Tom H. C. Anderson - More Than Market Research

Anderson Analytics’ Information Advantage

Tom H. C. Anderson - More Than Market Research header image 2

Conjoint & Discrete Choice

July 22nd, 2008 · 2 Comments

This week Conjoint and Discrete Choice are on my mind. I read somewhere recently that about 300 such studies are done in the US each year. Not too many I would say.

Made me wonder what proportion of market researchers have ever conducted a conjoint, and among those that do use them what the most popular method is? It seems our clients often prefer to use whichever technique they already have experience with.

Sawtooth is obviously the leader in this field. Surprisingly other larger analytics software providers haven’t seemed to challenge them much. Sort of surprising, especially since the math doesn’t seem too difficult.

Tom

Tags: Analytics · Market Research · Methods · Segmentation · Surveys

2 responses so far ↓

  • 1 Paul M. Banas // Jul 22, 2008 at 10:04 pm

    I think both conjoint and discrete choice are powerful tools to understand why people choose one thing over another, and I find them useful for many research issues. They can be used on anything from optimizing new product bundles, to packaging testing, to determining optimal price thresholds.

    Another choice model that I think has a wide range of applications is Max/Diff. When you have a battery of agreement statements that tend to cluster into non-discriminating lumps, Max/Diff choice models tend to tease out the real hierarchy that exists beneath the surface of universal agreement.

    Just curious on why Conjoint and Discrete Choice are on your mind these days?

  • 2 Tom H. C. Anderson // Jul 23, 2008 at 8:12 pm

    We do a couple Conjoints per year (often together with Segmentation - a key strength for us). Each time the client has a specific way they are used to/feel comfortable with. We’ve done them a few different ways, and I did have a preference, but decided it was time for a refresher. Everyone on our team have different levels of experience with this methodology, so this week two of us dedicated our time to learning EVERYTHING possible about Conjoint, Discrete Choice, MaxDiff etc.

    It’s not Friday yet, so won’t say if I’ve changed my mind, though I’m starting to lean slightly toward CBC HB or CBC Latent Class. Still I’m really a fan of KISS (Keep It Simple Stupid), so Full Profile is always attractive too.

    Naturally it depends on the situation. But trying to find out what to recommend in general. I understand at least among Sawtooth customers CBC HB seems to be most popular.

Leave a Comment