khatabook Business Analyst Interview: Questions, Experience & Prep (2026)

Khatabook is one of India's best-known fintech startups, building digital bookkeeping and payments tools for small merchants and kirana shop owners. With 61 open roles listed as of mid-2026, the company is actively hiring Business Analysts to help decode merchant behaviour, improve product features, and grow its SMB user base across India.

The BA interview at Khatabook typically runs across multiple stages: an initial recruiter screening, one or two analytical or case rounds, a SQL or data skills test, and a final round with a senior manager or business head. Candidates report that the process puts real-world data thinking ahead of textbook theory. You should expect questions about merchant cohort analysis, product metrics, and how you would prioritise features for a user base that is largely offline and price-sensitive.

Across all 398 Business Analyst openings tracked on knok's job radar, Bangalore leads with 53 roles and Delhi follows with 48, so Khatabook fits a wider national hiring trend for this function.

These questions have been reported by candidates who interviewed at Khatabook for BA roles. Prepare answers that tie directly to Khatabook's merchant-first product context.

1. Walk me through a time you turned raw data into a business decision that had a measurable impact.
2. Khatabook's core users are kirana owners. How would you define and track 'engagement' for this audience?
3. Write a SQL query to find merchants who recorded transactions in one week but made zero entries the following week.
4. A product feature shows strong adoption in Tier-1 cities but almost none in Tier-2 and Tier-3 towns. What would you investigate?
5. How do you decide which metric to move when you have three stakeholders each pushing a different KPI?
6. Explain the difference between a leading indicator and a lagging indicator. Give one example of each from a payments product.
7. Describe how you would set up an A/B test for a new onboarding flow aimed at first-time app users.
8. Khatabook wants to reduce merchant churn. What data would you pull first, and how would you segment the problem?
9. A dashboard shows daily active users dropped sharply one morning. Walk us through your investigation approach step by step.
10. How would you build a prioritisation framework for a backlog that has both merchant-requested features and internally identified gaps?
11. Tell me about a time you had to push back on a senior stakeholder's assumption using data.
12. If you had to pick one north-star metric for Khatabook's bookkeeping product, what would it be and why?

Use the STAR format (Situation, Task, Action, Result) for every behavioural question. The three examples below are written for the Khatabook context.

Q: Tell me about a time you turned raw data into a business decision.

*Situation:* At my previous company, the operations team believed support tickets were rising because of app bugs.

*Task:* I was asked to validate this claim and recommend a fix.

*Action:* I pulled ticket data, tagged each ticket by root cause, and found that the majority were related to a confusing payment confirmation screen, not bugs. I mapped the funnel drop-off to that exact screen and built a one-page summary with annotated screenshots for the product team.

*Result:* The design team redesigned the screen in the next sprint, and support tickets for that category fell significantly in the following month. The exact figure was tracked internally, but I can share the methodology if helpful.

---

Q: A DAU drop appeared one Monday morning. Walk us through your investigation.

*Situation:* At a previous role, we saw a sudden drop in daily logins on a Monday.

*Task:* I had to identify whether this was a real user-behaviour change or a data pipeline issue.

*Action:* I first checked whether the pipeline had run correctly and whether any events were missing. Once the data was confirmed clean, I segmented the drop by platform (Android vs. iOS), by city tier, and by user cohort age. I noticed the drop was isolated to Android users in Tier-2 cities who had received a push notification the night before.

*Result:* We found that the notification content was mistranslated in a regional language, which caused users to uninstall. The team pulled the notification series and the metric recovered within a few days.

---

Q: Describe a time you pushed back on a stakeholder using data.

*Situation:* A senior product manager was convinced that adding a loyalty points system would improve merchant retention at my previous company.

*Task:* I was asked to size the potential impact before committing engineering resources.

*Action:* I ran a cohort analysis on retained versus churned merchants and found that the top predictor of retention was frequency of bookkeeping entries, not any gamification feature. I put together a short presentation showing that merchants who recorded transactions regularly churned far less, and that a loyalty layer would not address the core drop-off point.

*Result:* The team shifted focus to a guided onboarding flow that helped merchants build a daily bookkeeping habit. The loyalty points idea was shelved for the quarter, saving engineering time.

For product metric questions, use the funnel approach: define the user journey step by step, identify where users are dropping off, and then ask what you can measure at each step. For Khatabook, a typical funnel runs from app install to first transaction recorded to weekly active bookkeeping.

For prioritisation questions, candidates report that Khatabook interviewers respond well to frameworks that weigh merchant impact against implementation effort. A simple two-by-two grid works: high impact and low effort items get built first, low impact and high effort items get dropped. Always tie your prioritisation back to the north-star metric.

For SQL questions, think aloud as you write. Interviewers typically care more about how you decompose the problem than whether your syntax is perfect. Break the question into: what table do I need, what filter do I apply, what do I group by, and what do I compute.

For case or estimation questions, structure your answer as follows: clarify assumptions, state what data you would want, build a rough model, then sense-check the output. For a merchant-focused product like Khatabook, ground your assumptions in the reality of a small kirana: a shop with one or two people running it and low digital literacy.

For behavioural questions, use STAR but keep the Result grounded. If you do not have a precise number, describe the direction and magnitude ('fell significantly', 'improved by a meaningful amount') and offer to share the method. Invented numbers hurt your credibility more than honest hedging.

Candidates who clear Khatabook BA interviews consistently report a few traits that interviewers value.

Merchant empathy. Khatabook's users are small business owners, many of whom are not highly tech-savvy. Interviewers want to see that you understand this context and will not propose solutions that assume high digital fluency.

Comfort with messy data. Fintech data for SMBs is often incomplete, inconsistent, or sparse. Interviewers typically probe whether you can work with what you have rather than waiting for a perfect dataset.

Clear communication of analysis. You will likely be asked to present findings to a hypothetical non-technical stakeholder. Interviewers look for candidates who strip out jargon and lead with the 'so what'.

Structured thinking under pressure. When a question is ambiguous, interviewers want to see you ask clarifying questions and frame your approach before diving into an answer. Jumping straight into numbers without framing is a commonly reported reason candidates do not progress.

Ownership mindset. Khatabook is a startup environment. Interviewers look for candidates who treat business outcomes as their own responsibility, not just analysis tasks handed to them.

Week 1: Product and company context

Use the Khatabook app yourself if possible, even briefly, to understand the bookkeeping and payments experience from a merchant's point of view. Read publicly available news about the company's product direction. Form your own opinion on which metrics matter most for a bookkeeping product targeting kiranas.

Week 2: SQL and data skills

Practice queries involving window functions, cohort segmentation, and retention logic. Focus on questions like 'find users active in one period but not the next' since these are commonly reported in Khatabook rounds. Write your queries without looking up syntax so you can think aloud confidently.

Week 3: Behavioural stories

Map your past projects to the most common themes: turning data into a decision, pushing back on a stakeholder, defining or changing a metric, and cross-functional collaboration. Write out each story in STAR format and time yourself to under three minutes.

Week 4: Mock rounds

Do at least two end-to-end mock interviews covering a case question, a SQL question, and a couple of behavioural questions. Ask for feedback specifically on whether your communication is clear to someone who has not seen the data.

If you are tracking openings while you prepare, knok checks 150+ job sites nightly, applies to jobs matching your resume, and messages HR for you, which frees up time for actual interview prep.

Citing numbers you cannot defend. If you used a rough estimate in your answer, say so. Interviewers at data-driven companies will probe your numbers, and getting caught inflating a result damages trust more than any honest approximation.

Jumping to a solution before defining the problem. When asked 'how would you reduce churn', many candidates immediately propose features. Interviewers typically expect you to first define what churn means, what data you would look at, and what hypotheses you would test.

Ignoring the Tier-2 and Tier-3 context. Khatabook's growth is closely tied to non-metro merchants. Answers that only consider urban, digitally fluent users miss a large part of the product reality.

Treating SQL as a memory test. If you forget the exact syntax for a window function, say 'I would use a window function here, let me think through the logic' and keep going. Stopping cold is worse than working through it imperfectly.

Giving generic behavioural answers. Saying 'I am a team player who loves data' adds nothing. Every story must have a specific situation, a concrete action you took personally, and an outcome you can describe even if you cannot share the exact number.