5 Questions to ask when choosing between Conjoint Analysis and Survey Tools

by Julia Kastner

For my field course project, I wanted to test a set of hypothesis around customer preferences. I’m launching a women’s jeans company, Kismet Denim, making better-fitting jeans for sophisticated women using Fair Trade and Organic materials. In order to afford the Fair Trade and Organic premiums, I’d like to be able to cut out some of the expensive processes jeans go through as they turn from a dark blue indigo to lighter, more royal shades. But which processes could I cut out? To answer this question, I wanted to know which jeans attributes where important. I developed a set of hypotheses but I had to choose between two tools to test them:

Conjoint analysis: a type of survey that breaks a product concept into relevant attributes and estimates the value potential customers attach to each attribute

Survey tools: showing photos of two alternate products to customers and asking the survey taker to choose the preferred picture 

Here are five questions that can help you choose which option is best for you: 

  1. Do you have a survey sample available? If so, how big is it? Do you have access to a pool of people to answer your survey? I knew I had access to a small number (I ended up with 110 responses) but not the 400-person sample size required to achieve 95% confidence in a conjoint survey.
  2. How much money are you willing/able to spend? According to Survey Analytics (I spoke with Jack Geiger), a student subscription costs $1,500 for a quarter of a year. This does not include potential costs involved in paying for survey takers. As a very early stage entrepreneur, I did not feel I could justify this expense.
  3. How tolerant are your survey takers to repetition? What are your survey-takers’s incentives? Can you predict how many COMPLETE surveys you will have at the end? My survey was only 19 questions long and I received comments from test takers that the found the survey repetitive. Conjoint, as a necessary function of it’s design, is even more repetitive, so even if you have access to a large pool of survey-takers, they may not actually complete a conjoint survey.
  4. How complex is the hypothesis you’re testing? How specifically do you need to disaggregate product features In one question, for example, I was intending to test the hypotheses that women don’t prefer “whiskers”– false creases along the hip area. When testing this, however, I used pictures of jeans that not only had different whiskers but also different color denim underneath. The answers disproved my hypothesis. In later questions, however, answers were aligned with my hypothesis – questions that better isolated the feature I was trying to test. A conjoint, which tests attributes individually by showing packages of features, would have made my question more clear. Similarly, it was difficult to test customer willingness to pay – in a conjoint you can add a price to packages of attributes but in the survey I had difficulty testing this.
  5. Can you explain product attributes in text or do you need photos or other media? Conjoint works best with attributes you can describe in words – are you testing for computer memory or car gas mileage? Conjoint tools may not have photo capabilities, and some attributes (like “whiskers”) are easier shown than described. Survey Analytics did have a photo-ranking tool, where customers could rank pictures in order of preference. By including multiple pictures, this tool might be better equipped to isolate features, but not as well as conjoint. 

Overall, I have learned how conjoint analysis and survey tools can be extremely useful and powerful. I learned about my customers by conducting a survey but I can see the shortcomings with this tool. As a very early entrepreneur, I do not have the resources for conjoint, but I can imagine companies with more resources investing in this service to research customer preferences – either for introducing new products into the market or for monetization of a product that has achieved market fit. 


Popular posts from this blog

Quiz Time 129

TCS IT Wiz 2013 Bhubaneswar Prelims

The 5 hour start-up: BrownBagBrain