How to Audit Your Campaigns for Invalid Traffic: A Step-by-Step Guide

In short: you can find the first signs of invalid traffic in the campaign reports you already have — check frequency, geo, hour-of-day, conversion deltas and placements — but only device-level verification proves whether a real person was behind the impression.

You don’t need special tools to start spotting invalid traffic. Your existing campaign reports contain the fingerprints. Here’s a self-audit any brand team can run with the platform access they already have.

1. Check frequency anomalies

Pull frequency distribution by user. A small cluster of “users” seeing your ad dozens or hundreds of times in a day is rarely human — it’s a bot, or a duplicated device ID being counted again and again as fresh reach.

2. Look for geo mismatches

Compare where you targeted with where impressions were actually served. Spend leaking to countries or regions you explicitly excluded is a classic sign of location spoofing or corrupted geo data.

3. Map hour-of-day patterns

Human audiences follow daily rhythms. Flat overnight delivery, or spikes at 2–4am, points to automated activity running on a schedule rather than an audience going about its day.

4. Compare conversion-rate deltas

Break results down by placement, publisher and device. A source delivering huge impression volume and near-zero conversions is buying you volume, not people. The funnel collapses because there was never anyone in it.

5. Read the placement report

Scan the actual domains and apps your ads ran on. Unfamiliar, generic or churn-heavy sites are the hallmark of made-for-advertising inventory — pages built to arbitrage ad spend rather than attract readers.

What this audit can’t show you

These five checks catch the obvious cases, but they all rely on aggregated platform data — the same data the platform grades itself on. They can’t tell you whether a given impression reached a device that was genuinely live with a real person behind it, or a dormant, recycled or machine-to-machine device that merely looks legitimate in a report.

That takes device-level verification: filtering to devices seen active with a real human in the target location within a recent window. The self-audit tells you that you have a problem. Verification tells you exactly how big it is — and, done pre-bid, stops it before you pay.

Make it a trend, not a snapshot

Run this audit quarterly and log the results rather than the readings. A single quarter can mislead: seasonality, a new publisher, one bad placement. A rising invalid-traffic rate across campaigns and placements is the real signal — and the point at which it’s worth moving from spotting the problem after the fact to preventing it before the spend.

If you’d rather see the number than estimate it, we’ll verify a live campaign for you.