You launch a feature and usage is lower than expected. How do you investigate?
Model answer
Model answerLow usage can mean the feature isn't valuable, isn't discoverable, or isn't working correctly. I'd start with the discovery funnel: how many eligible users saw the feature entry point? Of those, how many tried it? Of those, how many completed it? The drop-off tells you where the problem is. If discovery is low, it's a placement problem. If completion is low, it's a UX or value problem. I'd then run 5 qualitative sessions to understand the 'why' behind the data.
Framework · Metrics Quality
- AAwareness
How many users saw the feature entry point? Low awareness = placement problem - CClick-through
Of those who saw it, how many tried it? Low CTR = value proposition problem - TTask completion
Of those who tried, how many finished? Low completion = UX or friction problem - SSatisfaction
Of those who completed, how many came back? Low return = value problem