How to improve products? Survey consumers with “active machine learning”

MIT Sloan Prof. John Hauser

When you buy a house, it would be irrational to search every possible house on the market. Instead, you narrow down your choices based on things like price, location, and number of bedrooms. The same thing happens when you buy a car. You might only look at sporty coupes or hybrid vehicles. Everyone has their own individual methods – or heuristic decision rules — for screening products, usually based on the item’s key features.

This presents a significant question for companies:  How do you determine what these decision rules are? Managers are increasingly interested in this topic as companies focus product development and marketing efforts to get consumers to consider their products or prevent them from rejecting the products without evaluation. If they better understood consumers’ heuristic decision rules, they could use this information in the design and marketing of new products.

Prior methods of addressing this issue have involved using a fixed number of survey questions. While those methods might be sufficient for some products, they don’t work for products with a large number of features. It’s exponentially complicated to determine heuristics when consumers could be screening based on 50 or more possible product attributes.

For these types of products, we found that it’s far more effective to apply active machine learning methods. Using an algorithm, we created a web-based survey that actively learns how to ask consumers questions based on prior answers. Our main criterion was that each question needed to provide the most information possible about a consumer’s decision rules. The method was fast enough to ask the next question in under a second and identified the features consumers want in a product that are most likely to make it into their consideration set.

We also showed that the method can be used to see if advertising and other communications change the way people evaluate products. For example, if a car company believes it has superior interiors to its competitors and emphasizes this in an ad, active machine learning methods can determine if the ad will actually get more people to consider the car.

The breakthrough here is that without machine learning methods, you couldn’t really identify the decision rules within a reasonable number of questions. But with this method, you can quickly figure out consumers’ heuristic rules for a wide range of product categories – basically anything that has a large number of features like cars, smart phones, computers and even real estate.

This will help managers do “what if” projections when they are thinking about introducing a new product or launching a new advertising campaign. They’ll get a much more accurate projection of how consumers will respond. And with more information on how consumers make decisions, they’ll have a competitive edge when it comes to designing and marketing new products.

John Hauser is Kirin Professor of Marketing

Note: The paper, Active Machine Learning for Consideration Heuristics,  was written by MIT Sloan doctoral student Daria Dzyabura and Prof. John Hauser

Share your thoughts

Post a Comment

Required fields are marked *