In some cases, fraudulent traffic sources may attempt to avoid discovery by purposefully not loading FraudSensor’s detection tags. This process, known as tag evasion, affects our ability to collect traffic signals and flag invalid traffic.
HUMAN can identify tag evasion by using Impression Sync to compare the number of times a FraudSensor tag was loaded versus the number of requests triggered by your backend system. If there’s a discrepancy between these values, we can then determine that users are evading our tags.
How Impression Sync works
Every time a FraudSensor tag is loaded on an end user’s device (like a computer or cell phone), HUMAN can see that event in real time. If you’ve configured Impression Sync correctly, your backend system should also trigger a separate request informing HUMAN that our FraudSensor tag was delivered—regardless of whether our tag detected that impression or not.
The comparison between these two values is how we identify tag evasion. For example, if our FraudSensor tag only recorded 80 total impressions, but the secondary Impression Sync request on your server recorded 100 total impressions, we know our tag was evaded 20 times.
The data we collect through Impression Sync will also yield a limited number of traffic signals, which provides us with contingency data and helps us prevent future instances of tag evasion.
Get started with Impression Sync
The primary method for detecting tag evasion via Impression Sync is the server-to-sever (S2S) approach. S2S sync requires you to make a postback call to HUMAN every time you record an impression in a location where HUMAN’s detection tag is deployed, regardless of whether our tag detected that impression or not.
Every time you’re notified of an impression event, your server should make a call to the provided postback URL with information about that event. Similar to our detection tag, this information is passed via URL parameters. Any fields passed via S2S sync come from the same FraudSensor tag that was delivered with the ad, with the exception of the IP, User Agent, and Timestamp parameters (which are typically collected from the headers of the FraudSensor tag rather than passed through it).
If you’d like to use this Impression Sync integration, please reach out to your HUMAN representative. For a full list of parameters, consult our FraudSensor Parameters guide.