Publications and growth trend

According to the results obtained in the review, Figure 1 shows that the generation of knowledge related to the engagement is growing. Beginning in 1996 (n = 1) and increasing in subsequent years at a constant pace, in 2006 it presents 15 publications and it continues until 2009, which rises to 29, in 2010 it skyrockets to 55 publications, continuing a growing trend that is increasingly marked in the years 2011 (n = 86), 2012 (n = 106), 2013 (n = 154), 2014 (n = 209), 2015 (n = 267), 2016 (337), 2017 (n = 385) and the period that goes on in 2018 (n = 95).

Figure 1 Number of publications about engagement per year. 

Considering what Dereck J. de Solla Price (1956) mentioned that "the growth of scientific information is exponential and occurre at such a rapid rate that every 10-15 years the existing global information doubled (Price's Law)" (^{Ardanuy, 2012 page 14},). Although it is not necessary to mention that each discipline suffers its own evolution, going through different stages. ^{Price (1963)} explai-ned that scientific knowledge begins with precursors, where a small group of scientists begins to make a point in the subject, through exploratory studies and where many aspects of the subject must be studied. Then it passes to exponential growth where a considerable number of researchers are attracted by many aspects to study the subject that still has to be explored and analyzed. Subsequently, there is a consolidation of an academic body and a solid knowledge and finally maturity is given, characterized by the saturation of studies on the subject by covering all aspects, which manifests itself in a significant reduction of publications (^{Dabi et al., 2016}, ^{van Nunen et al, 2017}).

In the case of the subject about engagement it is easy to notice the established tendency by Prince, in the first years (1996 to 2005) there are few studies on this subject and with totally different approaches as well as the denomination, the used term and the conceptualization are different.

Authors and cooperation

Of the 1791 publications written since 1996, 1874 have been written by 159 different authors and 7 do not have an identified author. The largest proportion of authors (55.97%, n = 89/159) is credited with three publications on the subject. 20.75% (n = 33/159) have four publications, a 10.06% (n = 16/159) with five publications, 33.89% (n = 20/159) between 6 and 10 publications, and a single author representing the 1.13% (n = 1/159) with 18 publications in this regard.

Table 1 shows the list of the most productive authors regarding the study topic. The classification is based on the total number of publications of the author and not on the order of authorship. Hollebeek is the most productive author on marketing-oriented engagement with 18 publications, followed by Graffigna (n = 10), Malthouse (n = 9) and Cheung (n = 9).

Similarly, a very different range can be observed regarding the average number of citations per author (ranging from 189.5 to 0.16) and the number of publications as the first author (for example, Hollebeek has 10 publications as first author, Hamilton has 6 as the first author, while Rahma does not have any publication as the first author of the 8 publications he has made, coinciding with Wei, Suberamanian and Zanuddin, who are in the same case).

The average number of authors per publication was 0.088 (n = 159/1791). The average number of publications per author was 11.26 (n = 1791/159). 15.74% (n = 282/1791) of the publications has been written by a single author, 29.03% (n = 520/1791) of the publications has two authors, 27.58% (n = 494/1791) has three authors, and 27.64% (n = 495/1791) has four or more authors (with a maximum of 42 authors). It can be concluded that there is a lot of joint and collaborative research within and between institutions, since publications in co-authorship represent more than three quarters of the total.

^{Wang et al., (2014)} mentions that a large number of publications with several authors indicates a close relationship between themselves within the same research topic and a greater opportunity for future joint work.

The pattern of cooperation (ie: co-authorship) between the authors was analyzed in VOSvie-wer. The criteria used for this analysis was that the authors to include at least 3 publications of the subject and at least one citation. Authors who are in a network means that they have published together and those ones who are not in a network or connected means they have not published together. The size of the circles represents the cooperation between themselves, the colors represent the groups or clusters of collaboration. The results are shown in Figure 2. It is observed that 75 working clusters were formed, the cluster number 1 is made up of 16 authors, that appears in a red color (the names of Tang, Tucker and Wong stand out because they have more publications of the group), the second cluster is integrated by 10 elements shown in green color (the name of Kumar appears, as he has the most published documents of the group). In cluster three, highlights Holle-beek and Malthouse, it is composed by 8 elements and is shown in blue.

Figure 2 Authors cooperation network on marketing-oriented engagement. 

It can also be concluded that 34 authors are not integrated into collaborative groups, which are the circles that appear around separately; 22 groups are binas (two authors as a team), 10 groups of three people, two of 4 people, three of 5 authors and one of seven.

It must be mentioned that in terms of authorship, a possible bias must be taken into account. Authors with the same name cannot be distinguished from each other, and authors who use different names in their publications (for example, changing surnames by marriage or other reasons) could not be merged. As recommended by other authors such as Chiu and Ho (2007) or ^{van Nunen, Li, Reniers and Ponnet (2017)} it is important that researchers use a unique digital identity (such as ORCID) to solve these problems.

Journals that have published about marketing-oriented engagement

In total, the 1791 documents have been published in 160 different sources including magazines, congresses, books, among others. This high number indicates a wide variety of research approaches on engagement in marketing.

The magazine that has published most on the subject is Computers in Human Behavior with 32 documents, followed by Journal of Product and Brand Management (n = 21) and Journal of Services Marketing (n = 19).

Likewise, the citation indexes are presented (with the highest score is Journal of Interactive Marketing), the Scimago Journal Rank (SJR), which calculates the impact factor based on the information included in the Scopus database. Elsevier company (being the highest Journal of the Academic of Marketing Science); the Source Normalized Impact per Paper index (the highest one again) and the study area in which the journal is located.

Geographic and institutional distribution and joint cooperation

As it appears in the data of Scopus, each publication was assigned to a country or territory and to an institution according to the direction or affiliation of the author.

Publications by territory

87 territories or countries were identified, 12 are from Africa, 33 from Europe, 2 from Oceania, 10 from America and 30 from Asia. An inequality is observed in the geographical distribution regarding the generation of knowledge on the study topic, which is reflected in the number of publications from developed countries and those in emerging economies. Regarding the number of publications, 37.93% of the countries have published one or two documents on marketing-oriented engagement, 13.79% between three and five publications, 16.09% between six and ten publications, 32.18% more than 11 publications. The United States is the leader in quantity (n = 626), followed by the United Kingdom (n = 240) and Australia (n = 222), as shown in Figure 3.

Figure 3 Top-15 of most productive countries or territories publishing on marketing-oriented engagement research. 

The network of cooperation or co-authorship between territories was carried out in VOS-viewer. For its realization, countries with at least 3 publications and at least one citation were considered. The result yields 12 working groups. The size of the circles represents the amount of documents, the larger is which has more publications on the subject; the thickness of the links represent the strength of collaboration between territories and the color groups or clusters of work. The result is presented in Figure 4. Two main groups can be distinguished, one is led by the United States and the other one by England, although other smaller ones are also shown, such as Australia, Canada, Germany, China and Spain, among others.

Figure 4 Cooperation network between countries and territories in marketing-oriented engagement. 

Publications by affiliation

160 institutions were identified considering that an author may be affiliated with more than one institution, or a publication may be written by several authors from different institutions. Of all the institutions, more than 95% corresponds to universities, so it can be said that knowledge on the subject has been developed from the academic world. 25.62% of the institutions have published five documents, 54.37% between six and ten, and 20% more than eleven publications. Table 3 shows the most productive institutions, with Australia being the country with the most participation in this aspect. The three main ones are: the University of Auckland in New Zealand (n = 22), followed by City University of Hong Kong (n = 21) in China and Queensland University of Technology QUT (n = 19), in Australia.

Analysis of citation

The analysis of citations provides the number of times that the publications on a certain topic have been cited by other publications listed in the database used, in this case Scopus. In total, the 1791 publications have been used 23271 as a reference in other publications. The average citation per publication is 12.99 (n = 23271/1791).

Of the total of publications analyzed, 37.40% (n = 640/1791) have no citations. 40.87% (n = 732/1791) has between one and ten citations, 16.47% have received between 11 and 50 citations, 2.62% between 51 and 100 citations, and also 2.62% have received more than 100 citations.

It should be noted that there is a general assumption that the number of citations reflects the influence and notoriety of a publication, and hence its quality (Smith, 2007, Ugolini et al., 2015). However, some authors (for example, Walter et al., 2003; Chiu and Ho, 2007) state that the number of times a publication has been cited by others does not really indicate the quality of that publication, but it measures its visibility. Also, there is an increasing recognition that publications of open access journals are increasingly being cited (Whipple et al., 2013).

The top-20 citations per publication can be observed in Table 4. The most cited article is The social influence of brand community: Evidence from European clubs in Algesheimer, Dholakia and Herrmann in 2007 with 770 citations and an annual average of citations of 64.17. However, if the average number of citations per year is analyzed, it will be discovered that there are other articles that, even if they do not outnumber the citations received to that one of ^{Algesheimer et al. (2005)}, they surpass it in average annual citations. Such is the case of the article Consumer engagement in a virtual brand community: An exploratory analysis by ^{Brodie, Ilic, Juric and Hollebeek in 2011} that has received 421 citations but evidently its annual average of citations is 105.25 higher than the most cited article, as well as the one done by ^{Brodie, Hollebeek, Juric and Ilic (2011)} with an annual average of 77.5 and the article by van Doorn, Lemon, Mittal, Nass, Pick, Pirner and ^{Verhoef (2010)} with 75.29.

Concluding that it is noticeable that the more years a publication has, the more citations it will have, however, when reviewing the impact through the average number of citations per year, it is observed that recent publications may have an important relevance in the field, such is the case of the publications mentioned above.

Citing and co-citation analysis

As mentioned by Li and Hale (2015), the co-citation analysis focuses on the relationship or interaction between two publications, and offers a general description of the publications that have been cited in common within other publications. So it is assumed that the more publications are cited together, the more similarities can be found between them (^{van Nunen et al, 2017}).

The analysis of shared citations was made in VOSviewer. It was done by specifying those references shared more than 25 times, resulting in 32 items divided into four clusters, as shown in Figure 5.

Figure 5 Co-citation analysis of highly-cited references used in publications on marketing-oriented engagement. 

The size of the circles means the number of citations, that is, the larger a circle, that publication has been used more often as a reference in the articles of engagement. The distance between the circles means a stronger relationship and greater similarity between them. The color suggests a similar theme among these publications.

The map shows how the references of the engagement publications are grouped and are illustrated in four clearly identified groups where each one represents a research field in the engagement, without ceasing to interrelate.

Based on the examination of the titles of all the individual publications of each group, it could be concluded that the blue group is made up of the articles referring to the conceptualization, exploration, evolution and analysis of the term engagement, which are being mixed with the rest of the groups. The yellow group includes the documents related to the co-creation of value, service-dominand logic and the relationship with loyalty. The blue group is made up of publications related to consumer engagement in online environments creating value through a brand community, as well as some documents that focus on the structural equation model tool, as a technique that predominates in the quantitative analysis of studies empirical on the subject. And the green group focuses on engagement in social media, much related to the previous group.

Terms analysis

Finally, the trend analysis of the terms that were used as a filter was carried out. The classification was made manually of each one of the articles that contained in its title, summary or key words some of the terms related to engagement, it was separated, mentioning that some were included in several concepts because they had both. Similarly, the revision of the trend in the number of publications was measured by considering the number of articles of the year 2015 to the current date in relation with the total.

Firstly, as was observed in the trend of publications on the topic that is increasing, where 60.5% are from January 2015 to the cutoff date of March 2018, it is to be expected that if they are separated by subtopics or by the different conceptualizations, as it has been given the engagement in marketing, the result is the same.

Table 5 shows that effectively from 2015 to date the trend is growing in each one of the conceptualizations of engagement, currently focusing on online engagement, social media engagement and community engagement.

And it also identifies a field that is being born and brings with it great challenges and areas of opportunity for research and the professional field that is mobile app engagment and app engagement, since they do not represent even 1% (n = 7/1791) of all publications, and this amount is from 2014 to date.