Edward Malthouse
Erastus Otis Haven Professor and Research Director of Spiegel Research Center
Northwestern University
Edward Malthouse is the Erastus Otis Haven Professor at Northwestern University's Medill School and research director of its Spiegel Research Center, where he builds statistical models from large sets of consumer and audience data. In 2026 he led a study of which brands six large language models recommend, and found that many established brands are left out.
Measuring Which Brands AI Recommends
"A brand omitted from an LLM's recommendations may never be considered, even when it fits the consumer's needs," Malthouse and his co-authors wrote in Evaluating Brand Retrieval and Ranking in Large Language Model Recommendations, posted in September 2026 with Kun-Yu Lee, Jing Yang, Sanchary Pal and Xueyan Feng. They described this as "a measurement problem for marketers," because the set of brands a model may recommend is neither fixed nor visible, and the same question can return different brands in a different order.
The team asked two models each from OpenAI, Google and Anthropic for up to five brands of cordless drill, hiking jacket, coffee maker, cat food and boat cruise, 40 times per model and category, and scored how often and how high each brand appeared. These category-only prompts often left out large, established brands, and prominence followed conventional brand popularity only weakly. Google search interest was the strongest predictor of prominence, followed by online brand conversation, while news mentions, advertising spending and Wikipedia page views added little once the measures were considered together. The authors cautioned that these links are exploratory and should not be read as causal.
Positioning Brands for AI Answers
When prompts described a shopper's goals, budget and intended use, different brands surfaced, and brands missing from ordinary answers could appear once distinctive cues were supplied. The paper advised managers to "communicate the intended positioning consistently across the touchpoints they control and cultivate congruent associations across those they can influence." In the authors' words, "The objective is not simply to maximize the volume of information about the brand, but to establish a coherent set of associations between the brand and the consumer needs for which it should be considered." The team released its software and data so others can repeat the measurements.
From AT&T Laboratories to Medill
Before academia, Malthouse worked in software engineering for AT&T Laboratories, and he has since run analytics training for companies including Accenture, BNSF, Digitas and Capital One. At Medill his research covers media marketing, database marketing, advertising and integrated marketing communications, and he also holds an appointment in Northwestern's Department of Industrial Engineering and Management Sciences. He co-edited the Journal of Interactive Marketing from 2005 to 2011 and won the IMC Teacher of the Year award in 2010 and 2014.
Through the Spiegel Research Center, he has also analyzed reader data from the Chicago Tribune, the San Francisco Chronicle and the Indianapolis Star for Northwestern's Local News Initiative, to learn what content and habits lead people to pay for local news.