Market Expansion Study to Understand Consumer Responses in 5 Markets

Posted by | Fuld & Company

Background

A manufacturer, marketer, and distributor of well-known consumer and commercial brands across a wide range of products was considering a particular line of scented candle products aimed at Gen Z and Millennials and approached us to conduct a Market Expansion Study. The company had already tested consumer attitudes and receptivity to the collection in the US, UK, Australia, and New Zealand and now wanted to understand attitudes towards the brand idea and the collection in 5 additional markets – France, Italy, Germany, China, and South Korea. The client also wanted to optimize its product line for expansion into each market, so a crucial element of the research involved identifying which named products should be initially offered in each market.

The client had developed visual mood boards for the brand concept, and additional visual stimuli for each of the scented candle items in the collection, and wanted to get feedback on the packaging, name, and descriptions of each of the products.

The client had already created an online survey for the 4 English-speaking markets (US, UK, Aus, and NZ), using a top-3 and bottom-3 ranking of the product descriptions. However, whilst a frequency analysis of the top-3 and bottom-3 choices can provide a ranked list of the product descriptions, the procedure is not entirely reliable because a) respondents are given a long list of 12 complex product descriptions so may end up selecting randomly to minimize the effort required, and b) they may just click the first 3 products in the list order that they like without reading the complete list, a response strategy which decreases the reliability of the results.

Approach

Since the prioritization of the planned products was a critical part of the research, we optimized the base version of the questionnaire. In place of ranking questions, we used a MaxDiff experiment with multiple versions. Each version had 12 tasks with 4 options per task, where the options were the names and descriptions of the products in the target scented candles collection. On each task, respondents were asked to indicate the name and description that they found most appealing and the one they found the least appealing. This MaxDiff was designed using the essential principles of a balanced design.

In addition to the MaxDiff experiment, the survey questionnaire also sought information on:

• Receptivity to the brand idea and the general brand message – using visual and textual stimuli
• Response to the collection using a 5-point semantic differential scale (Dislike à Like, Boring à Exciting, etc.)
• Response to the collection design in terms of desired imagery attributes (association)
• Respondents were asked to select one Appealing and one Unappealing product from the name and text descriptions and indicate the reasons for their choice
• Response to product packaging with name, text, and visual stimuli – for each product, we asked respondents to indicate overall liking, fit to brand concept, name and description fit, and purchase intent, using a 5-point Likert-type agreement scale

The questionnaire also included questions on demographics (age-screener, education, annual household income before tax) and respondents were screened on whether they had made any purchases from a list of home fragrance products in the past 12 months. For China and South Korea, where the client had less prior exposure, we also included open-ended questions on spending on home fragrance products and scented candles (if used) during the past 12 months.

In each country, we collected data from 580 respondents with quotas by age, with approximately half the respondents in each country from the 18 – 24 age group (Gen Z and Millennials). We translated the questionnaires into the respective languages of each country. For France, China, and South Korea, this included translations of the textual stimuli about the brand concept, the product names, and the product descriptions as well. Appropriate care was taken to ensure cross-cultural concordance in the meaning of product names and descriptions, especially where colloquial or affective phrases were used in the original. The data from open-ended verbatim responses were translated back into English by expert translators (native speakers for Korean and Mandarin).

The results of the survey were analyzed using a mix of quantitative and statistical analysis, and quantitative analysis of textual data using the “quanteda” package in R. (“quanteda” stands for “quantitative analysis of textual data” and was developed by Kenneth Benoit of the London School of Economics.) The analysis of numerical data involved cross-tabulations by age, ANOVA, and ranking of preference shares computed from the MaxDiff data using Hierarchical Bayes. The focus was on the target segments – Gen Z and Millennials with other age groups also looked into as potential segments for specific products in a particular market, so at the overall level in each country, splits by age were applied. Further analysis was also conducted to look at consumer responses by income and education (overall and within each age category).

To get a better understanding of consumers’ affective responses to the brand stimuli, we further corroborated numerical responses to Likert-type scales using the results from the textual analysis. This was also carried out for consumer responses to each product in the collection. Word frequencies/clouds, co-occurrences, n-grams, and keywords-in-context analysis were used for the textual data. For understanding responses to specific items in the target collection, the analysis of textual data was vital due to lower variance in responses arising from the use of the original 5-point Likert type scales and also for cultural differences in scale use as discussed below.

Cultural differences in scale use are a well-known issue in cross-cultural consumer research and have been commented on extensively in the academic and practitioner literature. In our study, these differences did not always directly correlate with earlier studies in these markets, and we noticed, for example, that Chinese consumers gave much higher positive scale ratings on all questions, followed by South Korean consumers. The variation in their ratings was also low.

In addition, we also found that French consumers gave lower ratings in general to all products on all measures. The cross-cultural differences were explored further through the following steps:

• Reference to other survey questions on household usage and expenditure on home fragrance
• Background desk research to understand the usage of scented candles in a cultural context
• Any possible “country of origin” effect noticed from the open-ended verbatim responses
• Background research into some of the usage occasions – particularly the way leisure time is spent in the two East Asian markets
• Explain differences in response styles using Hofstede’s famous cultural difference dimensions, most notably the “Collectivist- Individualist” dimension and “Uncertainty Avoidance”
• Ipsatization of scale responses (standardization across respondents) to remove any potential differences arising from a cultural bias in the use of the rating scales

Throughout the analysis of numeric scale- data, we used both the mean values and a boxed-difference score on all measures.

Outcome

The research helped the client to prioritize the products in its collection for launch in each market. It also helped them understand the positive and negative consumer responses to specific fragrance notes and usage occasions that were included in the product descriptions for future tweaking of these descriptions or product names in specific markets.

Based on the results, the project team made recommendations for consideration by the clients’ Marketing Insights team:

• We recommended that the client go to market with 4 product concepts in France, Italy, and Germany and 5 each in China and South Korea
• We suggested specific changes in the product descriptions of some concepts for each of the markets. These changes mostly involved changes in the product descriptions aimed at communicating usage occasions to potential consumers. We also suggested changes in the way usage occasions were described, adapting them to the particular cultural context. There were 6 such products in France, Italy, and Germany; 2 in China, and 2 in South Korea
• There were some products that we recommended holding off until the product line had taken off as the proportion of people likely to be receptive to them in a particular market was predicted to be small

Project Team: Michele Sachar, Yusof Ahmad, Daniela Vojvoda, Abhishek Nigam, Abhishek Kaushik, Rajesh Biswas,  Abhaya Das

Tags: Market Analysis, Market Intelligence, Market Research, Marketing Strategy