table.
Regarding the Edelman and Singer model [52], at the consideration and evaluation stages, respondents equally prefer the digital channels e-mail, search engines, and the website; for the buying stage, the most appropriate channels are the website and e-mail; for the enjoy stage, e-mail and search engines are equally prominent; in the advocate stage, social media, e-mail, and search engines are equally preferred; and for the bond stage, the most appropriate channels are e-mail, social media, digital advertising, and search engines.
As for the micro-moments [51], one finds that for the I want to know and I want to do micro-moments, e-mail and search engines are the most appropriate digital channels; for the I want to go micro-moment, the website and search engines are the most appropriate channels, and for the I want to buy micro-moment, the website and e-mail are the preferred channels.
For the conversion process, in the coverage construct, e-mail and search engines are the preferred channels; for the frequency construct, digital advertising is the most appropriate channel; for the persuasion construct, search engines are the most appropriate channels; for the conversion and acquisition constructs, the most appropriate channels are e-mail, search engines, social media, and website; and for the loyalty construct, the most appropriate channels are e-mail, social media, search engines, and digital advertising.
When compared to the other studies mentioned in the biographical review, the results of this research present relevant differences.In the studies of Duarte [76] and Ribeiro [77], social media and website were the preferred channels for most stages.In the present report, most of the research hypotheses for social networks have been invalidated, while also only a few research hypotheses for the Website have been partially validated.
It is also possible to perform an analysis per digital channel, using Table 4. Starting with the website, considering H1.1, H1.2 and H1.3, it can be seen that this channel is more suitable for the initial stages of the purchase decision making process up to and including the purchase stage.For search engines, and evaluating H2.1, H2.2 and H2.3, it is found to be a suitable digital channel for most stages except the purchase stage.For e-mail, checking H3.1, H3.2 and H3.3, we can see that this digital channel is also suitable for most steps, but respondents pointed out the limits of not wanting frequent messages and messages with the purpose of persuading.As for social media, and attending to H4.1, H4.2 and H4.3, it is found that this digital channel is more suitable for the final stages of the purchase decision making process, showing to be symmetrical to the digital website channel.For digital advertising, gauging H5.1, H5.2 and H5.3, no pattern was found across the three models under analysis and consequently no conclusions can be drawn about it.
This research delves into an area of knowledge that has been little explored.The framework provided has contributed both to the state of the art, i.e., at the scientific level, and to the business level, offering significant guidelines for subsequent studies.In the scientific realm, few studies relate digital channels to the stages of the consumer buying decision process.At a business level, important contributions are offered to brands, which commit a significant percentage of their budget to digital marketing.Selecting the right digital channels and sending the right messages to consumers allows organizations to firstly optimize their resources and secondly meet the expectations that consumers have about the information they want to receive.The implications are relevant.First, it is known that there are two digital channels that can be used for most steps in the purchase decision making process, and these are search engines and email.Next, the website is suitable for the initial stages up to and including the purchase stage, while social media is suitable for the stages from purchase to loyalty.This framework allows organizations to optimize their communication with customers, understand which channels customers favor throughout their journeys, and helps companies optimize the allocation of resources to different channels.This is an important issue for future research.Given the benefits that this framework brings to organizations and the divergence of results between the different studies performed so far, further research should be conducted to investigate the results described in this paper, which can reconcile the diverging conclusions or bring new knowledge to this subject, and for this purpose future studies on the current subject are recommended.

Table 1 .
Formulation of the research hypotheses.
Search engines are the most suitable digital channels for the micro-moments: (a) "I want to know", (b) "I want to do";(c) "I want to go"; (d) "I want to buy".

Table 3 .
Stages of consumer behavior and digital channel preferences.