Local Inventory Ads Built on Verified Store Stock

Where this work is won: A local inventory ads programme is decided by one number: the share of store availability records that match what a customer finds on the shelf. Google audits that match itself, by phone call and by site check, and a failed audit takes local inventory offline for the whole account. We hold a 95 percent per-store match across a sample of 300 items before a campaign is allowed to serve. Below that line the ads are buying trips to stores that cannot fill the order, and the store manager stops answering emails about the channel inside a fortnight. How we set it up: We start at the point of sale, not at Merchant Center. Most retailers publish availability from a nightly ERP snapshot, which means a fourteen-hour-old promise on any item that turns more than twice a week. We trace the path from POS to the local product inventory feed and decide whether the account can carry a direct feed submission or needs a Local Feed Partner because the source systems will not produce anything cleaner on a timescale that matters. Where the retailer runs a modern stock ledger, we move availability onto an intraday refresh and tier it by quantity, so a store with one unit left is not presented as equal to a store with forty. Store codes come next, and they break more accounts than bidding ever does. Codes in the feed have to reconcile one to one with Business Profile locations. Closed stores still listed, duplicate profiles left behind by a franchise migration, and stores trading under a second name each produce a silent mismatch that suppresses the listing rather than throwing an error. We reconcile that mapping against the Business Profile location export before launch, and again at every change to the store estate. Then the promise itself. Pickup today, pickup later and on-display-to-order are three different commitments to a customer, and each is only worth annotating if the store can honour it. We set pickup method and pickup SLA to what operations will actually stand behind, which in practice means writing the annotation down to a slower promise than marketing wants it to be. Measurement, stated honestly: Store visits are modelled. We report them as context and never as the outcome a budget is defended with. What we instrument instead is countable: click-and-collect orders carrying the store code, and in-store transactions matched through a hashed offline conversion upload from the POS or the loyalty file. That upload is the piece most setups skip, and it is the only thing that makes this channel arguable in front of a finance director. What we do not do: We do not launch on an availability source refreshing less often than every four hours for fashion, grocery or anything on promotion. We do not take accounts where head office cannot get a stock file out of franchise stores, because the verification audit will find exactly those stores and the whole programme pays for it. We do not manage the Shopping or Performance Max campaigns sitting alongside this unless that is agreed separately, and we do not write your product titles. Who this is not for: Single-location retailers get very little here and should spend the money on their Business Profile instead. Retailers whose in-store range differs from the online catalogue by more than a third need a catalogue reconciliation before any of this is worth starting, and we will say so in week one rather than take the fee.