Join us at Sawtooth Research Conference 2025: Uncover the latest innovations shaping data-driven decision making

Join us at Sawtooth Research Conference 2025: Uncover the latest innovations shaping data-driven decision making
6 - 9 May 2025
New Orleans, Louisiana

SKIM is excited to announce that its team will share their expertise and connect with industry professionals at this year’s Sawtooth Research Conference.

During the summit, our experts will lead a series of academic sessions, workshops and tutorials covering a range of topics, with a special focus on AI in research and the application of MaxDiff. Join our sessions to explore cutting-edge techniques, gain practical insights and stay ahead of the curve in the evolving insights world.

Comparing Between STAN’s HMC No-U-Turn Sampler Versus Sawtooth Software Gibbs Sampler for Complex Conjoint Studies

Date & Time: May 6th, 9:45am

Kevin Lattery

Kevin Lattery

Senior VP, SKIM

Daan Schopmeijer

Daan Schopmeijer

Manager, SKIM

As market research becomes more complex, accurately understanding consumer preferences is crucial. This session will explore how two different techniques—STAN’s No-U-Turn Sampler and Sawtooth Software’s Gibbs sampler—compare when analyzing complicated conjoint data. We’ll highlight how these methods help improve results in studies with limited respondent data.

Working with Subsets of Utilities

Date & Time: May 6th, 2:45pm

Kevin Lattery

Kevin Lattery

Senior VP, SKIM

This session will explore how to work with subsets of utilities in conjoint studies, addressing questions like which subsets are most informative and whether we can predict a full set of utilities using only a subset from one study. We’ll introduce the Conditional Distribution of a Multivariate Normal to model these relationships, helping researchers estimate and predict utility values more efficiently, even when full data is unavailable.

Post-Hoc Constrained Point Estimates Using Distribution of Unconstrained HB Draws and Empirical Bayes

Date & Time: May 6th, 3:00pm

Kevin Lattery

Kevin Lattery

Senior VP, SKIM

In this presentation, we will examine how to handle constraints in Hierarchical Bayes (HB) models, especially when the usual assumption of normality doesn’t apply. Instead of focusing on the average of HB draws, the approach involves working around problematic constraints during the process and applying them later using the distribution of unconstrained HB draws. This method helps address issues like small differences in ordinal levels, leading to more accurate results.

Blending Historical MaxDiff Claim Studies and Using AI to Predict Claim Success

Date & Time: May 7th, 11:00am

Kevin Lattery

Kevin Lattery

Senior VP, SKIM

Nino Hardt

Nino Hardt

Senior Director Methodology and Analytics, SKIM

Howard Huang

Howard Huang

Lead Analyst, SKIM

Predicting the success of claims is a key component of strategic decision-making, and with the growing influence of AI in research, leveraging advanced data models has become more important than ever. Our experts partnered with P&G to conduct a Sparse MaxDiff study and train AI models on a database of 1,554 unique claims to predict claim success. In this session, we will showcase AI’s ability to forecast claim outcomes and highlight the need for a robust claims database to ensure accuracy.

The Will of the Many: Generating Novel Concepts Using AI-Enhanced Respondent Feedback

Date & Time: May 8th, 4:45pm

Peter Li

Peter Li

Lead Marchine Learning Engineer, SKIM

Joris van Gool

Joris van Gool

Manager, SKIM

In traditional product design, designers often limit their concepts based on what they think will appeal to consumers. Our SKIM experts will introduce an AI-driven approach that expands the design process by directly incorporating respondent feedback, allowing us to explore a wider range of product ideas and unlock more innovative possibilities.

Determining the Value of Price Thresholds in Pricing Conjoint Studies

Date & Time: May 9th, 11:00am

Michael Smith

Michael Smith

VP Methodology, SKIM

Juli Pham

Juli Pham

Senior Manager, SKIM

Many clients are now exploring the idea of price thresholds—where demand drops significantly at certain price points—as part of their pricing strategies. In this session, we’ll test both post-hoc and modelled threshold options to determine if these thresholds improve our pricing models, which types work best and the kinds of pricing studies where they’re most useful.

Can’t make it to the Sawtooth Research Conference?

Get the presentation in your inbox: Sign up to receive the event highlights and presentation via email.

About Sawtooth Software

Sawtooth Software provides advanced analytics and insights to help organizations understand what’s important to their customers and to predict what they will buy or choose. They do this through their solutions platform, consulting, and educational services.

The Sawtooth platform handles traditional survey questions, but they are best known for an integrated predictive analytics solution called Conjoint Analysis, or Choice Analysis.

Want to know more?

If you’re interested in learning how new research techniques can help you better uncover, predict and leverage the hidden truths behind decision making, please reach out. We’re happy to talk things through with you!