Why your ads dashboard can't see your in-store customers

01

Start with what the dashboard is actually counting

An ads dashboard is a record of things that happened on screens. It counts impressions, clicks, website visits, form submissions, and — if you have set it up — phone calls and requests for directions. Every one of those events happens inside Google’s systems or on your website, where a tag or a tracking number can see it.

A local business earns most of its money somewhere else: at a till, in a treatment room, across a service counter, or on a doorstep after a quote. Those outcomes happen inside your point-of-sale system, your booking diary, or your CRM, and no tag fires when a customer pays. Unless a recorded connection exists between the two systems, the dashboard and the till describe the same business in two languages that never meet.

This is why the two most common readings of an ads report are both wrong. “The ads show twelve conversions, so they work” ignores every customer who arrived without leaving a digital trace the dashboard could count. “The ads show nothing, so they don’t work” makes the opposite error, treating unmeasured as zero. The honest starting point is narrower: the dashboard shows the part of the journey it can see, and for a local business that part usually ends at the phone call or the directions request. 02

02

The journey your customer takes, and where the record stops

Take a typical path. Someone searches for a service nearby, sees your ad, and taps it. That tap is recorded. They look at your opening hours, tap for directions, and drive over; the directions request is recorded too, because it happened inside Google Maps. They walk in, ask about the service, and pay at the counter. That payment sits in your POS system with a date, a time, and an amount — and nothing that ties it to the tap.

Now vary the path and watch the record change. If the customer calls instead of visiting, the call can be recorded — but only if call reporting is switched on: Google assigns a forwarding number to the ad, routes the call to your real number, and logs details such as the time and duration. You can count calls over a length you choose, long enough to suggest a real enquiry, as conversions. If call reporting is off, the ad shows your direct number and the call leaves no trace in the dashboard at all.

If the customer saves your number and calls a week later from their contacts, the connection is gone. If they see the ad on a laptop, then visit from their phone, the devices may never be joined up in a way you can inspect. Each of these customers is real revenue influenced, or not, by the same ad — yet the dashboard treats them differently based on accidents of the path they took.

03

03

Which offline signals can be collected, and what each one really is

Four kinds of offline signal exist in Google’s measurement system. They differ in who qualifies for them and in how directly they measure anything, so keep them separate. Signals a local business can collect

Store visits deserve a careful explanation, because the name sounds like a headcount. Google builds them from aggregated and anonymised signals — for example, users who are signed in and have location history on — and then extrapolates to represent the broader population of your customers. The result is an estimate of how many visits your ads influenced, useful for comparing campaigns against each other. It is not a list of visits, and it will never reconcile exactly with the number of people who walked through your door. Store sales go one step further and attach values to offline purchases, but they sit at the far end of the scale: accounts must be allowlisted, upload transaction data regularly, and meet first-party data volumes that run to tens of thousands of transactions in a quarter. That is chain-retail territory. A single shop should understand the concept — sending till data back is how the loop closes — without expecting access to that program.

04

04

Why most small advertisers never see store visits in their account

Store visit reporting is not a setting you switch on. Google turns it on automatically for accounts that meet its eligibility criteria, and those criteria are built around data volume and privacy. Your locations must be in a country where store visits are offered. Your ads need to run with location assets, and the locations behind them must be set up and verified in your Business Profile. Your account needs enough ad clicks or impressions, and your business enough foot traffic, to pass Google’s privacy thresholds — and Google is explicit that the numbers vary by advertiser rather than being a single published bar.

Locations Google considers sensitive are excluded altogether; its examples include locations related to religion, sexual interests, children, and certain health issues.

Put together, this means a busy multi-location business in a large market may see store visits appear in its reports without asking, while a single-location shop with modest ad spend will usually see nothing — not because its ads send nobody to the shop, but because its data volume is too small for Google to model visits without risking individual privacy. The missing column is a statement about measurement thresholds, not about whether customers came.

05

05

The other direction: sending your own outcomes back

Because platform-side measurement has thresholds, the route most local businesses can actually use runs the other way: record outcomes in your own systems, with enough information attached to join them back to marketing later.

The join needs an identifier captured at the moment of first contact. In Google Ads, the standard one is the click ID attached to the ad click. If your website form or call system stores that identifier alongside the enquiry, and the record then travels into your CRM, a later outcome — a booked job, a signed contract, a completed sale — can be imported back into Google Ads as an offline conversion, matched to the click that started it. Google’s import routes accept uploaded files and direct CRM connections; its Salesforce import, for example, reads lead and opportunity stages from the CRM and counts the events you define as conversions. Where a click ID is unavailable, enhanced conversions for leads match on hashed first-party contact details, such as the email address or phone number the customer gave you.

The same discipline works without any import at all. A source field in your POS or CRM — “Google ad,” “Google search,” “walk-in,” “referral” — filled in at the moment of booking, plus a weekly habit of comparing that field against spend, gives a rough but honest picture of which channels bring people who pay. It will not match the dashboard, and it is not supposed to: it measures the end of the journey the dashboard cannot see, and the two records are most useful read side by side, with their different scopes understood.

06

06

Three working steps for a shop-sized business

Step one: make the countable things countable. Turn on call reporting so calls from ads run through a forwarding number and are logged with their duration, and set a minimum call length that reflects a genuine enquiry for your trade. Use one consistent set of tracked links and landing pages for your ads, and ask your website form to store the click identifier with each submission rather than discarding it.

Step two: record outcomes where they happen, with the source attached. Every booking, quote, or sale in your POS or CRM gets a source and, where you have it, the stored click identifier. The habit matters more than the software: a source field left blank half the time produces a report that misleads with confidence. Review the previous week on a fixed day, while the staff who served those customers still remember them.

Step three: judge campaigns on store outcomes, in layers. Compare campaigns first on the signals closest to revenue that you actually collect — calls of real length, direction requests, recorded enquiries, and the outcomes matched back to them — and only then on clicks and impressions. A campaign that produces fewer clicks but more recorded bookings is the better campaign, and you can only know that if step two happened. 07

07

The misreadings to avoid

The first misreading is treating unmeasured as zero: closing a campaign because the dashboard shows no conversions, when the dashboard was never equipped to see the shop-floor outcomes the campaign existed to produce. The second is the mirror image: crediting ads with every walk-in during the campaign period, including the regulars, the referrals, and the passing trade who would have come anyway. Both errors come from asking one record to describe the whole journey.

The third misreading is mixing kinds of number. A modelled store visit estimate, a counted phone call, and a sale recorded in your till are three different strengths of evidence. Reports stay honest when each figure keeps its label — observed, modelled, or recorded in your own system — and no figure is silently promoted into a stronger claim. That labelling habit is the same discipline used when comparing competitors’ public data: record what the evidence is, not what you would like it to prove. FAQ