Event Risk Explained

What it is

Event Risk is a scoring layer that sits between a conversion landing in your account and you actually paying commission on it. Every time a sale or lead is tracked, impact.com assesses how "trustworthy" that event looks — based on signals like the click that preceded it, the partner's history, repeated patterns, and other behavioural fingerprints — and assigns it a risk level. High-risk events get surfaced so a human can take a look before money changes hands.

Think of it like the fraud team at your bank. When you tap your card for a coffee on your usual street, nothing happens — the transaction sails through. But when that same card suddenly buys electronics in another country at 3am, the bank freezes it and asks "was this really you?" Event Risk does the same thing for affiliate conversions: routine, normal-looking sales pass quietly, while the odd-looking ones get held up at the gate for a second opinion. It isn't deciding guilt — it's raising a hand and saying "this one's worth a closer look."

Why it matters

Affiliate fraud is rarely dramatic. It's not usually one giant fake order — it's a slow drip of small, plausible-looking conversions that quietly siphon budget month after month. A partner stuffing cookies, a bot farm clicking links, a "coupon" partner intercepting checkout traffic they never earned: each individual event looks fine, but in aggregate they can eat a meaningful chunk of your payout before anyone notices the numbers don't add up.

The reason Event Risk matters is timing. Catching a bad conversion after you've paid is a clawback nightmare — awkward partner conversations, disputed invoices, money you may never recover. Catching it before you pay is just a click in a review queue. Every flagged event you reject is budget that goes back into rewarding partners who actually drive incremental sales, instead of subsidising the ones gaming the system. For a programme manager, that's the difference between defending your budget and watching it leak.

How it works

From your side as the advertiser (the brand running the programme), the flow looks like this:

1. A conversion is tracked. A partner drives a sale or lead, and impact.com records the event along with the click trail, partner, device, and timing data behind it.
2. The event is scored. Before the conversion is locked in for payment, the platform evaluates its risk signals and assigns it a level — low, medium, or high — reflecting how likely it is to be fraudulent or low quality.
3. Risky events are flagged. Conversions that cross your risk threshold are held back from automatic approval and routed into a review state instead of sliding straight toward payout.
4. You set the rules. Decide what should happen at each risk level — auto-approve the clean ones, hold the borderline ones for manual review, and optionally auto-reject the worst offenders.
5. You review the queue. Open the flagged conversions and look at the detail: what triggered the flag, which partner it came from, and whether the pattern matches a known issue or is just an unusual-but-legitimate sale.
6. You approve or reject. Clear the genuine conversions through to payment and reject the ones that don't hold up — keeping that commission in your account.
7. You feed back patterns. Use what you learn to tighten partner terms, pause repeat offenders, or adjust your thresholds so the system gets sharper at separating good traffic from bad.

Common mistakes

• Treating every flag as guilt. A high risk score is a prompt to investigate, not a verdict. Legitimate sales sometimes look odd — a deal-hunting customer, an unusual device, a burst of traffic from a viral post. Reject blindly and you'll punish good partners and sour relationships. Always review before you reject.
• Ignoring the queue entirely. The opposite failure: flags pile up unreviewed, defaults quietly approve everything, and the safety net does nothing. Event Risk only protects you if someone actually works the queue on a regular cadence.
• Setting thresholds and never revisiting them. Fraud tactics evolve, and so does your partner mix. Thresholds you set at launch may be too loose a year later. Treat them as a dial you adjust, not a switch you flip once.
• Not documenting your decisions. When you reject a conversion, note why. Without a record, partner disputes become your word against theirs, and you can't spot the repeat offenders forming a pattern over months.

Reporting tips

The dashboard is where Event Risk turns from a vague safety feature into hard intelligence about your programme's health. Start by tracking the share of conversions flagged each week — a stable baseline is normal, but a sudden spike usually means a single partner or a new traffic source has started misbehaving, and that's your cue to dig in.

Break flagged events down by partner: a handful of partners almost always account for the bulk of risky conversions, and isolating them tells you exactly where to focus. Watch your rejection rate too — if you're rejecting most of what gets flagged, your thresholds may be too loose and letting obvious junk through; if you're approving nearly everything, they may be too tight and creating busywork. Finally, keep an eye on the gap between conversions tracked and conversions approved over time. A widening gap is an early warning that quality is slipping before it shows up in your margins.

When to use / when not to use

Related guides

← Back to impact.com hub

Frequently asked questions