In addition to the expert interviews, online observations were conducted. The subsections to follow will present the processes related to conducting and analyzing these online observations.
5.4.1 Online Observation Planning
In order to confirm and fill out some of the knowledge gained through the literature review, we conducted blind peer review inspired online observation on ten social media influencers.
The influencers were chosen based on their accounts’ focus on fashion and their popularity among users (number of followers). An account’s focus in fashion was considered as an important criterion due to the focus of the research. Accounts not focusing on fashion would not represent the influence of fashion influencers on social media properly, and hence fashion influencers were seen as the right target for the observations. The popularity of the influencer’s account in terms of followers was seen as an important criterion as the number of followers naturally reflects upon the influencer’s reach – thus, how many people are seeing the influencer’s content and possibly affected by it (Convince & Convert, 2020).
Also, the popularity in the number of followers tends to have significant influence upon the influencer’s collaboration with brands – thus, if an influencer possesses a lot of followers on his/her account, it is likely that the number of collaborations in terms of paid content, ads and other types of marketing is more significant (Backaler, 2018, pp.92-93).
In addition to the popularity and focus of the accounts to be observed, the influencers’
nationality was set as a criterion. In order to decrease the cultural differences, and to obtain as comparable data as possible, we focused on Scandinavian fashion influencers. The
selection process of the influencers to be observed began from finding the most popular account in each of the Scandinavian countries. Based on this criterion, and the previously mentioned ones, the following influencers were selected: Kenza Zouiten (@kenzas) from Sweden, Mathilde Gøhler (@mathildegoehler) from Denmark, and Celine Aagard
(@celineaagaard_) from Norway. As the purpose was to find a total of 10 influencers, we then chose the remaining seven influencers purely based on the number of followers – thus, not having the same number of influencers per country. This said, four of the remaining influencers are from Sweden (Angelica Blick, Janni D, Victoria Tornegren, and Alexandra Bring), while three of them are from Denmark (Josefine HJ, Pernille Theisbæk, and Emili Sindlev). (See Appendix 10, pp.219.)
5.4.2 Observation Process
Concerning the content that was taken into consideration in the online observations, Instagram offers a few different types of content, as previously mentioned in the section about Social Media. When it comes to fashion related content, fashion influencers use all of the possible types of content available on Instagram: posted content (pictures and videos), stories, highlights, and IG-TV videos. For the purpose of this study, all content is relevant, as it includes a lot of fashion related content. However, as the goal with the online
observations was to confirm or deny the hypothesized phenomenon of posting outfits only once, it was appropriate to look into the type(s) of content that would offer a useful point of comparison. In order to assess whether or not, and to what extent, fashion related content encourages users towards a certain type of behavior, it was necessary to be able to compare content not only between different social media influencers, but also to compare the
content of the same influencer in a longitudinal way.
For the longitudinal comparison, the posted content (pictures and videos), the month of May was chosen as the focus point, as it was considered a somewhat neutral month, unlike December for instance, which is a month that would most likely have a lot of Christmas related content and thus be extremely influenced by paid collaborations to the extent that it would be out of the normal. Also, the summer high season, June-August was considered as being exceptional, as content within this time period tends to focus on vacation related content, which also does not picture the nature of influencers’ ordinary content. This said,
from the months that were left, the month of May was chosen randomly as the focus month. In addition to that, the years 2017, 2018 and 2019 were selected as the years within which the month of May would be observed, as the intention was to get the latest possible insights from Instagram.
During the selected time period, observations were made concerning the following factors:
total number of posts, number of fashion posts, number of posts including the same outfit, number of posts including items re-used, number of posts including product tags, and number of posts using filters. The purpose with looking into the total number of posts during the selected time period was to compare the obtained number with the one of the fashion posts. This was done in order to find out how much the fashion related content actually covers when it comes to the Instagram feed of the fashion influencers. Once the fashion content was separated from the non-fashion related content, it was observed in more detail in terms of the number of the same outfits in order to either confirm or deny the existence of the OOTD trend, thus whether or not an outfit can only be posted once on social media.
Also, it was considered important to not only find out the relevancy of the OOTD trend, but also to investigate the extent to which it extends. By this we refer to the question whether the entire outfit should be different in each post or can some parts of the outfits be re-used in several posts. This was observed by looking into the number of posts that include a re-used item – a handbag, shoes, or a jacket, for instance. In addition to these, the number of product tags in the fashion content, in the form of direct links to product pages or tags mentioning the brand or the product, was observed to understand the role of fashion companies in terms of fashion influencers' Instagram content. Finally, as filters are considered relevant and important based on the findings of the literature review, the number of posts including a filter was observed. This insight was considered as possibly useful for the purpose of combining it with the other primary data findings, as filters are a feature supported by the technology of AR and could be used for other purposes than tone and color fixing.
5.4.3 Analyzing
As the online observations were conducted in a blind peer review inspired manner, the analysis of the observations begun with the comparison of the obtained results. In other words, the results obtained by each researcher were compared in order to see whether there was any difference in the findings, and their possible significance. In practice, the figures of the observed six factors from the month of May in 2017, 2018 and 2019 were held up against each other, which resulted in the finding of a very small difference in the figures.
This difference was found in the number of posts using a filter, which can be argued as being a challenging estimation to make from the researcher’s point of view, as filters used to modify tones and colors on pictures can be difficult to observe. Nevertheless, out of the 923 fashion posts, a total of 5 percent of the posts were not assessed as using a filter by both researchers. After the comparison of the findings, they were put into a table to visualize the results and the development that has occurred during the observed years.
6 Findings & Analysis of Expert Interviews
In this chapter the findings derived from the primary data collection done by expert interviews will be presented and analyzed in the context of the chosen theoretical
framework TAM. Thus, this first part of the analysis focuses solely on the findings related to the themes related to TAM, while findings related to the key concepts will be discussed in the next chapter.