TrafficGuard Support avatar
Written by TrafficGuard Support
Updated over a week ago

Game changing mobile ad fraud protection

TrafficGuard is the only solution blocking fraud before it hits your mobile advertising budget.

How TrafficGuard Works

Where conventional fraud tools operate at or after app install attribution, TrafficGuard boasts three layers of formidable fraud protection. By blocking at the click, attribution and post-attribution levels, TrafficGuard safeguards mobile advertising budgets from known and unknown tactics.

Click Level

TrafficGuard blocks all General Invalid Traffic (GIVT) and lots of Sophisticated Invalid Traffic (SIVT) at the click. Blacklists intercept suspicious IPs, bots, server traffic, and crawlers. TrafficGuard identifies SIVT through analysis of click behaviours such as multiple campaigns being clicked at the same time by the same user; and click characteristics, such as traffic arriving via VPNs.

TrafficGuard blocks over 90% of the IVT it intercepts at the click, meaning:

  • Reduced measurement and analytics costs because IVT is removed before that traffic goes to 3rd parties

  • Reduced opportunity for misattribution (AKA organics poaching) as these clicks don’t get to the MMP

  • Faster, more confident campaign optimisation because invalid clicks don’t skew performance data

Attribution Level

TrafficGuard blocks SIVT before install attribution occurs. When TrafficGuard receives an attribution from a measurement platform, it uses a variety of click and installs characteristics to validate that attribution. Distribution modelling is used at the attribution level to ensure organic installs are not mistakenly attributed to paid, fraudulent sources.

Post Attribution Level

Emerging fraud tactics, making up a small portion of overall IVT, are commonly diagnosed post-attribution.

Analysis of post-install activity from 1000s of campaigns every minute, gives TrafficGuard a massive data set and broad perspective to detect fraud at the post-attribution level. Sophisticated machine learning analyses app engagement for abnormal behaviours including events happening on mass, or irrational engagements such as illegitimate in-app purchases going through MMPs.

Always Learning

Ad fraud is constantly mutating to avoid detection. TrafficGuard, and the boffins that train it, utilise machine learning fueled by big data to find and validate new fraud tactics as they emerge.

Each detection level feeds data back into TrafficGuard’s machine-learning algorithms to help it get even stronger. Subsequent traffic from sources of SIVT detected at attribution and post-attribution is flagged at the click so that low-quality traffic is optimised out.

Did this answer your question?