Skai and Stratably just published the number that should end the dashboard era in retail. Only 15% of marketers say they are very or extremely effective at measuring retail media performance. The barrier that almost nobody named was access to data. Zero percent said they lacked retailer or platform data.
Sit with that. The data exists. The pipes are connected. The platforms are sharing. And we still cannot turn measurement into a decision.
That is not a data problem. It is an architecture problem. Most omnichannel measurement stacks are dashboards in a trench coat. They report beautifully. They do not close the loop.
A loop is not closed when a chart updates
The word “closed-loop” gets thrown around like it means something. In practice it usually means a SKU-level sale got stitched back to an ad exposure, rendered in a quarterly readout, and admired. That is not a closed loop. That is a slower mirror.
A loop closes when the output of measurement becomes the input to the next allocation decision, automatically, without a human translating a chart into a media plan three weeks later. The test is simple. If your measurement layer cannot move a dollar of spend on its own, you do not have a loop. You have a report.
The Skai data exposes exactly where the loop breaks. 75% of brands name incrementality as their hardest measurement problem. 59% struggle with cross-channel measurement. Only 20% are good at both measuring incrementality and applying it to decisions. The break is not in collection. It is in the gap between knowing and acting.
What closing the loop actually requires
From the forward-deployed seat inside enterprise retail, a stack that closes the loop has three layers, and most stacks are missing two of them.
First, identity resolution. You cannot measure what you cannot tie to a person. Deterministic matching on captured email, phone, and authenticated customer IDs has to be the spine, with probabilistic as the backup, not the other way around. If a customer browsing on mobile and buying in-store reads as two strangers, every downstream number is fiction. Fragmentation is the enemy. Amazon measures at SKU level, Meta by audience, Google by keyword, and your stack has to reconcile all three to one human.
Second, incrementality, not attribution. Attribution tells you which touch got credit. Incrementality tells you what would not have happened without the spend. Those are different questions, and only one of them justifies a budget. Half of brands measure incrementality at a basic level only. Holdouts and controls have to be standard practice, always-on, not a once-a-quarter lift test. An iROAS that is modeled instead of measured is a guess wearing a lab coat.
Third, and this is the layer almost everyone skips: an action layer. Identity and incrementality produce a signal. The action layer turns that signal into a change in spend, inside the platforms, on a cadence the market actually moves at. Re-allocation should be a rule the system executes, not a slide a human builds. When incrementality says a retail media network is buying sales you would have won anyway, the budget should already be moving before the readout deck gets formatted.
The Cost of Doing Nothing is not flat, it compounds
Here is what the 15% number really tells me. The market has accepted measurement that cannot act as the price of doing business. That acceptance is the most expensive line item nobody puts on a budget.
The Cost of Doing Nothing in measurement does not stay still. It compounds. Every cycle you measure without acting, you fund non-incremental spend you have already proven is non-incremental. You let identity gaps misattribute a store sale to a channel that did not earn it. You widen the distance between the brands building control systems and the brands buying another dashboard.
And the gap is structural. The 56% who say they lack internal analytics resources will not close it by hiring three analysts. You do not staff your way out of an architecture problem. You build the loop, or you keep paying the tax.
The brands that win the next phase of retail media will not have prettier dashboards. They will have measurement that behaves like a thermostat instead of a thermometer. Reporting reads the temperature. A loop changes it. Stop buying the thermometer. Build the thermostat.