Return to site

Primary Market Research Part 2

Recap of Webinar 6 by Michal Gilon-Yanai

· Product Definition,Market Research,Quantitative

Marketing Professors Duncan Simester, MIT Sloan, discusses the importance of customer behavioral research and Olivier Toubia, Columbia University reviews quantitative methods, specifically conjoint Analysis.

Our May webinar on Primary Market Research (PMR) focused on qualitative methods and featured both representatives from academia and practitioners. Our speakers provided a rich set of frameworks and tools that can be used to effectively conduct interviews, observations, lead user innovation and other interactions with potential users in the initial stages of product design (view here). A discussion of primary market research cannot however be complete without quantitative methods, and conjoint-analysis in particular. This was our main area of focus on July 8 2015. But before diving into it, we had the pleasure of hosting Prof. Duncan Simester, Marketing professor at MIT Sloan, who highlighted an often overlooked area of research. He also explained the value of obtaining behavioral rather than perceptual data from target customers at a very early stage.

In a presentation titled “Why great products fail”, Duncan explained that many products fail not because they fail to deliver value. In fact, he says, many companies today “get that right”. They do speak to their customers in depth about their needs, and they do go on to design products that create real value.

However, customers’ purchasing processes are often too little understood, and therefore, while great products are being created, customers are often unaware of them, and companies do not succeed in helping them find out and learn about their products. No matter how great, they fail.

Keeping this in mind when we conduct primary market research can have significant impact. According to Duncan, there are two main ways in which customers make purchasing decisions:

  1. They search for information
  2. They make inferences based on cues

When thinking about search, there are several points to note. First, there is a question of the cost vs. the benefit of the search. Different products have different corresponding cost/benefit ratios, depending on the importance of the decision and how difficult it is to find information. For example, if you’re an airline trying to decide which aircraft to purchase, the search process is long and expensive whereas if you’re getting into a taxi at an airport in NYC, you don’t search at all despite knowing that some taxis are safer than others to ride in.

The relationship between customers’ prior expertise and how much they search is interesting. Customers with high expertise don’t search much, since they believe that they already know what they will find. Customers with very little expertise, on the other hand, also hardly search since they typically don’t know where or which questions to ask. This is important when launching a new product and planning the communication around it. Another point to note and address is that at times, the perceived benefit of searching for a new product or service is less than the actual benefit, especially in markets with no recent innovation.

When searching is impossible, or the information collected incomplete, customers form inferences using what can be observed to make assumptions about what is too difficult or expensive to search for. Price and brand are the two most common cues to inferring quality. The role of the brand depends on the level of expertise of the customer, and may also vary across product features. Features that can be found on a spec sheet can usually be found through search, decreasing the reliance on brand. However, with less tangible features such as reliability, brand plays a more central role.

In the second part, Duncan provided a couple of examples to show how behavioral rather than perceptual data (what people do vs. what people think they will do) with regard to a new product can be obtained at a very early stage. Citing one example of a startup developing a unique new material, he described the high interest shown by large companies in the new technology. In order to narrow down the list of potential venture partners to a short list of real ones, the startup required an upfront $50K investment in the joint venture. This strategy allowed them to identify the more serious potential partners, and to begin working with a few of them early on.

Our next speaker was Olivier Toubia, Marketing professor at Columbia University. Olivier teaches a course called “Customer-Centric Innovation”, in which students carry out both qualitative and quantitative primary market research. He started out by outlining the main differences between these two approaches. For example, while qualitative research requires time-consuming, face to face interaction, quantitative research can be done online, with large samples that are easily scalable through services like Mechanical Turk. The type of information collected is different of course. While qualitative research is very helpful for exploring and uncovering user needs, quantitative methods are best suited for measuring customer preferences and willingness to pay.

The quantitative tool used in Olivier’s course is Conjoint Analysis, one of the most commonly used marketing research methods. This method is useful for quantifying how much consumers value specific features of a product or service. The method Divides a product into attributes, and assigns attributes different levels. Then, a survey is designed to quantify how much each level of each attribute is worth to consumers. The data is then analyzed using statistical methods, primarily regression. To help students and entrepreneurs conduct this analysis, Olivier has co-developed online tutorials with Prof. Elie Ofek of HBS (Conjoint Analysis: Online Tutorial ; Conjoint Analysis: A Do it Yourself Guide).

In answer to questions about the number of attributes that can be included in a survey, the type of products studied and whether the method applies both to incremental and disruptive innovation, Olivier emphasized the importance of including only attributes which customers are able to visualize and evaluate. If you include features that are too abstract or completely new, this type of survey is not ideal. Furthermore, it is important to consider cognitive load, which can also make the survey less effective. For this reason, it is recommended to test no more than 10 attributes. When deciding on features to include in the survey, you should think strategically – skipping the obvious features that you will surely include in the product, and including only those that you believe will make the difference – in other words, the areas in which you will innovate. The focus should be on things that matter to consumers, and that will help you make better decisions in the product design process. Olivier reviewed a case which he developed named VerTerra, based on a new venture launched by one of his past students. The student, who develops eco-friendly disposable plates used conjoint analysis to make critical product design decisions. Learning about his different considerations is useful for understanding both the nuts and bolts and the potential benefits of the method.

Olivier provided some guidelines and recommendations for designing the conjoint questionnaire, including ways to make the task of responding easy and to reduce biases, using objective descriptions and mutually exclusive attribute levels.

He then concluded his presentation with a short overview of additional methods. When building complex products, where customer needs or “the job to be done” are more important than the actual features of the product, it makes sense to focus on uncovering these needs, then structuring and analyzing them. This can be done using the “Voice of the customer” method, or one of the methods described in our previous May webinar. Once you have a clear picture of customer needs, you can use a method called “The house of quality” to tie between them and the attributes of the product that you are building. More information on this can be found in Olivier’s slides.

As a final point, Olivier highlighted the importance of concept testing – after conducting conjoint analysis or other research to help decide on a high level “package” of product features, using online tools to present the idea and ask for feedback is an easy yet highly effective way to gain valuable insight, when it isn’t too late.

All Posts

Almost done…

We just sent you an email. Please click the link in the email to confirm your subscription!

OKSubscriptions powered by Strikingly