Stop 15 · Interview PreparationBusiness & Product Sense · 3/3

Root Cause Analysis Questions

Part of the Data Analyst Roadmap

In short

"Orders dropped 20% yesterday. What happened?" This classic question tests whether you investigate in a calm, structured way instead of guessing the first thing that comes to mind.

A close cousin of business case questions, focused specifically on diagnosing why a metric moved — testing hypothesis generation and how you'd rule causes in or out with data.

The jaadu holds up a pressure cooker lid and inspects its rubber ring with a puzzled look.

Think of it like this

The pressure cooker that won't whistle

When the cooker doesn't whistle, you don't throw it away. You check the obvious things first: is the flame on, is there water, is the rubber ring fitted, is the vent blocked. One by one, until you find it. Root cause analysis is the same checklist thinking.

Watch it happen

Why did orders drop 20% yesterday?

Orders -20%
Data issue?
Internal
External
Tracking bug
App release
Price change
Holiday, rain
Competitor sale

1/5Clarify first: which metric exactly, compared with what, and is it all users or some?

The big ideas

What you need to know

Rule out data issues first

A surprising number of "drops" are broken tracking. Check another source before panicking.

Use an issue tree

Split causes into data, internal and external. It shows the interviewer you're being thorough.

Segment to narrow down

Cut by platform, city, app version and channel. A drop in one segment points straight at a cause.

Check the calendar

Compare with the same day last week and last year. Festivals and long weekends explain many dips.

Save yourself the pain

Common mistakes

Oops

Jumping straight to "probably a competitor" without checking anything.

Do this instead

Work through the tree: data, internal, external, then segment to confirm.

Oops

Exploring every possible cause in equal detail.

Do this instead

Start with the most likely and easiest to check, and say why you're ordering them that way.

Your mission

Build your own issue tree

~25 min

Quick check

Orders dropped sharply, but only on Android, only on the newest app version. What's the most likely cause?

Go deeper

  1. 1Practice listing hypothesis buckets: internal changes, external factors, seasonality, data/tracking issues.
  2. 2Prepare a real example from a past project (or Phase 14 portfolio piece) if you have one.
  3. 3Practice narrating your reasoning out loud — the process is what's being evaluated.

More business & product sense at this stop

Stuck on this topic? Ask an Insider

Get 1:1 guidance from people who've walked this exact path — on the InsideEdge app.

Download
InsideEdge