Great experiments start with great hypotheses.

Hypothesis format:

If we [change], then [metric] will [improve] because [reason].

Backlog prioritization:

• Impact: How much could this move the needle?
• Confidence: How sure are we it will work?
• Ease: How hard is it to implement?

Prioritize high impact, high confidence, low effort first.