The Company That Reads the Return Line

The return counter is the one place in a store where the deal runs backwards. Money leaves the register, product comes back, and someone has to decide in a few seconds whether the person holding the receipt is a loyal customer or working an angle. For more than 60 of the 100 largest retailers in the United States, that decision is not made by the associate at the desk. It is made by Appriss Retail.

Appriss Retail is an enterprise software company built around a single, unglamorous question: is this return legitimate? Its models sit behind the point of sale, pull from more than twenty years of transaction history, and return a recommendation - approve, warn, or decline - before the customer has finished sliding the box across the counter. The company says its decisions land with 99.99% accuracy, and that its systems touch roughly 40% of all U.S. retail transactions. Most American shoppers have been scored by it and never knew the name.

01 / What It DoesThe math of trust, at the counter

Retailers live inside a genuine dilemma. Tighten the return policy and you turn away the loyal customers who buy three sizes and send two back. Loosen it and you invite the people who wear a dress once, "wardrobe" it, and return it - or who run refunds against receipts that were never theirs. Appriss Retail's pitch is that you do not have to choose. Its platform scores the behavior behind each return rather than the person, so a first-time shopper and a serial abuser can be told apart in real time.

That distinction - behavior, not identity - is the heart of the product. Appriss Retail says it can recognize roughly 250 million shoppers by their return patterns without needing a name attached to them. The counter associate does not see a dossier. They see a recommendation, and the model carries the memory.

02 / The ProblemA $796 billion number nobody was measuring

In late 2025 the company published what it called an industry-first Total Retail Loss Benchmark Report, drawn from hundreds of retailers and more than a thousand consumers. The headline figure was blunt: $796 billion in total retail loss for the year. The more useful finding was where it hid. Returns abuse - people exploiting lenient policies rather than committing outright fraud - cost retailers about $86 billion, roughly six times the size of return fraud itself. Only 21% of total loss, the report argued, is actually preventable.

Publishing the number is itself a strategy. If the industry accepts loss as a fixed cost of doing business, there is nothing to sell. Measure it, break it into fraud, abuse, shrink and process error, and suddenly there is a line item a software company can attack. The benchmark report, a public loss calculator, a newsletter called The Takeback and an executive community are all built to make the same argument - the problem is bigger than you think, and most of it is hiding in the returns.

03 / Products & ServicesEngage, Secure, Incident

The platform underneath is called Appriss Intelligence, and it feeds three product lines. Appriss Engage handles the moment of the return itself - in-store and online return, claim and appeasement authorization, plus an Order Protection layer that flags high-risk transactions at checkout. Appriss Secure is exception-based reporting: it surfaces suspicious patterns across transactions, inventory and employee activity, the kind of quiet leakage that never reaches a return counter, and the company says it typically returns 6x its cost. Appriss Incident rounds it out with case and audit management for shoplifting, organized retail crime and store safety.

In 2025 the company added Sidekick, an AI collaborator that lets loss-prevention teams question their own data in plain language and act on what comes back. A year earlier it had pushed an Exception Analytics app onto Shopify, a signal that the same models built for the top 100 chains can be sized down for merchants who will never have a 150,000-store footprint.

04 / Business ModelPriced against the loss it removes

Appriss Retail is B2B enterprise SaaS in the most literal sense - it sells subscriptions to very large retailers, and the contract is justified by the fraud, abuse and shrink the software takes off the books. That is why the numbers it markets are return-on-investment figures rather than seat counts: 6x on Secure, up to 10x on average. When the pitch is "we will hand you back more than you pay us," the sales conversation happens with a CFO, not just an asset-protection director. Winn-Dixie's finance chief is among the named customers who have vouched for it publicly.

The customer list runs across grocery, specialty, luxury and general merchandise, and the company puts its footprint at roughly 150,000 store locations in 45 countries. Those are not businesses that swap out infrastructure lightly. A retailer that wires Appriss Retail into its point of sale, its e-commerce checkout and its case-management workflow is not shopping for a replacement next quarter, which is part of why a decades-old data asset keeps compounding: every additional retailer makes the behavioral models sharper for all the others.

05 / Where It Came FromSpun out, then bought

The retail data science goes back more than twenty years inside the broader Appriss data business. In 2023, backers Clearlake Capital and Insight Partners carved the retail unit out as a standalone company focused entirely on total retail loss. Two years later, in March 2025, the middle-market private equity firm Gemspring Capital acquired it; terms were not disclosed. The company framed the deal as fuel for new growth, and by the end of the year it had crossed $1 trillion in annual transactions protected and expanded its executive bench, bringing in Sarah Cascone - who had helped scale the martech firm Bluecore - as chief marketing officer.

06 / The CompetitionNarrow on purpose

On the online fraud side, Appriss Retail brushes up against names like Signifyd, Riskified, Forter and Sift. On the store-analytics side it competes with exception-reporting vendors such as Agilence and the workforce-and-loss tools inside larger platforms. Its edge is not breadth but the opposite - it does one narrow thing, the return decision, with a data set almost nobody can match, built from decades of scoring returns across a huge share of American checkouts. The moat is the history.

Expertise here is less about any single algorithm than about the accumulated labeling that comes from watching hundreds of retailers process returns for twenty years. A new entrant can buy the same machine-learning tools; it cannot buy the record of which return patterns quietly preceded a chargeback, an employee-fraud case or an organized-crime ring. That is the asset Gemspring Capital bought in 2025, and it is the reason the company leans so hard on research - the benchmark report, the loss calculator, the executive community - to remind the market that the data underneath is the point.

07 / Where It FitsThe invisible decision layer

There is a version of retail technology that gets written about - the checkout-free stores, the delivery drones, the consumer apps. Appriss Retail is the other kind: infrastructure that most shoppers will never see and most retailers would rather not advertise, because admitting you score returns is an awkward thing to put on a sign. Its position in the market is that of a utility. When something like 40% of the country's transactions pass through your models and more than a trillion dollars of value leans on your yes-or-no, you do not need a consumer brand. You need to keep being right.

The 2025 benchmark report put the stakes plainly. Loss is not shrinking, returns are getting more complicated as online and store channels blur, and the honest customer and the abuser increasingly walk up to the same counter. Appriss Retail's whole business is standing between them and getting the call right - fast enough that neither one notices the software was ever there.