In order to see how much attention the links in the policy popover get, we want to calculate a CTR for each of them.
Acceptance criteria
The next links have a CTR instrumentation and the rates are shown in one of Superset/Growthbook (see open question)
- offensive
- misleading
- promotional
- not an organisation
- real name
Open questions:
- Are these links worth instrumenting within an experiment given they do not exist for the control group?
- If so, should the outcome be part of TestKitchen's automated analysis report? Or we're using GrowthBook for all the metrics in this experiment?
- If not, should we capture the CTRs as a long-lived instrument, do manual analysis?
