Golden rule: Pick low-hanging fruits first
Three principles for prioritization:
- Business priority
- Estimated commercial benefit
- Ease of implementation
Things to consider when assigning priority:
- Ease of implementation
- Potential business benefit
- Which part of the funnel the experiment is focusing on. Later stages of funnel usually easier to optimize
- The scale of the changes. Avoid meek tweaking. Bigger change → potentially higher affect on the metric → easier to detect change in the test
- Consider weather the experiment supports a business objective
- Calculate what increase in CVR (or other metric) would do to revenue. If the test would detect MEI how the revenue will change?