The 3-Question Framework for Deciding What to Fix First in Your Retail Data

Every brand we talk to has the same moment. Someone pulls the store-level data — orders, shipments, visit logs, whatever mix you've got — and asks the question that actually matters: what are the 3–4 things we should do immediately based on this. Not a dashboard. Not a QBR deck. An answer.

Most teams don't have a bad data problem. They have a triage problem. You're staring at a thousand doors, a dozen KPIs, and a sales team with limited field hours, and everything looks like it needs attention. It doesn't. Almost every retail execution issue collapses into one of three buckets: coverage, activation, or velocity. Figure out which bucket you're actually in before you send anyone into the field, because the fix for each is completely different, and fixing the wrong one wastes a quarter. Most brands run this off sell-in data — orders and shipments, not point-of-sale scans — so the framework below is built for what that data can actually tell you, with reps as the tool that resolves what it can't.

Question One: Is This a Coverage Problem?

Coverage means product was never actually stocked in a door you're authorized in. Pull your authorized door list and check how many have EVER received a shipment. If you're authorized in 800 doors and only 550 have ever received product, that's not an execution issue, that's a rollout that stalled — a reset that never happened, a DC that never started shipping, a broker who said yes and never followed through. If that gap is over 10–15%, stop looking at anything else; velocity numbers on doors that were never stocked are meaningless. Fix: a phone call to the buyer or distributor, not a field visit.

Question Two: Is This an Activation Problem?

Assume coverage is fine — every door has received product at some point. Now ask which of those doors have gone quiet relative to their OWN normal ordering pattern: a door that reorders every 6 weeks showing 12 weeks of silence is a real flag; one that's always ordered quarterly showing 10 weeks isn't. This tells you who to look at, not why — a door gone quiet looks the same whether it ran out of stock and nobody noticed, or it's fully stocked and genuinely not selling, and sell-in data can't tell those apart. That's what the visit is for: a rep sent to the shortlist, not the whole territory, with one job — check the shelf. Empty means it's a coverage problem that started late. Stocked and untouched means it graduates to question three. Fix: a targeted rep sweep of the shortlist.

Question Three: Is This a Velocity Problem?

Coverage is good, activation is good, product is on shelf and selling — and it's still underperforming. This is the one people jump to first because it's the most visible number, but it's the last thing to check, because low velocity caused by a coverage or activation problem looks identical to low velocity caused by real demand weakness. Once you've ruled out the first two, compare velocity against a comp set in the same store class, not your own network average. A door doing 2 units/week against a category comp of 6, with full distribution and steady ordering, is telling you something real — price, assortment mismatch, or a genuine demand gap. Fix: trade, pricing, or assortment, not another rep visit.

The Shortcut

Run the three questions in order — coverage, activation, velocity — and stop at the first one that's broken. We had a brand come to us convinced they had a velocity crisis in one region: sales down 18% versus plan. Two hours in the order data showed 22% of their "authorized" doors in that region hadn't received a single shipment in six weeks. Not a velocity problem. A coverage problem wearing a velocity costume.

The takeaway: don't start with the KPI that's screaming loudest. Start with the sequence. Coverage first, activation second, velocity last. Whichever one breaks first is where your next two weeks of field time and phone calls should go — not all three at once, and not whichever one happens to be on the slide someone already built.