One of the significant pain points customers interviewed repeatedly raised was that legacy and black-box solutions had high rates of false positives. To address this, TrafficGuard has by design included a vast number of controls, legitimate baselines and control group(s) of data, supporting the continued minimisation of false positives/negatives and the ongoing analysis of accuracy.

Many of our customers have cited false-positive prevention as a critical reason for choosing TrafficGuard over our competitors.

Unlike single-point solutions, TrafficGuards multi-point full-funnel approach allows measurement ​​across the whole advertising journey. Additional data points, behaviour and corroboration of verification supports the minimisation of false positives.

TrafficGuard provides unparalleled transparency to support invalidation. With transparent reasons for each invalidated transaction aligned with MRC/IAB standards and empirical support via best-in-class self-serve reporting.

Advertising specific rules and thresholds that reliably identify invalid traffic. Rules are the first line of defence that helps to eliminate KNOWN invalid traffic.

TrafficGuard analyses transactions as they occur from an individual machine or device over time; from a supply source over time; across all campaigns; and as a part of each campaign. We evaluate and compare all traffic against norms and explore statistical outliers, to minimise false positives and find early indicators of unknown invalid traffic (false negatives).

TrafficGuard creates a session score on each individual based on behaviour such as the quality and frequency of engagement and whether the traffic abandons the session quickly or goes on to convert.

By measuring and verifying billions of ad engagements across the world’s top ad networks, apps, and advertiser websites, TrafficGuard has one of the best vantage points into the reputation of IPs, domains, apps and devices.

Every transaction is analysed and stored in our proprietary Reputation IQ engine based on its quality and propensity to convert. That is, the likelihood that the click is from a human with the intention of engaging with the advertiser in some way, whether it is to buy a product, fill out a form, or subscribe to a service.

TrafficGuard has developed controls that consider reputation for minimising false positives/negatives.

TrafficGuard has over 200 sophisticated algorithms using an array of independent methods. Traffic is tested by many individual rules asynchronously. While in most cases, for clarity of counting, only a single reason is reported, invalidation of traffic occurs from multiple rules.

Separate rules collectively identifying traffic as invalid is indirect evidence of the performance and correctness of invalidation.

TrafficPath is one of the routes where we send invalid traffic to interrogate the device further and provide a user challenge to test for false positives. Any user who passes the challenge is redirected to the app store or landing page not to damage the real-user experience. The solved challenges (false-positive) rate is fed back to our detection algorithms in real-time for optimisation.

Data quality and completeness are critical elements of invalid traffic detection and filtration, often overlooked by other solutions. Missing or incorrectly populated tracking parameters means that some traffic is never eligible for measurement or invalidation. Often other vendors erroneously invalidate traffic with missing information. However, TrafficGuard ensures this traffic is pre-filtered and reported unmeasurable.

Our team of dedicated data scientists and engineers live and breathe fraud prevention.

Our team conducts periodic risk assessments of measurement, including assessing the sufficiency of the internal controls, continued relevance and effectiveness of IVT procedures and the ongoing analyses of accuracy. When our team identifies undetected IVT, they use the data to improve our automated detection to improve the real-time prevention of similar threats.