Who This Helps
You are a Junior Analyst who just saw a key number tank. Maybe it's conversion rate, maybe it's daily active users. Your boss wants answers by Friday. No pressure, right? This is for you.
This approach comes straight from the Product Portfolio Strategy course. It helps you size the problem, find the real cause, and recommend a fix without drowning in data.
Mini Case
Imagine you track a portfolio of features. Last week, feature A had a 12% drop in engagement. You have 7 days to explain why. You pull a quick report and see the drop happened on Tuesday. That's your first clue.
You check the release log. Feature A got an update on Monday night. The update changed the onboarding flow. New users now skip a step that used to show them the feature's value. That's your root cause.
Do This Now (5 Steps)
- Isolate the drop time. Look at your KPI by day. Find the exact day it changed. If it's a flat line before Tuesday and a cliff after, you have a clear window.
- Check recent changes. List every deployment, campaign, or external event in that window. Feature updates, pricing changes, competitor moves. Write them down.
- Match the change to the KPI. Ask: does this change logically affect the metric? If you changed onboarding, expect engagement to drop. If you changed pricing, expect conversion to shift.
- Talk to one person who knows. Ping the product manager or engineer who made the change. They might say, "Oh yeah, we removed that step to speed up signup." That confirms your hypothesis.
- Write a one-page fix recommendation. State the root cause, the impact (12% drop), and one clear action. Example: "Add the value step back into onboarding. Expected recovery in 3 days."
Avoid These Traps
- Blame the data first. Don't assume a bug or bad tracking. Check the change log before you check the code.
- Over-analyze. You don't need a 50-slide deck. One page with the cause and fix is enough. Your boss wants speed, not perfection.
- Ignore the portfolio view. A drop in one feature might be a trade-off for a gain in another. Check if the team intentionally shifted focus.
- Skip the human step. A quick chat with the engineer saves hours of digging. People know things dashboards don't.
Your Win by Friday
By Friday, you will have a one-page analysis that says: "Here's the root cause, here's the impact, here's the fix." Your boss will see you as the analyst who gets to the point. And you will have learned a repeatable method for any future KPI drop.
Plus, you'll feel like a detective who cracked the case. That's a good feeling.