← Back to blog

Junior Analyst · Metrics & Dashboards Basics

Diagnose a KPI Drop Like a Junior Analyst

Find the root cause of a metric drop in one focused session. Ship clean analysis with clear recommendations.

Who This Helps

This is for junior analysts who get a sudden KPI drop and need to figure out what happened fast. You want to ship a clean analysis with clear recommendations, not a messy report full of guesses. The Metrics & Dashboards Basics program is built for exactly this moment.

Mini Case

Imagine you support a subscription service. Last week, the weekly active users dropped 12% in 7 days. Your boss wants a root cause by Friday. You have a dashboard with 20 numbers, but you need one primary metric and a clear story. In the Metrics & Dashboards Basics course, you learn to pick a North Star metric and build a metric tree. Let's apply that now.

Do This Now (5 Steps)

  1. Pick your primary metric. Don't track 20 numbers. Choose one that matters most. For our case, that's weekly active users.
  1. Define it clearly. Write down exactly what counts as an active user. Include the time window. For example, "a user who logs in at least once in 7 days."
  1. Build a metric tree. List 3 supporting metrics that feed into your primary. For weekly active users, those could be new sign-ups, returning users, and churned users.
  1. Set realistic targets. Use last month's average as a baseline. If weekly active users were 10,000 last month, your target is 10,000. The drop to 8,800 is a 12% problem.
  1. Run a focused session. Spend 45 minutes checking each supporting metric. Look for the one that changed most. In our case, new sign-ups dropped 30% while returning users stayed flat. That's your root cause.

Avoid These Traps

  • Don't chase every number. Stick to your metric tree. If you look at 20 metrics, you'll find noise, not signal.
  • Don't skip defining your metric. Vague definitions lead to wrong conclusions.
  • Don't forget to check the data source. A bug in the tracking can look like a real drop.
  • Don't present raw numbers without context. Always compare to a target or baseline.
  • Don't blame the data without evidence. Find the specific supporting metric that broke.

Your Win by Friday

By Friday, you'll have a clean one-page analysis. It will name the root cause (new sign-ups dropped 30%) and recommend a fix (check the sign-up flow for errors). Your boss will see a clear story, not a data dump. That's the win: ship analysis that drives action, not confusion.