Who This Helps
You're a team lead who wants to scale a repeatable analytics routine. You already track a bunch of numbers, but picking the next experiment feels like a coin flip. This is for you.
Mini Case
Meet Maya. She leads a product team that tracks 20 metrics every week. Every Monday, her team debates which experiment to run next. Last quarter, they ran 4 experiments, but only 1 moved the needle. Maya realized she needed a better way to prioritize.
She took the Metrics & Dashboards Basics course and built a weekly scoreboard. Now, she ranks experiments by potential impact. Her team runs 3 experiments per quarter, and 2 of them show real results. That's a 12% lift in their North Star metric in just 90 days.
Do This Now (5 Steps)
- Pick your North Star metric. Choose one primary metric that matters most. Maya picked "weekly active users."
- Define 3 supporting metrics. These help you understand why the North Star moves. Maya chose "sign-up rate," "session duration," and "feature adoption."
- Set realistic targets. For each metric, write a target number for next week. Maya set a target of 5% growth in sign-ups.
- Build a weekly scoreboard. List your metrics, their current values, and targets. Update it every Monday. This becomes your experiment launchpad.
- Rank your next experiment. Score each idea by how much it could move your North Star metric. Pick the one with the highest score. Run it.
Avoid These Traps
- Tracking too many metrics. Stick to 4-5 key numbers. More than that, and you'll drown in noise.
- Skipping targets. Without a target, you can't tell if you're winning or losing.
- Running experiments without a scoreboard. You'll waste time on low-impact ideas.
- Changing metrics every week. Consistency helps you spot real trends.
- Forgetting to celebrate small wins. A 2% lift is still progress.
Your Win by Friday
By Friday, you'll have a clear answer to "What experiment should we run next?" Your team will stop debating and start executing. You'll focus effort on the highest-impact move, every single week.