Tying online behaviour to in-store purchases, making loyalty data usable in paid media, and keeping product feeds and order data in step across every channel.
Retail has more customer data than almost any other sector and the hardest time making one coherent thing out of it. The same shopper browses on mobile, buys in store with a loyalty card, returns online and complains on WhatsApp, and in most estates those are four unrelated records. Everything downstream, from personalisation to paid media efficiency, is limited by how well that gets resolved.
Retail also carries the heaviest operational data load in martech: product feeds, stock, price changes, promotions and order status all have to stay in step across the website, the app, email, the ad platforms and the marketplaces. When those drift, you advertise products you cannot ship and personalise around prices that no longer exist.
It is also the sector where the stack gets judged in public. A tracking gap that costs a little in March costs a great deal in the ten days before Christmas, and a feed that drifts on Black Friday morning advertises stock you do not have to the largest audience of the year. We design for peak, not for the average Tuesday.
Usually the people who own a number and are being asked to hit it with data that does not join up.
The majority of retail revenue still completes in store, where the tracking cannot see it. Without a loyalty or payment identifier tying that purchase back to the digital journey, your ad platforms optimise against a fraction of your actual sales, usually the least profitable fraction.
Most retailers have a rich loyalty database that never reaches the channels where it would earn money. Getting lifetime value, category affinity and lapse risk out of that system and into paid media suppression, lookalikes and lifecycle journeys is often the single largest uplift available.
Product, price and stock feeds fail silently far more often than anyone assumes. The symptom is not an error message, it is a slow decline in ad performance and a rise in complaints about out-of-stock items being promoted.
Heavy discounting makes attribution actively misleading. Without holdouts you cannot separate campaigns that drove incremental sales from campaigns that discounted sales you were going to make anyway.
A sale that is returned three weeks later is still a conversion as far as Meta and Google are concerned. In categories with high return rates the platforms are optimising towards customers who buy and send back, and nobody has told them. Feeding refunds and returns back as conversion adjustments is unglamorous work with an outsized effect on bidding.
You can tell a lot about where the money is from five answers.
The Martech & Adtech Health Check is a fixed-fee review of your tracking, consent, feeds and loyalty activation, run against how you actually trade. You get a prioritised plan you can act on before the next peak. Stack Teardown from £1,950, full Health Check from £6,500.
Usually through a loyalty identifier captured at the till, supplemented by card tokens or receipt-level matching where loyalty penetration is low. The realistic goal is not universal coverage. It is enough coverage to model the rest reliably. Retailers who get to 40-60% identified in-store transactions can usually infer the remainder well enough for both measurement and activation.
