The proposition of this study is the first effort to comprehensively concept a dimension for friendliness in the BBX scale and measure how consumers experience eco-friendliness as a part of an overall brand experience, and as such, it has its limitations.
Thus, the extended BBX scale with the eco-friendliness dimension needs to be still replicated, further developed and tested with other data sets.
As this is a new area of research, there was limited possibility to refer to existing research literature which would have helped to create a more precise theoretical framework on which to build the extended BBX scale including the eco-friendliness dimension. Consequently, the theoretical background of this thesis is a mixed combination of theories mainly from consumer behavior research and marketing, as there is still very limited research and theories on measuring eco-friendliness in the context of brands or experiences. On the other hand, also the research of Brakus et al.
(2009) was based on a wide spectrum of research areas starting from consumer and marketing research, and further covering philosophy, cognitive science and applied management. In this study, the items for the eco-friendliness construct were formulated on the basis of the existing BBX scale items which may not be the most optimal for describing eco-friendliness, even though in this study the factor loadings for the items proved to be well above the recommended values for factor loadings.
One of the delimitations that was decided in the beginning of this research was that only global brands would be examined and smaller local brands were consciously out-scoped
from this study. The reason for this is that there was a need to focus on some specific high-tech brands that are very familiar to most consumers due to their wide visibility.
Another delimitation is that the research focuses only on one country specifically that has a high penetration of mobile phones, and it would still be beneficial to confirm that the results apply in some other countries and cultures (Netemeyer et al., 1991). This in turn leads to one of the major limitations of this research which is that a single study with a data set from one country is not enough to support the proposition of the extended BBX with a fifth dimension for eco-friendliness. However, otherwise there are not that many limitations concerning the sample, as the respondents covered all age groups and were evenly distributed in the country. The sample size (N=506) was fairly large, and when the responses for the three brands included in this study were combined the sample size was tripled to over 1500, so the sample size effects are not statistically so relevant as in the case of small samples that have issues with reliability and factor structure.
The "Do not know" option could have been avoided by pre-testing more brands initially and then select only those brands that were familiar to the respondents, which would then have meant that the pre-tested sample and main sample would have had to be from a similar population. However, this was not done because in the scoping of this research it was decided that there would not be a lengthy pretesting phase. Even though the "Do not know" option offers a way to skip the pre-testing phase, the risk is that it may have created some noise in the measurement model. However, this was resolved by treating the “Do not know” answers as missing data. Also, it needs to be considered that if there would not have been the “Do not know” option, respondents could have selected some other option in cases where they did not truly know what to select, which could have in turn created some additional noise.
It also needs to be taken into account, that when a new parameter is added to a model that is tested with a fairly large sample, very often the factor loadings for this parameter are rather high just due to the fact that a new parameter has been added. However, this problem has been tackled in this study by examining different ways how the new parameters incorporated in the model, by also considering the possibility that the new parameter is embedded in the existing parameters of the original scale. With this approach, the results can be considered to be valid and supportive of the fifth new parameter for eco-friendliness in the BBX scale. If a new parameter added to an existing scale fits the data well it can often help to explain the data better, but this is not always necessarily automatically the case.
In addition, it may be considered to be a minor limitation that in the sample the majority of the respondents owned a Nokia/Lumia branded mobile phone which may have had some effect to the means and standard deviations in the responses, even though the factor loadings would not have been necessarily affected. The fact that Nokia is a Finnish brand may have an effect on the way the Finnish respondents have rated the
brand Nokia in this survey, so there clearly is the possibility that there is a so-called home country bias in the results. This phenomenon is due to affective brand processing which refers to the situation where local products are assessed more positively than imported products irrespective of the actual objective quality of the product (Riefler, 2012).
The number of brands included in this study was limited to three brands, from which two were the two current market share leaders (Apple and Samsung) while Nokia has lost market share considerably since then. However, at the time of the data collection for this study, the deal between Nokia and Microsoft had just only been announced, and the Nokia brand still had a rather strong position among consumers' minds, at least in the country where the data was collected. Now that the former Nokia Lumia branded phone (current Microsoft Windows phone) has lost market share considerably, it would be also important to include brands of some other phone brands that have lost market share, such as Blackberry by RIM and Motorola.
This study did not include any data on other brands than high-tech brands and therefore the original BBX model was only replicated for high-tech brands, which may also be considered to be a delimitation. There was a high correlation between the intellectual and sensory variables in all of the three tested models of this study in the case of high-tech brands. This study did not then further investigate the possibility of combining factors in the original BBX model to test if by reducing the number of factors to four, so that the sensory and intellectual factors would be combined, would have resulted in smaller correlations. In the original BBX model of Brakus et al. (2009) the correlation between the intellectual and sensory variables is not notably high. As this is a replication research, the high correlation between two variables is not yet considered to be an alarming issue, but this is an indication that more data would need to be collected and analyzed to do a deeper analysis of the correlations between the intellectual and sensory factors in the models. The highest correlation between the intellectual and sensory variables was in the case of the Nokia brand which may also be due to the home country bias partly that was already mentioned above.
There are also some demographic differences in the results, however, they have not been reported in detail in this thesis except when assessing the criterion validity of the measurement scale. The difference between men and women is statistically the most significant one, and there are some differences in the results according to the age groups of young and mature consumers, but these are not as significant. The educational background of the respondents has no statistical significance in how the respondents rate the eco-friendliness of high-tech brands.