Good design prevents bad decisions.

Design elements:

• Primary metric (one only)
• Secondary metrics to watch
• Control vs variant definition
• Audience segmentation
• Sample size calculation
• Runtime estimate

Document your hypothesis before seeing results.