Sawtooth Software Conference 2010, October 4-8, Newport Beach, California
March 04, 2011
RTi Research recently joined over 150 researchers and academics from around the world for Sawtooth Software’s fifteenth conference. Sawtooth Software is one of the leaders in online survey research software, with a focus on advanced techniques such as conjoint (choice) analysis.
Conjoint experts presented on practical and theoretical applications of current and future conjoint techniques over the design course of three days. Highlights of the conference included:
Presentations covered all aspects of how respondents make decisions in choice based tasks – the effect of the design (e.g., length, visual display, live vs. virtual stimuli) on respondents’ choice, how respondents narrow down the choice set, how respondents process the large amount of information presented to them during a typical conjoint study, etc. These discussions help keep RTi on the leading edge in conjoint study design.
Analytic discussions touched on technical theories and advancements in some existing techniques. We’ll leave the technical theorizations to the academics, but are excited about improved segmentation and line optimization capabilities using conjoint methodologies.
- Advances in Conjoint Techniques:
Sawtooth developers presented on the next breakthrough in conjoint techniques, Menu Based Choice (MBC). MBC promises to offer greater flexibility in design by allowing respondents to add on desired features to choice sets. These features may be priced to understand whether consumers are willing to pay a premium for one or more features. (As a side note, RTi Research has been involved with MBC as a beta tester over the past year and will remain on the forefront of this new technique.)
- Future Directions in Conjoint Techniques:
Presentations also led to discussions of future development directions, with particular interest in a new spin on discrete choice – Best/Worst Discrete Choice. As the name suggests, rather than picking one preferred option, the respondent selects the bundle that they like the most and the one that they like the least. This technique maximizes the learning from each discrete choice question to lead to more informed modeling while minimizing respondent fatigue.