Fast Moving Consumer Goods (FMCG) are a fundamental part of our daily life, often without even realizing how relevant they are, above all: food products, especially packaged ones.
As typical of such dynamic industries, market data (Euromonitor, 2021) suggest that constant growth is driven by frequent innovation, not only in the product portfolio, but also in the offer of additional services (e.g. different purchasing channels, sustainability requirements, etc.).
Precisely for this reason, consumer research to understand these behaviors and motivations has become a key element on which companies are intensifying their efforts and resources to offer products and services aligned to the final consumer’s priorities and win in an highly competitive scenario.
This research supports a new point of view in favor of the use of alternative analysis methodologies to conjoint, proposing for the first time the comparison of the famous Conjoint Value Analysis (CVA) and Maximum Difference Scaling (MaxDiff.), an avant-garde technique whose potential theoretical benefits have been empirically tested only in a few occasions and never in mass consumption or in a comparative way with the conjoint.
Research design – To offer this contribution, the experiment was carried out within the Italian chocolate confectionery market.
To test these assumptions, two distinct analyzes were conducted: the first of a qualitative and exploratory nature (7 interviews, 1 focus group, 1 protocol analysis) aimed at drafting the second, consisting of two quantitative surveys carried out through Qualtrics and Lighthouse Studio platforms, administered to a total of 654 responding individuals (with 465 valid interviews), divided into two homogeneous and independent samples to highlight which of the two techniques best represents the cognitive process of the consumer.
Coefficients of importance – In particular, for the CVA sample the most important attributes were found to be: brand awareness (26.61%), sustainability certifications (26.51%), contained ingredients (18.51%). On the other hand, for the MaxDiff sample: Content ingredients (20.06%), Brand awareness (19.27%), Price (15.68%).
Table 1 - Coefficients of importance deriving from CVA and MaxDiff.
It can be inferred a general tendency to prefer well-known brands and characteristics related to the spheres of health and quality that an offer profile possesses (and makes visible); for example, a reduced calorie content or sustainability certifications, which the literature demonstrates to be implicitly associated with better products (Magnier, Schoormans and Mugge, 2016).
Despite of that, it has to be wondered what is the cognitive explanation that led the two techniques to generate such accentuated differentials in the evaluation of the Price and Sustainability attributes, also to identify a strategic solution to mitigate this effect in the research phase and propose interventions related to the discipline of marketing capable of responding to the pragmatic manifestation of these consumer needs.
This difference can be traced back to the fact that there is a so-called 'unobserved variability' found in numerous empirical researches in which it is shown to be unlikely that the variances of importance between the components (or attributes in this case) do not depend on factors contextual to the experiment, such as the place, the time period or, as in this case, the task and stimulus administered.
For example, if the pack color was a relevant characteristic, but the colors that individuals think when asked how important that is were not known, the distribution of importance would be conditioned by the distribution of ambiguity/uncertainty about colors, used as an implicit reference by the interviewees (as it happens in this specific study within the MaxDiff. tasks with respect to the 'Sustainability' or 'Price' characteristic).
In the author's opinion, this phenomenon can be easily mitigated by preparing a specific presentation of each item of the MaxDiff. task (as it would be with a level of an attribute in the CVA) to take advantage of all of MaxDiff.’s benefits, without affecting its reliability.
Managerial implications – It can be said that on the total sample both techniques are adequate to capture what is defined as of greater importance at a macroeconomic level by the experts, that is the quality of the products and the brand awareness, thus clearly suggesting a strategy to optimize the investment allocation and product design.
At managerial level, it is therefore recommended to simplify and concentrate the distribution of marketing investments. In fact, in both questionnaires about 40% of the total importance was attributed to the reputation of the brand and the formulation of the ingredients.
In this perspective, the observed trends recommend the use of two extremely effective communication levers in spreading these characteristics:
- Advertising campaigns: compared to the importance attributed to brand awareness, maintaining a strong degree of brand awareness is crucial to carve out a significant market share in this sector.
- The packaging: beyond its aesthetic appeal, the packaging is essential to report the information of greatest interest to the consumer. As proof of that, in the presence of detailed information of this kind provided in the classic conjoint exercise, the monetary value of the product has been overshadowed. The resulting managerial suggestion is therefore to increase and enhance the investments made in these areas.
On the contrary, the price and promotion levers lose considerably their importance when explicated, since the perception of savings is rather limited and consequently has little impact on the management of the consumer's economic availability.
It is therefore assumed that in the presence of a concept showing both elements, consumers give priority to quality over price.
Variance – However, although both techniques have been able to identify preferences consistent with the main market trends, it must be specified that the literature agrees in defining socio-demographic characteristics as moderating factors in consumer preferences.
Consequently, for this industry it is advisable to prefer methodologies that best capture the variance of respondents' response, so that to be able to prepare more precise socio-demographic segmentation or split procedures, where each group is made up of individuals who have great internal homogeneity and high external rejection.
The comparison table between the standard deviations (Table 2) supports a net superiority of the MaxDiff. against CVA in this instance, generating a greater variance in six out of seven attributes analyzed, confirming what was hypothesized and allowing to more clearly discriminate the preferences expressed by the subjects.
Table 2 - Variability of the responses produced by the two methods (MaxDiff. - CVA): analysis carried out on the sample total.
In addition, for segmentation purposes, the author argues that the MaxDiff. is more reliable than CVA for two further reasons: first, the CVA algorithm is able to extract only artificially the individual utility coefficients, which are therefore estimated and not declared by the respondents.
Secondly, the recommended threshold of seven attributes at a time in a CVA exercise risks to become a critical and highly limiting problem to produce satisfactory segmentation even in relatively simple industries such as chocolate.
Indeed, if we then of more complex sectors, tracing a segmentation to such a small number of attributes risks nullifying the usefulness of the analysis and the transferability of the insight.
User Experience – In addition, this thesis also investigates the preferences of the participants with respect to the interview method, noting a statistically significant preference for the MaxDiff. exercise in all four dimensions tested - involvement, realism, complexity, level of attention - which favors a higher quality of obtainable information.
Table 3 - Evaluation of interviewees in the interview experience: comparison between samples.
Given this interpretation and the results obtained, this study argues as the MaxDiff. is more suitable for addressing product design problems in the chocolate confectionery sector, as it allows interviewees to be presented with a greater number of attributes to be evaluated, generates greater variance between the answers provided (resulting preferable for segmentation processes), is clearly preferred in terms of involvement, realism, (less) complexity and attention, also incurring a lower rate of abandonment by the interviewees.
By virtue of this research design, the incremental knowledge generated in support of the discipline can be summarized in the following points, which lead to a preference for MaxDiff. with respect to CVA:
i. Reliability of results
ii. Variability of responses
iii. Number of testable attributes
iv. Positive perception of the survey
v. Lower dropout rate
Table 4 - Duration (in seconds) and dropout rate of the two questionnaires.
Comments