Receipt validation becomes the bottleneck nobody budgets for

Your campaign launched on time.
The creative was approved.

The platform went live. The first claims started arriving. And then, quietly, the operating model started breaking.

Nobody put that in the brief.

We keep seeing the same pattern. Receipt validation gets scoped as a back-end process. It gets three lines in the SOW. It gets mentioned once in the kick-off meeting.

And then, somewhere around week two of a live campaign, it becomes the only thing anyone is talking about.

The belief we're challenging: validation is admin. It's not. It's where campaign scale becomes operationally real.

The volume assumption

Every campaign brief has a participation forecast. Usually it’s built on awareness reach, prize appeal, and whatever the last campaign did. Sometimes it’s built on gut feel and a growth target someone circled in a slide.

What the forecast almost never includes is what happens to validation workload when that number lands.

Here’s what we’ve noticed. A campaign designed for 30,000 claims has a different operational reality at 30,000 than at 80,000. Not proportionally different. Categorically different. Because the volume doesn’t just scale the expected work. It surfaces the edge cases. The receipts that don’t parse. The SKUs that weren’t on the original list. The claimants who uploaded a screenshot of a screenshot. The entries that look identical across three different email addresses.

OCR handles the easy receipts. The difficult ones don’t disappear. They queue.

And that queue is where the operating cost compounds.

What manual review actually costs

Most campaigns have a manual review threshold. Below a certain OCR confidence score, a human looks at it. Simple in theory.

We’ve seen 15% of claims require manual review. We've seen 40%. Both numbers came from campaigns that were briefed, approved, and launched with no manual review resource in the plan.

The cost isn’t just the reviewer hours. It’s the delay it creates in claim status updates. It’s the customer service contacts asking why their claim hasn’t been processed. It’s the escalations from brand managers who are watching a public-facing promotion appear to stall. It’s the reporting that can’t close because a chunk of the claim file is sitting in a queue.

Manual review is not the fallback. It's a product experience decision. Either you design it deliberately or it designs itself under pressure.

The SKU problem nobody briefs

A qualifying product list sounds like a solved problem. You know which products are in the promotion. You add them to the validation logic. Done.

Except the list changes. New pack sizes arrive. A retailer substitutes a product. A SKU that wasn’t on the original list turns up on a receipt because someone in procurement approved it after the brief was locked.

we keep seeing valid claimants get rejected because the SKU logic didn’t keep up with the campaign. Andwe keep seeing that rejection become a customer support issue, a complaints spike, and eventually a manual override workload that costs more than the original build.

The qualifying product list is not a launch document. It’s a live operational dependency. If nobody owns it after go-live, it drifts. And when it drifts, the validation system enforces a version of the campaign that no longer matches what consumers were sold at the shelf.

Fraud detection and the real cost of ambiguity

Duplicate detection is usually described as a rules problem. One person, one entry. Simple.

Duplicate detection in practice is a judgement problem. The same person rarely uses the same email twice. They use variations. They use their partner’s account. They submit receipts from the same shopping trip. They’re sometimes the same household and sometimes four different people trying their luck.

Blocking every ambiguous case is cheap operationally and expensive commercially. You’ll reject real customers. You’ll create support escalations. You’ll get complaints. You’ll potentially get a TIO issue if the promotion involves specific categories.

Not blocking anything means you’re approving fraud. That has its own downstream cost.

The trade-off has to be designed. Before launch. With actual decision rules. Not managed reactively by whoever is on review shift when the volume arrives.

Where validation planning actually belongs

Validation planning is not a technical question. It's a commercial planning question.

How you handle OCR confidence thresholds affects fulfilment cost. How you handle the qualifying product list affects rejection rates and support volume. How you handle manual review design affects turnaround time and customer experience. How you handle duplicate detection affects fraud exposure and brand reputation.

Every one of those decisions has a commercial implication. And every one of them is typically made after launch, under pressure, by whoever is watching the queue fill up.

You can brief a campaign in four hours. Validation logic takes four hours to brief properly too. The difference is that the campaign brief happens in the room with the marketing director. Validation logic happens in an email chain three days before go-live.

That imbalance is where the bottleneck starts.

What this means for your next campaign

• You don’t need to solve everything upfront. You need to make the decisions that are expensive to reverse.

• OCR confidence thresholds: what score triggers manual review, and who is resourced to clear it?

• Qualifying SKU list: who owns updates to it during the live period, and what's the process when something changes?

• Duplicate logic: what's the rule, what are the edge cases, who makes the call when it’s ambiguous?

• Exception handling: what happens when validation can’t return a clean decision?

• If those four questions don’t have answers in your campaign brief, they’ll find answers anyway. Just not the ones you would have chosen.

Where does validation planning sit in your campaign brief — before launch, or after the volume arrives?

Roilti is a claims validation, reward fulfilment and payout platform for complex promotional campaigns, class actions and financial remediation. We build the operating infrastructure that makes large-scale claims programmes run cleanly — from receipt validation and fraud detection to payment disbursement and audit trail.

Talk to the Roilti team about your next campaign