Store visits, the number nobody shows you.

Most retail reports stop at the click. Sessions, cost per click, maybe a form fill. Then the report goes quiet about the thing that pays your lease. Somebody walked in, tried on three coats and bought one. That walk-in has a number attached to it, and it rarely reaches the owner.

The number is a store visit, and it is modelled. It rests on shoppers who agreed to share location from their phones. A sample gets observed. A model extrapolates. What lands in your report is an estimate for the period. Aggregated, never a list of names.

Be clear about what that makes it. An estimate is not a headcount at your door, and it will not tie out to your door counter or your till. Nobody is following a shopper from an ad into your fitting room. So a headcount claim tells you more about the provider than about your door.

Direction is enough to decide with.

Precision is seductive. Click-through rate reads to two decimals. It is also close to useless for a furniture store, where the decision happens on a showroom floor days after the ad ran. You can measure clicks perfectly and still have no idea whether the media moved anyone through the door.

Take the trade. A rough read on visits, held steady across weeks, tells you more than an exact read on a step nobody cares about. Direction is enough to move money. Direction tells you the London store answered the spend and the Barrie store didn't.

A fashion store makes the gap obvious. Two shoppers walk in with the same jacket saved on their phones. One saw a video ad midweek. One got a delivery notice and came to swap a size. Your till can't tell them apart. The visit line at least tells you whether the video moved the floor.

This is where the budget argument changes. Without a visit line, every dollar gets judged on e-commerce revenue, because online revenue is the only outcome the report can see. So the plan drifts online. One sensible decision at a time. The stores keep paying rent and stop getting credit for what the advertising does.

Put a visit line in the report and both halves of the business are on the table. Store goals can run inside Performance Max. Business Profile feeds the local surfaces where a shopper checks hours and directions. None of that is exotic. It has to be asked for, and then it has to be read the same way every month.

Keep the visit line in its place, though. It is one outcome among several, and it belongs in the blended read next to online revenue, calls and direction requests. No channel grades its own homework on this desk. A visit number that gets to explain the whole season by itself has stopped being measurement.

Three habits make the number worth spending against.

Hold the geography still. Whatever radius or store list you drew in March is the one you keep in November. Widen it and visits rise, while nothing about your advertising has changed.

Hold the window still. Visits get attributed inside a lookback window. Change that window and you change the number without touching a single campaign.

Read the trend, not the day. Four weeks against the prior four weeks is a reading. Anything shorter is noise. This desk plans in weeks for exactly that reason, and measurement follows the same clock.

Now the ways it breaks. Thin volume is the first one. A single store with few observed visits produces a jumpy estimate, and a jumpy estimate invites confident nonsense. Small numbers move for small reasons.

Some locations never report at all. Volume sits under the threshold the platform will publish, so the cell comes back empty. That is an honest answer and you should want it. Take the empty cell over an invented one, and judge that store on the outcomes it can actually produce.

Store operations break it too. Cut Sunday hours in October and visits fall, and the media had nothing to do with it. Same for a renovation, a fenced-off parking lot, an ice storm, a competitor opening two blocks away. Write those down as they happen. Your report should carry them beside the line, in plain language.

Moving the goalposts mid-season is the worst one. Somebody widens the radius in week six, or swaps a campaign type, or quietly extends the lookback. The line jumps. The deck says the change worked. Nothing worked. The ruler got longer.

Ask your current provider four questions before the next invoice. Is this figure modelled or counted? What geography does it cover, and has that geography changed since we started? What lookback window built it? What else moved in the store during the period? Four straight answers and you can use the number. Hedging on any of them and you can't.

Cost per visit is a line an owner can argue with.

Divide the media spend by the modelled visits and you get cost per visit. Treat it as a trendline, never as a fact. The absolute figure is an estimate, and it will stay an estimate no matter how many decimals anyone puts on it.

An illustration, with hypothetical round numbers rather than benchmarks. Spend $10,000 across four weeks and read 2,000 modelled visits. Cost per visit lands at $5. Hold the geography and the window fixed, spend the same $10,000 over the next four weeks, and read 2,500 visits. Cost per visit lands at $4. Both figures are estimates. The gap between them is the argument.

That is a number you can push back on. Weigh $5 a visit against your average basket and what your floor staff actually close. Ask which store earns the next thousand dollars. A partner who wants to know why November carries more money than March gets an answer in one sentence.

And you can be wrong out loud, early, in weeks rather than quarters. That is the part owners like. A trendline gives you something to disagree with while the season is still running.

Open last season's reporting. Look for the weeks the ruler moved, and the weeks it held. That list is the agenda for a strategy call, and the next plan should be built on top of it.