January 2026 | Advanced Analytics

Predict the Future with Choice Modeling

Most marketing questions and issues are complex and involve tradeoffs among many variables. Choice modeling allows us to precisely measure these tradeoffs and create simulation models so that optimal decisions can be made, given a set of conditions and variables.

Here are some valuable uses of choice modeling:

• Strategy Optimization: What core concepts, positionings, and messages can form the nucleus of a grand strategic vision for a brand or a company?
• Set the consumer’s mental agenda (what consumers think about).
• Optimal Pricing: Determine price-demand curves and maximize revenue and/or profits. What variables (product quality, positioning, messages, images, distribution channel, etc.) allow a brand to command higher prices over the long-term?
• Technology Forecasting: What might the future look like, given a series of potential technological developments, environmental variables, and economic conditions?
• New Product Development: Thousands of new product concepts can be created and screened with choice modeling experiments. The cost per new product idea screened may only be $50, or $100, or $200. A simulator can be developed to forecast the relative market potential represented by each new concept.
• Portfolio Management or Product Line Optimization: Evaluate complex investment portfolios, product lines, package offerings, and feature bundles to understand how to optimize market share, sales revenue, or profits.
• Feature Optimization: Identify the specific product features, design elements, color patterns, etc., that maximize purchase interest while minimizing manufacturing costs.

Make Choice Modeling Better

Here are five tips to help insure that your company gets the greatest value from investing money in choice modeling:

1. Invest in Qualitative Research First
   Before designing the choice modeling experiment, use qualitative research to identify the key variables, set realistic ranges for those variables, and learn what combinations of variables to exclude. The goal is to measure and optimize what is really important to consumers, to make the choice tasks as realistic as possible, and to ensure that consumers can understand and accurately respond to choice tasks.
2. Simulate the Real Shopping Experience
   The more realistic the experimental stimuli, the more accurate the results. Use an interactive, digital shelf-set that mimics what consumers experience in front of a store shelf or while purchasing online. Allow participants to pick up and look at packages more closely, place and remove packages from a shopping cart, make purchases, etc.
3. Calibrate the Model
   Choice models can almost always be improved by calibrating the data to actual market shares, brand awareness levels, or distribution levels. This step grounds the theoretical model and improves predictive accuracy.
4. Use a Realistic Competitive Set
   Accuracy drops significantly when competitive brands are omitted. Include a realistic set of competitive brands so study participants are making choices that are more similar to the choices they make in the store or online.
5. Build a DecisionSimulator™
   Choice modeling experiments allow explanatory equations to be derived, and these equations can be combined into an interactive simulator, so that analysts can play “what is” games by varying prices and other inputs–to make optimal marketing decisions.

Choice modeling and AI

Choice modeling is mathematically precise and predictive. There is no hallucination, wonky results, or “workslop.” You know exactly what the variables are, precisely what the model outputs are, and you can fully trust the results. AI can be used on the front end to help identify possible variables or ranges of variables, as a complement to the qualitative research, but it is of limited value from an analytical standpoint.

A Forward Look

Choice modeling allows us to look into the future. By the variables, environments, and context of the choice modeling experiments, the future can be imagined and explored. What happens if the price of gasoline goes to $5.50 a gallon? What happens if the prices of electric cars falls below the prices of gasoline vehicles? What happens if the range of electric cars moves from 200 miles to 600 miles? Choice modeling creates the binoculars to help us see and navigate the future.

Jerry W. Thomas

Chief Executive Officer

Email Jerry

Bonnie Janzen

President

Email Bonnie

Elizabeth Horn, Ph.D.

SVP, Advanced Analytics

Email Beth

Audrey Guinn, Ph.D.

Statistical Consultant

Chris Hammack

Sr. Statistical Consultant

We Are Here To Help