Who This Helps
You're a Junior Analyst staring at a dashboard that just turned red. A key metric dropped, and your stakeholder wants answers by Friday. This guide is for you. It's pulled straight from the Data Storytelling for Stakeholders course, specifically the "One Key Message" mission.
Mini Case
Imagine you track weekly active users. Last week, the number fell 12% — from 10,000 to 8,800. Your boss, Li Wei, needs to know why and what to do. You have one hour to diagnose the drop and prepare a recommendation.
Do This Now (5 Steps)
- Check the date range. Was last week a holiday? Compare same period last month. If last week had a public holiday, the drop might be normal.
- Segment the users. Break the 12% drop by region, device type, or plan. You might find that 80% of the drop came from mobile users on Android. That's your first clue.
- Look for a single event. Did you push a new feature? Change a pricing page? Run a buggy release? A 12% drop in 7 days often traces to one change.
- Talk to one teammate. Ask the product manager: "Did we ship anything last Tuesday?" That 30-second chat can save you hours of digging.
- Write one key message. Summarize your finding in one sentence. Example: "The 12% drop in active users is driven by Android users after the login update on Tuesday." Then add your recommendation: "Roll back the update and test with 5% of users."
Avoid These Traps
- Don't chase every possible cause. Focus on the segment that explains most of the drop.
- Don't present raw numbers without context. A 12% drop sounds scary, but if it's only 200 users, it's less urgent.
- Don't skip the recommendation. Stakeholders want a decision, not just a diagnosis.
- Don't overcomplicate your chart. A simple bar chart showing the drop by segment is enough.
- Don't wait until Friday. Share your finding by Wednesday so the team can act.
Your Win by Friday
By Friday, you'll have shipped a clean analysis with one clear recommendation. Your stakeholder will know the root cause (Android login bug), the impact (12% drop, 1,200 users lost), and the next step (rollback and test). Plus, you'll look like the analyst who gets things done — and maybe even earn a high-five from Li Wei.