Problem-solving interview questions for analystsAnalyst interview questions
← Interview & exam prep
Analyst interview questions

Problem-solving interview questions for analysts

12 questions from beginner to advanced, each with the correct answer and a one-line explanation of why. Study them here, or take them as a timed test.

12 questionsBeginner 6 · Intermediate 6Answers + explanations
Take it as a timed test
Free PDF
All 389 questions & answers, in one file

DAX, Excel, Power BI, Fabric, certification practice and analyst interviews — every answer with the reason behind it. Print it, study it offline.

Instant download, plus the free Tuesday briefing. Unsubscribe in one click.

Beginner 6 questions

1A dashboard number suddenly drops to zero. The best FIRST step is to…

  1. Change the visual type
  2. Assume the source is down
  3. Ignore it and refresh later
  4. Rebuild the whole report
  5. Isolate where it breaks: source data → transformations → measure
Show answerHide answer

E · Isolate where it breaks: source data → transformations → measure Localize the failure systematically instead of guessing or rebuilding blindly.

2A good way to tackle a large, unfamiliar problem is to…

  1. Guess an answer
  2. Break it into smaller, well-defined sub-problems
  3. Do only the easy part
  4. Solve it all at once
  5. Wait for detailed instructions
Show answerHide answer

B · Break it into smaller, well-defined sub-problems Decomposition turns an overwhelming problem into tractable pieces.

3Estimating 'how many coffees this office drinks per day' is best treated as…

  1. A trick with no answer
  2. A trivia question
  3. A structured estimation: state assumptions and reason step by step
  4. Impossible without exact data
  5. A statistics exam
Show answerHide answer

C · A structured estimation: state assumptions and reason step by step It's a Fermi-style estimate — reasonable assumptions and simple arithmetic.

4Before optimizing a slow report, you should first…

  1. Measure to find the actual bottleneck
  2. Delete visuals at random
  3. Buy more hardware
  4. Rewrite everything
  5. Assume the cause
Show answerHide answer

A · Measure to find the actual bottleneck Measure first — optimize the real bottleneck, not a guessed one.

5Root cause analysis aims to…

  1. Restart the system
  2. Only document the issue
  3. Fix the symptom quickly
  4. Assign blame
  5. Find the underlying cause, not just the symptom
Show answerHide answer

E · Find the underlying cause, not just the symptom Fixing the symptom lets the problem return; the root cause prevents recurrence.

6When two data sources disagree, the sound approach is to…

  1. Average them
  2. Trust the newer file
  3. Pick the higher number
  4. Trace each to its definition and source of truth, then reconcile
  5. Ignore one
Show answerHide answer

D · Trace each to its definition and source of truth, then reconcile Reconcile by understanding each definition and lineage.

Intermediate 6 questions

7The '5 Whys' technique is used to…

  1. Repeatedly ask why to reach the root cause
  2. Collect five opinions
  3. Rank five tasks
  4. List five fixes
  5. Ask five people
Show answerHide answer

A · Repeatedly ask why to reach the root cause Asking 'why' iteratively drills past symptoms to the underlying cause.

8Sales are down 20%. The strongest analytical move is to…

  1. Blame the market
  2. Guess a reason
  3. Wait for next month
  4. Look only at the total
  5. Segment/decompose by region, product, channel and time to localize the driver
Show answerHide answer

E · Segment/decompose by region, product, channel and time to localize the driver Break the aggregate down to find where the change actually comes from.

9A hypothesis-driven approach means you…

  1. Explore data at random
  2. Avoid data
  3. Confirm what you already believe
  4. Ask the boss to decide
  5. Form a testable hypothesis, then seek data to confirm or reject it
Show answerHide answer

E · Form a testable hypothesis, then seek data to confirm or reject it State a hypothesis up front and let evidence settle it.

10When you can't get the ideal data, a pragmatic step is to…

  1. Use a reasonable proxy and state its limitations
  2. Wait indefinitely
  3. Abandon the analysis
  4. Hide the gap
  5. Fabricate the numbers
Show answerHide answer

A · Use a reasonable proxy and state its limitations A clearly-caveated proxy keeps analysis moving honestly.

11To guard against confirmation bias, you should…

  1. Only look for supporting data
  2. Trust your gut
  3. Pick the nicest chart
  4. Stop as soon as it looks right
  5. Actively seek evidence that could disprove your hypothesis
Show answerHide answer

E · Actively seek evidence that could disprove your hypothesis Trying to falsify your own hypothesis is the antidote to confirmation bias.

12Deciding which problem to solve first, a useful lens is…

  1. The loudest requester
  2. Random choice
  3. Newest request first
  4. Easiest first always
  5. Impact versus effort (and urgency)
Show answerHide answer

E · Impact versus effort (and urgency) Prioritize by value and feasibility, agreed with stakeholders.

More analyst interview questions