How store sales reporting works
You take the transactions your till or point-of-sale system already records — the customer’s email address or phone number, the amount, the date — and send them to Google. The identifying fields are hashed before they leave your systems, so what travels is a scrambled fingerprint rather than a readable address. Google hashes the details of signed-in users the same way, compares the two sets, and where a fingerprint matches a person who clicked an ad within the conversion window, the sale is credited to that campaign.
Some businesses upload the file themselves; larger retailers can route it through an approved data partner instead. Either way, only a share of transactions ever match, because not every customer gives a contact detail at the counter, not every customer is signed in, and details entered wrongly will not hash to the same value.
Why store sales matter
This is the closest thing paid search has to proof for an offline business. Unlike a modelled figure, a matched sale is a real transaction with a real amount attached, which means the campaign can finally be judged on revenue rather than on enquiries. That changes budget conversations: a keyword that looks expensive per lead can turn out to be the one bringing in the larger baskets.
It also lets bidding work towards value. Once genuine revenue flows back into the account, value-based strategies have something honest to optimise against instead of a flat number assigned to every form fill.
Common mistakes with store sales
Treating the matched total as your whole offline revenue is the biggest one. It is a matched subset, always smaller than the true figure, so use it to compare campaigns against each other rather than to calculate your real return.
Two data problems follow close behind. Dirty customer records — misspelled emails, phone numbers stored in inconsistent formats, a shared shop email typed in for everyone — destroy matching quietly, and nothing in the interface tells you that is why the numbers look thin. Uploading the same transactions twice, or running store sales alongside an online purchase conversion that already covers click-and-collect, produces double counting that inflates results.
There is also a consent obligation. Sending customer details to any platform needs a lawful basis and a privacy notice that actually describes it, which is a conversation to have before the first upload rather than after.
How to act on it
Fix the data before the plumbing. Ask for an email or phone at the counter as a normal part of the sale, standardise the formats, and clean out obvious rubbish. Then decide which conversion actions cover which purchases so nothing is counted twice, keep the upload on a regular schedule, and hold store sales in a secondary column until the match volume looks steady. Only then wire the revenue into bidding, alongside the rest of your conversion tracking.