TICKET TO BE EDITED

As a product manager for credibility signals, I want to lessen the burden of manual testing, so that I can better craft accurate credibility signals. As our signals mature and grow in scope, eventually being the only human in the loop will result in a massive blocker and could create errors.

While removing a human in the loop for checking would be a mistake, template usage at least on enwiki is consistent and can help test signals that employ them. Some templates are used more than others, but we don’t really understand usage with the granularity we should to make specific decisions on what templates to depend on and when.

I want to create testing sets that do not decay and are reflective of the community's usage of templates by looking back on and tracking template usage going forward, refreshing the set data over a certain time period.

This way, I can better understand our communities and more accurately derive contexts about template usage across projects and simultaneously and have an accurate data set to test against that can cut down manually checking results.