Lukewarm take: Observing 5 developers from a reasonable sample try to use your developer tool product will produce more actionable insight than most A/B-tests.
Conversation
Replying to
You’re backed up by plenty of UXers with that one:
Can confirm. Max 10 user interviews/over-the-shoulders. After that, you're just hearing edge cases that acting on those initial findings may negate
2