Forensiq works to constantly stay ahead of new types of invalid traffic by performing periodic reviews of existing methodologies and adding new ways to identify IVT. The general update process closely follows the scientific method:
Inquiry: During this phase, Forensiq’s Data Science team conceptualizes new ways to identify invalid traffic. This knowledge-gathering phase allows our team to better understand a topic by tapping into research from various fields, as well as online guidelines and best practices shared by MRC, IAB, TAG, ABC, and other industry bodies.
Hypothesis development: During this phase, the Forensiq Data Science team forms a hypothesis based on the gathered knowledge or based on reasonable assumptions about the data.
Experiments: Forensiq’s team performs experiments to prove or disprove a hypothesis. These experiments can involve browser testing, statistical data analysis, and other methods of validation.
Edge cases, biases, and controls: Through experiments, Forensiq’s Data Science teams strive to map out potential edge cases or biases that may result in false positives and create controls for these edge cases.
Peer review: Forensiq techniques are peer-reviewed within the company to validate their accuracy and the accuracy of the implementation.
For any new IVT techniques, Forensiq has a backward-looking assessment process. Prior to the release of a change to IVT methodology, Forensiq performs at least a 14-day assessment of the impact of the change.
Existing techniques are reviewed using the process above on at least an annual basis.