Constructing the composition of the focus groups

The main objective of the focus group interviews conducted in this study was to gain information about customers’ beliefs and insights into designing a customer-friendly elec-tronic insurance servicescape. For this purpose, three focus group interview sessions were organized in March 2006. Each group consisted of six participants. In order to ensure that the saturation point in collecting the empirical data is reached, an additional focus group session containing five participants was organized and conducted in February 2007.

No generally acknowledged rule for the amount of groups has been established. Vaughn et al. (1996, 49) suggests that focus group interviews should be conducted until partici-pants’ responses are predictable by moderator. However, it has been proposed that in most cases this phase emerges between two and four focus group interviews (e.g. McQuarrie and McIntyre 1987; Lyons 1991).

Hence, I decided to first conduct three focus group interviews, and then perform the data analysis. Further, I planned that after analyzing the data of the three conducted focus group interviews it is time to estimate whether or not additional focus group interview sessions were necessary. Due to the correspondence in the discussions, the fourth focus group interview, organized to ensure the saturation, confirmed that the amount of the interviews was sufficient to qualitatively describe and explain the investigated phenomenon.

In constructing the focus groups the principles Technology Readiness (TR) model and customer classifications (e.g. Parasuraman and Colby 2001; Colby 2002) discussed in chapter 5.3 were used as an instructive guideline. As to the customer classifications, the general principle in sampling the respondents to the focus groups was that they should be considered as representing one of the four customer segments possessing a somewhat posi-tive attitude towards electronic services (i.e. explorers, pioneers, skeptics, and paranoids;

laggards were excluded).

In addition, two specifying key principles were followed while constructing the focus groups. First of all, similar backgrounds, especially socioeconomic and educational factors, were considered as important determinants in constructing the groups. Demographic fac-tors, such as age and gender, were not considered as such critical determinants that they would have jeopardized the functionality of the group, although people, especially of dif-ferent ages, most likely have difdif-ferent attitudes towards computers and electronic services.

In fact, it was rather desired to include participants of different ages and genders to the same group in order to have various kinds of different opinions and discussion among group members. Instead, education and professional background were assumed to affect participants’ world of thoughts more. The above mentioned assumption was based on the principles of TR model. Therefore, people who had similar educational background (e.g. university versus comprehensive school or vocational school) and similar professional background (e.g. managers and specialists versus workers or officers) were put into the same group.

Second, one thing that was common to all the participants of the focus groups was related, in addition to their attitudes towards electronic services, to their skills and motiva-tion to use electronic services and a computer in general. Again, referring to the TR model as well as customer classification by Parasuraman and Colby (2001), and Consumer Readi-ness (CR) model by Bitner et al. (2002), all the participants should possess sufficient com-puter skills, and motivation to use a comcom-puter and electronic services. Thus, certain types of customers who can be classified as laggards were appropriately excluded from the study since it is not even of insurance companies’ interest to try to get all the customers engaged in the electronic service environment. Taking the perspective of research methods, the technique by which the groups were constructed in this study can be determined as pur-posive sampling often used in this type of research setting. It refers to the “procedure in

which the participants for the focus group interview are selected by researcher’s predeter-mined insights into participants’ ability to contribute to the study” (Vaughn et al. 1996, 58–59).

As to the composition of the four focus groups, each of them possessed similar char-acteristics to some extent (e.g. age distribution or residence), but still the nature of each group was clearly distinct from another. These four groups were named as: 1) Insurance science students; 2) Academics and managers/experts; 3) Non-academics and workers/of-ficers; and 4) Idiosyncratic group. The list of the interview participants can be found in appendix 5.

All the participants were living in Tampere or its neighbor towns (Pirkkala, Nokia).

The decision to select all the participants from the same region was mainly made due to practical reasons. In addition, regional factors were not considered to have such a crucial in-fluence on research results that it would have been essential to also consider regional factors as one of the key determinants of the group composition.

Students were selected among the people who are studying business studies and have insurance science as their major subject at the University of Tampere. In my opinion, this group provides an interesting surface to reflect the responses of other groups on since the students of insurance science are already familiar with insurance business and services, and therefore, they can also be considered as experts of insurance business. Further, since the students need computer and the Internet in their studies everyday they can be consid-ered to have particularly good skills and motivation to use a computer and electronic ser-vices. Moreover, as Lin et al. (2005, 691) point out using students as targets for a research is particularly useful in electronic service environment since students will eventually be-come the most active Internet users and influential consumers in the marketplace. Hence, they could be considered as potential customers. Understanding the needs and preferences of potential customers is naturally important. As the literature points out good friends should not be included in the same focus group (Koskinen et al. 2005, 127). At this point it has to be noted that, the members of this group knew each other at least on some level, since they were all studying the same major subject at the same University. Despite of knowing each other, the discussion, however, stayed at a professional level, and thus, I am confident that it did not decrease the quality of the results.

Three female and male members between the ages of 27 and 59 constituted the second group (Academics and managers/experts). One of them did not have an academic educa-tion but instead she was an entrepreneur and owner of a company. Thus, she was perceived to fulfill the criteria to be included in the group. None of the group members knew each other beforehand.

Third group (Non-academics and workers/officers) consisted of four male and two female members between the ages of 26 and 60. One of the group members was still a

student at a polytechnic but since he was not going to have an academic degree he was perceived to be suitable for the group. Two of the group members worked at the same place and knew each other superficially but they did not have a closer relationship. Hence, it was not perceived to be harmful to the research execution.

The fourth group was named as idiosyncratic group since it can be considered to some extent as a mixture of the other three groups. The idiosyncratic group contained five participants of which three were male and two were female. The age distribution of the participants was not as wide as, for instance, in groups two and three. The ages of the par-ticipants varied from 23 to 32 years. Two of the parpar-ticipants studied at the university. One participant having engineering background in terms of education is working as consulting engineer. One participant, who had graduated from vocational school, was working in the metal industry as a regular “worker”. One participant having academic education was working as assistant group controller.

Getting people to participate in the research was not an easy task. One reason for re-luctance may be the complexity of the topic (electronic insurance services). The first strat-egy to collect group members was to send a letter of invitation for the group of people who fulfilled the above discussed predetermined criteria, and who had responded to the feedback survey, which was attached to “insurance cover evaluator” service concept after its launch in the early 2005 (see chapter 5.3). By responding to the feedback survey the potential research participants would have proved their natural interest towards comput-ers and electronic services, as well as motivation. As good as the idea was, unfortunately it did not work. Altogether 27 (16 to women and 11 to men) letters of invitations were sent.

A movie ticket was promised as a compensation for everyone who would participate. Only four responses came back of which only two responded positively to the invitation.

Accordingly, another strategy had to be followed. It was not as organized, and it was based on “word-of-mouth” marketing. I contacted my friends and acquaintances and asked them to think whether they had a friend(s) and/or acquaintance(s) that would be interested in participating in the focus group interview session. Naturally, I was hanging onto the predetermined criteria for the group members all the time. These friends also helped me by ensuring that the people they recommended fulfilled the predetermined criteria for participants of the focus group interviews. For this reason, trying to recruit participants randomly, for instance at the front door of a grocery store, would not have been a functional strategy for the purposes of this study.

Finally, by following the above mentioned “snow-ball” strategy, the groups were con-structed. At this point it has to be noted that I recognized (or knew at some level) some of the participants. However, I would not consider that this fact would have skewed the results especially since the group interaction was in question. In addition, a research as-sistant was present all the time and was also asking questions from the respondents.