Fraud detection
Same receipt. Rotated. Edited. Uploaded again. Yes, it happens.
That's why fraud checks have to run before winners are drawn, not after. Recefly checks every entry automatically, the moment it arrives.
Definition
How does fraud detection work in a promotional campaign?
Recefly checks every entry on two levels. Fixed checks catch what's certain: the same receipt image uploaded again (exact match), the same receipt cropped, rotated, or re-photographed (visual similarity matching), entry limits, and your campaign rules. A second level looks at each entry in context, gives it a risk score with the reasons, and recommends accept, reject, or review. Both run on every single entry, automatically.
Duplicates a crop won't hide
Every receipt image is fingerprinted twice. One check catches the exact same photo uploaded again. The other catches the same receipt cropped, rotated, brightened, or photographed a second time.
- Exact match for identical uploads
- Visual similarity matching for near-duplicates and edits
- Checked across the whole campaign, not just per participant
A risk score on every entry
Beyond fixed rules, each entry is assessed in context and gets a risk score, the reasons behind it, and a recommended action: accept, reject, or review.
- Each check shown with its pass or fail result
- A confidence level on every recommendation
- Risk thresholds you set per campaign
Your rules, applied every time
Set the limits that fit your mechanic. The platform applies them the same way, around the clock, without reviewer fatigue.
- Entry limits per person, per day, per week
- Automatic rejection for clear violations
- Changes to the data read from a receipt are tracked and flagged
Every decision explained
Every flag, score, and override is recorded. When a participant disputes a rejection, or a client asks why an entry won, you have the full story.
- A log of status changes and manual overrides
- A review queue with the evidence side by side
- Fraud patterns flagged for you, on the Pro (AI) package
What happens to a suspicious entry
- 1
Signals fire on arrival
A duplicate, a rule violation, or a high risk score flags the entry the moment it comes in.
- 2
The system recommends
Clear violations can be rejected automatically. Borderline cases go to review, with the evidence attached.
- 3
Your team decides with context
Reviewers see the receipt, the data read from it, the checks, and the participant's history in one place.
- 4
Everything is recorded
Accepted, rejected, or overridden: every decision has a reason and a log, for disputes and compliance.
Frequently asked questions
What teams ask before their first campaign on Recefly.
The same receipt entered again (identical or visually altered), edited receipt data, entries over the participation limits, receipts outside the campaign rules, and suspicious patterns in how people enter.
Recefly is built to avoid that. Clear violations are rejected automatically, but borderline cases go to a person instead of being guessed. Every automatic decision shows its reasons, and reviewers can override it with a logged justification.
Yes. Each campaign sets its own risk threshold, entry limits, and automatic-rejection policy. A campaign with a car as the prize can run stricter checks than a low-stakes sampling promotion.
It depends heavily on prize value and mechanic, but duplicate and edited receipts show up in almost every campaign whose prizes are worth the effort. Patterns range from re-uploading and cropping the same photo, to edited or generated receipts, to several accounts run by one person. We don't publish figures from client campaigns. In the demo we'll show you which checks run and what the rejection reasons look like in the dashboard.
Next campaign coming up? Don't start it from zero.
Tell us the mechanic. We'll show you how it runs on Recefly and what you'd reuse next time. Reply within 24 hours.