Can a Shopping Cart Make You Spend More? We Checked the Data

Every retailer is chasing the same shiny idea right now: put a screen on a shopping cart, add some AI-driven personalisation, and call it “the future of retail.”

But does it actually work? Or is it just another gadget that excites trade press and gets ignored by real shoppers?

My co-authors (Sabrina Gottschalk, Annika Lorenz-Kornfeld, Nils Eisler) and I set out to answer this question properly, with real data, not assumptions. Our paper, “Customer responses to smart shopping carts in supermarkets,” was just published in the well ranked, high impact factor Journal of Business Research. We analysed 12,418 real shopping sessions over one month, from a major German supermarket piloting smart carts equipped with screens, in-store navigation, and personalised offers.

Here’s what we found.

The Numbers Don’t Lie

Shoppers who actively used the smart cart screen, compared to those who didn’t:

• 💶 Spent 32% more per basket (€22.89 → €30.18)

• 🛒 Bought 25% more items (9.61 → 12.02)

• ⏱️ Stayed 23% longer in-store (32.75 → 40.40 minutes)

That’s not a small nudge. That’s a meaningfully different shopping trip.

(If you want help turning findings like this into an actual AI and customer experience strategy for your business, that’s exactly what I do in my corporate training and consultancy work. More on that at the end.)

But Here’s the Twist: More “AI Engagement” Isn’t Always Better

This is the part that should make every marketer pause.

We also looked at “superusers”, shoppers who tapped the cart’s screen 20 or more times per trip. You’d assume they’re the biggest spenders, right?

Not exactly.

• Superusers did buy more items and stay in the store longer.

• But their actual basket value levelled off. And past a certain point of screen taps, it actually dropped.

Beyond a certain point, extra taps look more like curiosity than buying intent. The relationship between engagement and revenue isn’t a straight line. It’s a curve, with a sweet spot in the middle.

This same pattern showed up twice in our data, independently. Conversion rate (whether someone acted on a promotion shown on the screen) followed the exact same shape: it rose with engagement, then fell off again at very high engagement levels. When a result repeats itself across two different outcomes, that’s a much stronger signal than a one-off finding.

We saw the opposite pattern with shoppers who uploaded a shopping list before they shopped: fewer items, but higher value, and faster trips. Classic efficient, task-focused shopping. Interestingly, list uploaders still had a higher conversion rate than other shoppers. They may not browse much, but they still notice and act on a relevant offer when it’s in front of them.

Two completely different shopper types, using the exact same piece of technology.

A Quick Word on the Method (Because the Details Matter)

I get asked a lot in my workshops: “how do you know this isn’t just coincidence, or busier shoppers naturally spending more?”

Fair question. Here’s the short version.

We didn’t just compare two averages and call it a day. We ran regression models for different outcome variables since different outcome variables need a different statistical approach.

Every model controlled for the context that naturally shapes shopping behaviour: time of day, weekday versus weekend, and even weather, since temperature and rainfall both affect how long people stay in a store and how much they buy.

We then went a step further and compared smart cart users to non-users only within identical time-of-day and day-type windows, so we’re not comparing a Tuesday morning shopper to a Saturday evening shopper and calling it a fair fight. Even after that stricter comparison, the gap held: smart cart users still spent about €5 more in the afternoon and €3.41 more in the evening, bought 1.3 to 1.5 more items, and stayed 5 to 7 minutes longer than otherwise similar shoppers.

None of this proves causation. Shoppers who choose to use a smart cart might simply be different kinds of shoppers to begin with, and we’re upfront about that limitation in the paper. But it does mean the patterns we found are not a statistical accident, a busy Saturday illusion, or an artefact of one specific model choice.

Why This Matters Beyond Grocery Carts

This is the part I keep coming back to in my AI workshops: the technology itself is rarely the differentiator.

A smart cart, a chatbot, a recommendation engine, an AI-generated ad. None of it automatically creates value. What creates value is the marketing and behavioural design wrapped around the technology: what content you show, to whom, at what moment, and how well you read the signals coming back.

The cart didn’t make shoppers spend more. The ability to use real-time behaviour as a marketing signal did.

That’s the argument I keep making: in the age of AI, your competitive advantage isn’t the tool. It’s how well you design the experience and the marketing strategy around it.

What This Means for Your Business

If you’re exploring AI-enabled customer touchpoints such as in-store tech, apps, chat, or recommendation engines, here are four practical takeaways:

• Track engagement quality, not just volume. More interactions don’t automatically mean more revenue. In our data, the relationship had a clear sweet spot.

• Segment by behaviour, not just demographics. “List uploaders” and “superusers” need completely different content strategies.

• Context changes everything. The same technology performed differently by time of day and day of week. Your AI strategy needs to account for when, not just what.

• Validate before you scale. We controlled for context and ran robustness checks, and the patterns still held. Before rolling out an AI feature company-wide, it’s worth testing whether the effect survives the same scrutiny.

If you’d like to explore how to apply findings like this to your own organisation’s AI and customer experience strategy, that’s exactly what I help companies do through corporate training and consultancy. Get in touch via www.yusufoc.com.

📄 Full paper: Gottschalk, S.A., Lorenz-Kornfeld, A., Oc, Y., & Eisler, N. (2026). Customer responses to smart shopping carts in supermarkets. Journal of Business Research, 215, 116337. https://www.sciencedirect.com/science/article/pii/S0148296326003723?dgcid=coauthor