Smart Cart Study Links Screens to Higher Spending

The reported differences extended beyond basket value. According to the university release, smart-cart users:

• •Spent an average of EUR30.18, compared with EUR22.89 for non-users.
• •Bought 12.02 items on average, versus 9.61 items, a 25% difference.
• •Spent 40.40 minutes in the store on average, compared with 32.75 minutes, a 23% difference.

The research team collected basket value, basket size and time in store for each session, then linked cart activation to the time of day and day of week. The release reported larger spending differences during afternoons and weekends, while basket-size differences were more pronounced in afternoons and evenings. It also found that higher temperatures shortened average shopping-trip duration across customers.

FoodNavigator reported that the data came from a one-month period in March 2025 at a German supermarket chain piloting the carts. That outlet described the devices as a possible retail-media surface, where brands could seek visibility close to the point of purchase.

The study separately examined "superusers," defined as customers who made more than 20 screen interactions in a session. The university release found that these customers bought more items and spent longer in the store, but did not spend more money overall.

Yusuf Oc, a study co-author and senior lecturer in marketing at Bayes Business School, told FoodNavigator that heavy users may use the display to compare products and prices. "They are actually finding better deals, spending more time with this tablet because they can ask for different products and see the prices without actually going to the shelves directly," Oc said.

Lead author Sabrina Gottschalk summarized the aggregate result in comments to CNET: "Smart cart users in our data spent more, bought more, and stayed longer than non-users."

The reported results compare smart-cart users with non-users, rather than describing a randomized experiment. As a result, the differences are associations and do not independently establish that the cart interface caused higher spending. Customers choosing a smart cart could differ from other customers in ways that also affect trip duration or basket size.

For retail data and ML teams, the distinction matters when evaluating recommendation, navigation and advertising systems. Comparable deployments commonly require controlled tests that separate interface effects from selection effects, while measuring outcomes beyond revenue, such as substitutions, promotion exposure, trip completion and consumer choice. The superuser finding also illustrates why interaction count alone is not a sufficient proxy for monetary value.

CNET noted that Instacart and Weis Markets had announced a rollout of AI-driven Caper Carts to selected stores. The Bayes study provides field evidence that such screens can correlate with substantial changes in shopping behavior, while leaving open which product features, customer segments and promotional designs account for the observed differences.