To validate these new features, we set up a series of tests that focused on optimizing both the
learning and
exploitation phases of CPA campaigns:
- Feature A (a traffic filtering tool) disabled underperforming traffic segments, allowing the platform to focus resources on higher-converting opportunities.
- Feature B (a traffic prioritization tool) prioritized incoming traffic based on historical performance, ensuring that the best-performing segments were explored first.
We implemented the following testing phases:
- Initial Learning Phase: The tools worked together to create a blacklist of underperforming traffic segments and a whitelist of high-potential segments for further exploration.
- Exploitation Phase: The platform then focused on traffic segments most likely to convert, while further filtering out any segments that did not perform well.
These tests were run with several
key clients, including a US-based performance marketing company and another long-standing client focused on high-volume CPA campaigns. Both clients were keen to trial these features to improve conversion rates and better manage their campaign spending.