Receipt verification pricing: what counts?
Compare receipt verification pricing by upload, extraction, or verdict. Learn how retakes, duplicates, and review affect the bill.
Receipt verification pricing models can count uploads or completed checks. Imagine two vendors quoting the same price per receipt. One counts every upload, including blurry images it asks the customer to retake. The other counts submissions that end with a decision. The unit price looks the same. On a campaign with unusable uploads, the bills will differ.
When comparing receipt scanning pricing, write down what each rate counts. “Per receipt” is not a billing rule. It might mean a file, a page, or an extraction: fields read from evidence. It could mean a completed check against your campaign rules.
For a budget owner, that matters more than another decimal place in the quote.
Compare receipt verification billing units
The table compares possible ways to charge, not specific vendors. Contracts can combine them, so check the actual terms.
| Unit | What counts | What the model rewards | Cost to check outside the headline rate |
|---|---|---|---|
| Upload or page | Each file sent or page processed, as the contract defines it | More input volume | Unusable photos, retakes, extra pages |
| Extraction | A call that returns fields from evidence | Returning data in a set format | Decision rules, duplicate checks, review tools, and upkeep |
| Seat or licence | Users or access for an agreed period | Selling access and capacity | Extra reviewers during peaks, usage limits, unused capacity |
| Verdict | A submission that ends the checking process with a decision | Completing an evidence check | Which outcomes count, which refusals are excluded, reviewer labor |
An extraction price can suit a team that already owns the decision rules. You may only need seller’s name, date, total, and line items sent to an existing system. Buying that part is reasonable. Just add the work your team will keep when comparing its rate with a full verification workflow.
A licence can also suit stable volume and a fixed review team. How well you can plan costs depends on what it includes. A peak campaign is a poor time to find that temporary reviewers need separate seats.
The build-or-buy guide maps those duties. The guide to reading receipts versus checking claims explains why fields alone do not settle a claim.
What Steve counts as a verdict
A verdict is a decision on the evidence. Steve’s public pricing counts a submission that leaves the checking process with one. It may be approved, flagged for review, blocked, or rejected by your rules. The rate depends on the evidence class, or type of evidence. Enterprise terms are agreed separately.
A receipt verification verdict counts the check, rather than the reward it may trigger. A valid receipt can still fail the next system’s campaign rules for earning rewards. Finding that someone submitted a receipt before is a verdict and is billed too. The duplicate check did work even though no new purchase was accepted.
Evidence refused by the capture quality gate is not billed. This gate checks whether evidence is usable. Processing failures are not billed either. A refusal means the evidence could not be assessed. A blocked verdict means the process assessed it and reached a decision. Give these separate rows in any comparison.
The evidence class sets what one unit covers:
- Receipts and documents: one image or PDF that receives a verdict.
- Photos: one checked photo. A store visit with five photos is five verdicts. Photos that cannot be matched to a required store or task are not billed.
- Recordings: confirm the supported call length and billing terms for your workflow before budgeting. See call compliance and the current pricing terms. Recordings too poor to turn into text, silence, and dead air are not billed.
- Extraction only: the submission is resolved when its fields are delivered. Check the current pricing terms for that scope.
A flagged submission later cleared by your reviewer stays one verdict. The review queue is included. So are seats, workflows, rules, the API, and the Open Loyalty connection. Your team’s review time still belongs in your operating budget.
Use the pricing page for list rates and options based on committed volume. Keep the evidence classes separate when estimating a mixed program.
How retakes and duplicates affect your bill
If a contract bills every upload, a refused photo can still earn the vendor money. Stopping an unreadable image early may not reduce that bill. A retake may add another charge, depending on the contract.
Under Steve’s public verdict rules, an unusable capture refused at the quality gate is free. This gives the service a reason to spot bad input before processing. It also makes retake billing easier to explain. The refused attempt is free. The new one counts if it produces a verdict.
Neither structure proves a vendor makes accurate decisions. A cheap bill with wrong approvals or needless review is still costly to operate. Test the decisions and the capture experience alongside the bill.
Seat pricing creates a different tradeoff. Adding reviewers during a launch can raise licence costs just when the queue needs help. Unlimited seats remove that charge. They do not remove the cost of staffing the queue.
Questions to ask about a receipt verification quote
Ask for written answers using your own submission examples. A broad promise of “successful processing” leaves too much room for debate.
- What is one billable unit? Ask how pages, multi-photo visits, and long recordings are counted. Ask which outcomes finish the unit.
- Are retakes billed? Separate an unreadable capture refused before processing from evidence assessed and found not to qualify.
- Are duplicates billed? Separate a repeated purchase check from a technical retry of the same request.
- Is review extra? Separate queue access, reviewer seats, and human-review labor. Ask whether clearing a flag creates another charge.
- Are seats metered? Include admins, API users, and temporary reviewers in the answer.
- What changes when a rule changes? Ask about workflow setup, upkeep, and any effect on the agreed rate.
- Is the pilot priced differently? Record its scope, duration, exclusions, and what carries into production.
- What commitment and exit terms apply? Ask what happens if volume is lower than expected or the test fails.
Use the evaluation scorecard to assess the workflow beside these answers about terms and costs. Use the verification cost model to estimate receipt verification cost per receipt. Include review time and setup work in monthly sums.
Agree on pilot limits and exit terms
One buyer told us high scan volume was a good sign. It meant repeat purchases. Usage growth was welcome. An unclear commitment was harder to accept.
A pilot with set limits lets you measure capture refusals, review rate, and handling time. Do this before committing to production volume. Steve’s public terms describe short fixed-term pilots with no volume commitment. Agree the scope and exit in writing. Do not assume a trial and production have the same terms.
Bring your current quote or invoice and a sample month of submissions to a discovery call. We will price it per verdict alongside the current unit, with exclusions visible.
Count the decisions. Make every upload charge explain itself.
