And therein lies the opportunity.
I recently spent time in conversation with Sean Turner, co-founder and CTO of Swiftly, and it clarified something that’s been sitting just below the surface in grocery retail.
Spend a few minutes in a typical mid-market store and everything looks functional. Checkout moves. Shelves are stocked. Promotions are doing what they’ve always done. There’s no obvious breakdown.
But if you follow a single shopper across a full trip, the seams show up quickly. The ad they saw earlier rarely connects to what happens in-store. The offer they may have clipped doesn’t always surface at the right moment. The choices they make in the aisle don’t consistently feed back into the next interaction. By the time the transaction is complete, most of that context is gone.
What you’re left with is a system that resets itself after every purchase.
Turner’s perspective is grounded in tracking that behavior more continuously. The emphasis isn’t on adding more channels or features – it’s on making sure the signals from one moment carry forward into the next. Once you start looking at it that way, the traditional structure around acquisition and retention feels a bit artificial. The same mechanisms are doing both jobs. A campaign meant to drive a visit is also shaping what happens during the trip and influencing whether the shopper comes back.
The challenge is that most retailers still experience those as separate systems. Media, loyalty, in-store execution – they operate with different data, different timing, and different measures of success. Connecting them requires something more basic: the ability to recognize the same customer across those moments and tie that identity to what actually happens at checkout.
That single capability shows up everywhere once you start looking for it. It affects how offers are targeted, how campaigns are measured, how budgets get allocated. Without it, a lot of what passes for personalization or attribution ends up being educated guesswork.
The same kind of shift is happening in how incentives are being used. Turner walked through the alcohol rebate example, which on the surface is a practical workaround for regulatory limits. The mechanics are simple – a shopper selects a rebate, completes the purchase, and receives value shortly after through familiar rails like PayPal or Venmo.
What’s more interesting is how that changes the structure of the promotion. The value can be tied to specific behaviors, timed differently, and measured against comparable shoppers who didn’t receive it. That creates room to understand what actually drove the trip and what else came with it. In some cases, a single promoted item ends up pulling through a much larger basket.
That kind of visibility has been harder to get in physical retail. Which is why the measurement conversation matters as much as the incentive itself. The move toward comparing exposed and unexposed groups, and focusing on incremental outcomes, brings a level of discipline that hasn’t always been present in store-driven media. It forces a clearer answer to a simple question: what changed because this ran?
The AI discussion stayed relatively grounded. There’s interest in how it might reshape planning and discovery, especially around meal ideas and product suggestions. But the more immediate issue is access. These systems can only work with what they can find and interpret. If store data, inventory, and pricing aren’t structured in a usable way, the retailer doesn’t meaningfully show up in those interactions.
There’s a familiar pattern in that. Retail went through a similar phase when search became a primary entry point. Visibility depended on how well information was organized and surfaced. The same dynamic is starting to form again, just through a different interface.
Taken together, these pieces point to a broader shift in how grocery operates. The transaction still matters, but it’s no longer the only moment that defines the relationship. The signals around it – before, during, and after – carry more weight than they used to.
What Turner and the team at Swiftly are doing is working through how to connect those signals into something usable. It’s not a finished model, and it’s not the only approach, but it does make the direction more concrete. You can see how acquisition, engagement, incentives, and measurement start to line up when they’re tied to the same underlying behavior.
There’s still a fair amount of friction in the system. Data isn’t always clean. Integrations take time. Economics are still being worked through in places.
But the retailers that start to operate with a clearer view of that full loop – rather than isolated moments—are going to have an easier time making sense of where to invest and what’s actually working.
That’s the shift Turner is pointing toward. And it’s already starting to take shape.
