30 interview questions, with the answers that get the offer

Strong answer

"I'm a data analyst, 6 years in, the last 3 at a logistics company where I own the pricing dashboards 40 account managers use daily. The project I'm proudest of this year: I found that our discount approvals took 4 days on average, built the approval queue into the dashboard, and cut it to same day. That saved the sales team about 2 hours a week each. I'm looking at this role because it's the same work at 10 times the data volume, and your post says the team owns pricing experiments end to end. That's the part I want to do more of."

• Now, one result, why here. Done in 90 seconds.
• Number on the result (4 days to same day).
• Quoted the job post back to me.

Strong answer

"Three chapters. Chapter one, 2 years at an agency, where I learned to ship weekly for clients who changed their minds on Friday. Chapter two, 3 years at a fintech, where I moved from analyst to lead and ran the team's first A/B testing program, 30 tests in year one. Chapter three, the last 18 months at a startup that was acquired in June, which is why I'm talking to you. Each move was for more ownership, and this role is the next step in that line."

• Three chapters, one sentence each.
• Explained the layoff in one neutral clause.
• 'More ownership' is a pattern I can score.

Strong answer

"Three concrete reasons. Your post says this team owns the pricing experiments end to end, and I've only ever owned analysis, so this is the ownership step for me. Second, I read your engineering blog post on the experimentation platform rebuild from March, and the problem it solved, metric drift between teams, is the exact problem I fought at my last company. Third, the role is hybrid in Austin and I'm already there. I'd rather be in the room on launch days."

• Quoted the post, named a blog post with a date.
• Connected a company problem to their own experience.
• Honest practical reason (location). I believe the other two more because of it.

Strong answer

"I've done the job I came to do. I joined to build the reporting layer from nothing, and 3 years later it runs itself, with 2 analysts I trained maintaining it. The next version of my job here would be more of the same, and I want to move from reporting into experiments. Your post is that move."

• Moving toward something. I believe this reason.
• Trained 2 people: a collaboration signal I didn't even ask for.
• Ties the exit to this role.

Strong answer

"You sell scheduling software to clinics, about 12,000 of them by your last press release, and the growth story since 2024 is the payments add-on, which I'd guess is why this role sits under the payments team. Your two biggest competitors are the practice management suites that bundle scheduling for free, so I expect retention is the metric this team lives on. Is that right?"

• Named the product, the customer, the number and the competitor.
• Guessed the team's metric and asked. That's how analysts talk.
• Ended with a question that invites me to correct them.

Strong answer

"Two that matter for this role. First, I finish ambiguous projects. Last year I was handed 'figure out why churn is up' with no owner and no definition. I defined churn with finance, found that 60% of it came from one onboarding step, and the fix we shipped brought churn down 2 points in a quarter. Second, I make numbers land with non-technical people. The sales VP still uses the one-page summary I wrote for that project in her board deck."

• Two strengths, each with a story and a number.
• Picked the strengths the job post asks for.
• The VP detail is proof someone else valued it.

Strong answer

"Last year my manager told me my written updates were too long. People skimmed them and missed decisions. That stung, because I thought thorough was good. I asked her for the one update she thought worked and copied its shape: decision first, three bullets, link to detail. I also started asking one stakeholder a month whether my updates were useful. Six months on, two of them told her unprompted that my updates were the ones they read."

• Real weakness, not core to an analyst's job.
• Named who gave the feedback and what they did about it.
• Measured the change (two stakeholders, unprompted).

Strong answer

"Cutting our discount approval time from 4 days to same day. It wasn't glamorous. I noticed account managers were emailing finance for every discount, mapped the 11 steps, and found that 8 of them were waiting. I proposed moving approval into the dashboard with a rules layer finance agreed on, built the first version in 3 weeks, and ran it with 5 account managers before rolling it out. Sales closed about 15% more deals in the quarter after launch, and finance said no to fewer requests because the rules were clear."

• 'I noticed, I mapped, I proposed, I built, I ran.' Five actions.
• A result for sales and a result for finance.
• Chose one. Didn't list three.

Strong answer

"My manager would say I'm the person she sends into a meeting when the numbers are going to be unpopular, because I stay calm and I bring the next step. My coworkers would probably say I over-document, and they'd be right. It comes from the written-updates feedback I mentioned. It's a habit I'm still tuning."

• A specific scene (unpopular numbers meeting).
• Named a flaw coworkers would name. Consistent with question 7.

Strong answer

"I launched an A/B test with the wrong success metric. We measured clicks on a new checkout button, clicks went up 20%, we called it a win and shipped it. Two weeks later finance flagged that completed orders were flat. I'd measured the step, not the outcome. I owned it in the team review, wrote a one-page metrics checklist every test now goes through, and I'm the reviewer on it. In the year since, 2 tests were stopped at design because the checklist caught the same mistake."

• Owned it, out loud, in front of the team.
• Made the fix structural (checklist, reviewer).
• Closed the loop: 2 tests caught since.

Strong answer

"At my last company I was told 'revenue per user is flat, find out why', with no owner and no deadline. I spent the first week on definitions: I sat with finance and product and we agreed what counted as revenue per user, because three teams had three numbers. Then I split it: price, frequency, and mix. Frequency was the drop, and it came from one customer segment. I sized the top 3 fixes by expected revenue and cost, and brought a one-page ranking to the product lead. We shipped the first fix in 6 weeks and frequency recovered in that segment by the next quarter."

• First action was alignment on the definition. Senior move.
• Decomposed the metric. Prioritized by ROI.
• Named the result and the time frame.

Strong answer

"An engineer on my team kept shipping tracking changes without telling analytics, and twice our dashboards broke on a Monday. I was annoyed, so I waited a day. Then I asked him for 20 minutes and started with his side: he had a release deadline and the tracking spec lived in a doc nobody read. We agreed on a 2-line change request in the pull request template, and I took on reviewing tracking changes within a day so he wouldn't be blocked. Dashboards haven't broken since, and he now pings me before big releases on his own."

• Waited a day. Asked for his side first.
• Fixed the system, not the person.
• Took on work to make it easy for him.

Strong answer

"My manager wanted to ship a feature before the A/B test reached sample size, because the quarter was closing. I asked for 30 minutes, brought the numbers: at the current traffic we'd have the answer in 9 days, and shipping now meant we'd never know if the 4% lift was real. I also laid out the risk if it was flat. She decided to ship anyway for the quarter, and I committed: I set up a holdback group so we'd still learn. The lift held at 3%, and she's used the holdback pattern on two launches since."

• Disagreed with numbers and a risk, not an opinion.
• Committed fully and still protected the learning.
• Changed how she works. That's influence.

Strong answer

"A week before a quarterly business review, the CFO changed the question from 'which regions grew' to 'which regions grew profitably', and our margin data lived in a different system. I called a 15-minute meeting with finance that afternoon, we agreed on a margin proxy we could pull in 2 days, and I cut two slides from the deck to make room. We presented on time. After the review I built the margin join properly, so the next quarter took zero extra time."

• Re-planned in an afternoon with the right people.
• Made a trade-off (cut two slides) out loud.
• Fixed it for next time.

Strong answer

"I needed the sales ops lead to change how discounts were logged, and he didn't report to me or care about my dashboard. So I stopped talking about the dashboard. I pulled 3 deals his team had lost that month where the discount history was unclear, and showed him the pattern. He asked how fast we could fix it. Two weeks later his team was logging it the new way, and he presented the change as his idea in the sales meeting, which was fine by me."

• Found what the other person cared about.
• Let him own the win.

Strong answer

"In March I had 3 'urgent' requests on the same day: a board slide, a bug in the churn report, and a sales territory analysis. I ranked them by who was blocked and by when the decision was being made: the board deck was Thursday, the bug was blocking finance's close today, the territory analysis had no decision date yet. I fixed the bug first, did the slide, and told the sales director the analysis would be Monday and why. He was fine with it. Saying the date out loud was the part that mattered."

• A ranking rule (who is blocked, when is the decision).
• Said no to someone with a date and a reason.

Strong answer

"When we moved from Redshift to BigQuery I had 3 weeks before the old warehouse shut off. I rewrote our 2 most-used queries first and compared results side by side to learn the dialect differences, kept a running doc of every gotcha, and asked the one engineer who'd used BigQuery before for a 30-minute review each Friday. By week 3 the team's top 20 queries ran, and the gotcha doc became the team's migration guide."

• A method: start with the highest-use cases, compare, document, ask for review.
• Left an artifact for the team.

Strong answer

"The head of customer support and I started badly: she thought analytics ignored her team. I asked for a standing 20 minutes every two weeks and spent the first three sessions only listening to what she needed. Then I built her one small thing: a daily ticket-volume alert. It took me 2 hours. After that she pulled me into planning, and her team's data became the source for the churn project I mentioned."

• Started from the other person's need.
• Small early win, then trust.

Strong answer

"Leading an analytics team of 3 to 5 that owns experimentation for a product area. The step between here and there is owning experiments myself instead of supporting them, and that's this role. I don't know the exact title in 5 years and I'm fine with that, as long as the next 2 years are spent on that step."

• A direction, with this role as the next step.
• Honest about not knowing the title.

Strong answer

"Small teams with a clear owner per problem, and written decisions. Your post mentions pods with a PM, a data scientist and engineers, which is exactly how my best year was set up. One thing I'd ask: how often do pods reshuffle? I do fine with change, but I do my best work after about 6 months with the same people."

• Specific preference, tied to our structure.
• Asked a real question back.

Strong answer

"Clear goals, then room. A weekly 30 minutes where I bring the decisions I need and the risks I see, and otherwise trust until I give a reason not to. When I was new to a domain I asked for more check-ins for the first quarter, and I'd do that here too."

• Named a cadence and a condition.
• Adapts to a new domain.

Strong answer

"Seeing a number I own move. The week the discount queue launched, I checked the approval time every morning, and watching it drop from days to hours is still the most fun I've had at work. The second thing is teaching: the 2 analysts I trained now do the reporting I used to do, and that's the part of leadership I want more of."

• A specific scene.
• Two motivations, both things this job offers.

Strong answer

"Yes, I'm in a couple of other processes at a similar stage. This role is the one I'm most interested in, for the ownership reason I mentioned, and I'd like to be able to make a decision with full information. What does your timeline look like from here?"

• Confirms options without playing games.
• Turns it into a timeline question.

Strong answer

"I'd like to learn more about the scope first, since it affects the number. Can you share the range budgeted for the role? If it helps, roles at this level in Austin that I've seen recently are in the 130 to 160 range, and I'd expect to be in that band."

• Asked for the range first.
• Gave a market range, not a current salary.

Strong answer

"First 30 days, learn: the data, the stakeholders, and the 3 reports people actually open. I'd ask every person I meet what they wish they knew. Days 30 to 60, one small win that someone asked for, shipped and announced. Days 60 to 90, a proposal for the first experiment I'd own, based on what I learned. I'd check this plan against yours in week one, because you know what's on fire."

• Learn, small win, proposal. Each with a time box.
• Offered to adjust to my priorities.

Strong answer

"I took 8 months off in 2024 after a layoff, 3 of them on purpose. In the other 5 I did two contract analytics projects for a nonprofit, which is where I learned dbt, and I've used it daily since. Happy to talk about either project."

• One sentence on the gap, two on what they did.
• A skill came out of it.

Strong answer

"Three. What would success look like for this role at 3 months and at 6 months? What's the biggest problem the team is facing this quarter? And, for you personally, what has kept you here?"

• Success criteria. The question I remember candidates by.
• Asked me a personal question. I'll talk for 3 minutes and like them more.

Strong answer

"Before I answer, two quick questions: is this metric drop sudden or gradual, and do we know if anything shipped that week? [Interviewer answers.] Then I'd split it into traffic, conversion and ticket size, check which one moved, and look at the segment it moved in before touching any hypotheses."

• Clarified first. Then a structure.

Strong answer

"One thing we didn't get to. Your post asks for experience presenting to executives, and I've presented at the monthly leadership review for the last 2 years. If that's a thing this role does, I'd be glad to walk through one."

• Checked the job post for the one requirement they hadn't covered. Prepared.

Strong answer

"I'd give my current team the standard 2 weeks, and I'd want to hand over the churn report cleanly, so about 3 weeks from an offer. I can be flexible by a week if you need me sooner."

• Finishes the current job well. That's how they'll leave us too.