Case Study / Business Problem Questions
Part of the Data Science Roadmap
In short
'Revenue dropped 15% last month. How would you investigate?' There's no trick answer. Interviewers want calm, structured thinking that moves from the metric to likely causes to the data you'd check.
'Revenue is down, how would you investigate using data?' — testing structured, hypothesis-driven thinking that connects business acumen (Phase 13) with your technical toolkit.
Think of it like this
The scooter mechanic
When a scooter won't start, a good mechanic doesn't replace the engine. He checks in order: is there petrol, is the spark plug firing, is the battery alive? Each check rules out a branch. Business case questions want the same calm, step-by-step narrowing.
Watch it happen
Breaking down 'revenue is down'
1/5First clarify: which revenue, compared to when, and is it sudden or gradual?
The big ideas
What you need to know
Clarify the metric
Pin down the definition, time window and comparison before guessing causes.
Rule out data issues
A surprising number of 'drops' are broken tracking. Say you'd check that first.
Split the metric
Break it into parts that don't overlap and together cover everything.
End with an action
Close with your top hypothesis, the data to confirm it, and what you'd recommend.
See it in code
Walk through the code
1SELECT2 DATE_TRUNC('week', order_date) AS week,3 COUNT(*) AS orders,4 COUNT(DISTINCT customer_id) AS customers,5 SUM(amount) / COUNT(*) AS avg_order_value,6 SUM(amount) AS revenue7FROM orders8WHERE order_date >= CURRENT_DATE - INTERVAL '12 weeks'9GROUP BY 110ORDER BY 1;1/3Bucket orders by week to see when the drop started.
A first query for 'revenue is down' (PostgreSQL syntax). It shows which part of revenue moved, week by week.
Save yourself the pain
Common mistakes
Oops
Jumping to 'it must be a competitor' in the first ten seconds.
Do this instead
Clarify the metric, rule out data issues, then break the metric down before naming causes.
Oops
Listing ten random hypotheses with no order.
Do this instead
Use a tree, go branch by branch, and say which data would confirm or kill each idea.
Your mission
Practise one case aloud
Quick check
'Daily active users fell 20% yesterday.' What's the best first step?
Go deeper
- 1Practice the structure: clarify the metric, list hypothesis categories, prioritize by likelihood.
- 2Prepare a real example from a past project or Phase 14 portfolio piece if you have one.
- 3Practice narrating your reasoning out loud — the process is what's being evaluated.
More case studies & product sense at this stop
ML System Design Questions
A/B Testing & Experimentation Questions
Stuck on this topic? Ask an Insider
Get 1:1 guidance from people who've walked this exact path — on the InsideEdge app.
