HackTheRounds Interview Experiences

Capital One Business Analyst Interview Experience (2026) - Pet Credit Card Profitability & DAU Drop Diagnosis, Offer

Capital One BA Power Day: pet owner credit card profitability case with CAC and LTV reasoning, iOS only DAU drop diagnosis, behavioral storytelling, and a first

By Anonymous · 2026-04-18

Background

Capital One's Business Analyst role is not a software role and should not be prepped like one. I want to be upfront about that because too many posts online conflate the BA loop with the SWE loop and mislead candidates. I am a recent econ and stats grad applying to the BA Development Program, aiming for the retail banking track. I had prior BA internships at a smaller regional bank, strong Excel and SQL, moderate Python for analysis. My case prep was mostly "Case In Point" plus mock cases with friends from the career center. This post walks through the Power Day I actually sat and what I learned.

Timeline

Online application: early March
Resume screen + recruiter call (30 min): 2 weeks later
Online logic and analytical assessment (CodeSignal-style, not coding): 4 days after recruiter call
Power Day invite: 6 days after the assessment
Power Day (virtual, half-day): 2 weeks after invite
Hiring manager final: 5 business days after Power Day
Offer call: next-day after the final
Total: ~6 weeks.

Loop Overview

Capital One BA is heavy on cases and light on technical. Across the full loop I did:

Recruiter call (behavioral, motivation).
Online logic and analytical assessment. Puzzles, chart reading, business-math word problems. No coding.
Power Day: two case rounds and one behavioral round, 45 minutes each, back-to-back with a 10-minute break.
Final HM round: 60 minutes, deeper case discussion plus strategic questions about the team.

The Power Day is the load-bearing piece. If you clear Power Day, the HM is usually a formality.

Online Logic and Analytical Assessment (60 min)

Not coding. This is Capital One's version of a consulting PST. Roughly 25 to 30 questions spanning pattern puzzles, business-math word problems, and chart and table reading. I averaged 100 seconds per question and had to skip two. Strategy: skip the two-paragraph word problems on first pass, clear the chart and pattern questions, come back to the word problems with remaining time.

This question is coming soon to HackTheRounds.

Power Day Round 1: Profitability Case (45 min)

Problem: Capital One is considering launching a new credit card targeted at pet owners. You are given: annual fee $95, customer acquisition cost $150, projected cards issued in year one 100,000, expected interchange revenue per card per year $80, and a default-loss estimate. Compute year-one profit and recommend whether to launch.

This is a textbook profitability case. The way to win is structure first, numbers second. I opened by writing the equation on the shared whiteboard: profit equals revenue minus cost, where revenue decomposes into annual fee, interchange, and optional interest income, and cost decomposes into CAC, default loss, and operating cost.

The interviewer did not give me operating cost up front. I asked. She smiled and said "good that you asked," then gave me a number. That one question was worth real signal.

Plugging the numbers in, the year-one result came out as a loss. The follow-up was the interesting part: "Given year one is negative, why would we still launch?" My answer was built around Customer Lifetime Value:

Year one includes the one-time CAC of $150. Years two onward do not, so cohort revenue crosses breakeven in year two or three depending on retention.
Cross-sell: a pet-card customer is a candidate for the broader Capital One credit portfolio. The real unit economics sit at the relationship level.
Segmentation: pet owners are a growing demographic with high discretionary spend, and this product positions Capital One earlier in their wallet.

She pushed on "what retention rate makes this break even in year two?" which is a clean algebra problem. I walked through the setup verbally and solved to a rough number. She nodded.

This question is coming soon to HackTheRounds.

Power Day Round 2: Product Metric Diagnosis (45 min)

Problem: A mobile app's Daily Active Users dropped 15% in the third week of last month. Diagnose why.

This is the canonical product case and Capital One runs it often. The mistake everyone makes is jumping straight to hypotheses. The right move is to clarify the data first.

My clarification volley:

Is the drop platform-specific? iOS only, Android only, or both?
Is it geography-specific? US only or global?
Is it user-segment specific? New users, power users, or across the board?
Is it a specific feature's DAU, or the app's overall DAU?
Is the DAU metric definition stable, or did the team change the definition recently?

She said the drop was iOS-only and affected all segments. That narrows the field immediately. I walked through an internal-vs-external split:

Internal: - iOS release around that week. A new app version with a crash bug. - Analytics SDK or event-logging bug that under-reports on iOS. - A UX change on iOS that confused users or broke a key flow.

External: - iOS 17.x OS update that broke compatibility. - A competitor iOS launch that pulled users away. - Apple Store policy change, like a privacy prompt, that suppressed signals.

The interviewer then revealed: there was an iOS app update that coincided with a new iOS OS release, and the combination broke the push notification permission flow. Users stopped getting their daily nudge, and DAU dropped. The fix was to hot-patch the notification flow and restore the nudge.

The tip I took from this round: ask clarifying questions one at a time, not as a list. The interviewer gives you a signal with each answer, and you can thread to the right branch faster.

This question is coming soon to HackTheRounds.

Power Day Round 3: Behavioral (45 min)

Prompt: Tell me about a time you used data to convince someone.

Classic STAR. I used an internship story where I persuaded my PM to re-prioritize a reporting feature by showing 70% of active users had requested the capability in support tickets. Situation: PM had deprioritized as "old-school Excel-style reporting." Action: pulled six months of ticket data, tagged by theme, plotted the request distribution, and presented a one-page memo. Result: the feature shipped two months later.

Every Power Day behavioral question I heard from peers was a variation of "translate data into a decision for someone who does not speak SQL."

Final Hiring Manager Round (60 min)

More strategic conversation than case. He asked what I would prioritize in my first 90 days. I leaned on the job description and gave a three-phase answer: ramp on SQL and the internal data marts, shadow one full reporting cycle, then pick one recurring report and propose one automation or improvement by day 60.

He also asked "speed or rigor on a rush report?" The truthful answer is it depends on the audience, but for senior stakeholders I pick rigor, ship a partial with explicit caveats, and follow up with the full cut next cycle. He liked that you ship something rather than nothing.

Result

Offer call came the day after the final. BA Development Program, retail banking track, starting compensation at the upper end of the published band. Signing bonus standard. I accepted.

Tips

Capital One BA is a case interview role, not a coding role. Do not over-prep SQL and under-prep cases. The platform test has zero coding. The Power Day has zero coding. The HM has zero coding. Your prep hours should match that weight.
For profitability cases, write the Revenue = ... and Cost = ... equations before plugging numbers in. Half of candidates dive to the math and miss a cost line like operating cost. Writing the structure forces you to ask for missing inputs.
Ask clarifying questions one at a time on product metric cases. Dumping five questions at once makes you look unfocused. One question, get an answer, thread to the next branch. It is closer to real diagnostic work and it reads that way.
Use Customer Lifetime Value as a follow-up anchor. When the year-one number is negative, the interviewer is testing whether you know why year-one losses can still be rational. LTV, cross-sell, segmentation, and cohort retention are the four hooks they are listening for.
Translate-to-stakeholder stories are gold for the behavioral. Capital One's unique selling point versus Google or Meta is that BAs translate data to non-technical business partners. Every behavioral should have at least one story where you simplified a complex analysis for an audience that does not speak your language.
On the HM round, expect a "what would you do in your first 90 days" question. Have a three-phase answer ready: ramp (shadow, ask questions), own (take one deliverable to completion), and propose (identify one improvement). This is a framework that fits almost any team description they throw at you.