PhonePe Business Analyst — Interview Questions
PhonePe Business Analyst Interview Questions
Business analysts at PhonePe sit between raw transaction data and decisions that affect merchant relationships, pricing, and competitive positioning in one of the most contested markets in Indian fintech. The interviews test whether a candidate can turn a vague business worry into a structured, data-backed diagnosis — and whether they know the difference between a correlation and an actual explanation.
PhonePe's Interview Process for Business Analyst
Typically includes a screening round, a case-study or SQL/analytics exercise, a business-reasoning round with a hiring manager, and a final round with a senior stakeholder from the business team. Expect questions grounded in merchant economics, regional performance, and competitive dynamics rather than abstract "estimate the market size of umbrellas" puzzles.
Question 1: The Regional Volume Drop
Merchant transaction volume in one city has dropped 15% over the past three weeks, while the rest of the country is flat to slightly up. Leadership wants to know why, and what to do about it. Walk through how you'd investigate.
Why interviewers ask this
This tests whether a candidate can structure an open-ended diagnostic problem instead of jumping to a single explanation. PhonePe wants to see a candidate rule hypotheses in or out systematically, using data they'd realistically have access to.
Example strong answer
"Before I look for a cause, I'd confirm the drop is real and not a data or measurement artifact — checking whether there was a reporting change, a merchant re-categorization, or a data pipeline issue around the same three-week window. That sounds basic, but ruling it out first saves a lot of wasted investigation.
Assuming it's real, I'd break the 15% drop down by segment: is it concentrated in a specific merchant category — say, small kirana stores versus larger retail chains — or spread evenly? Is it driven by fewer merchants transacting at all, or by the same merchants doing lower volume per transaction? Those point to very different causes: merchant churn versus reduced consumer spend versus a specific merchant-facing issue.
I'd also check for anything operationally specific to that city in the same window — a local competitor running an aggressive cashback campaign, a regulatory or local event disruption, or even something as simple as a payment-flow bug affecting devices common in that region. I'd pull support ticket volume from that city specifically, since a spike in complaints often correlates with — and sometimes precedes — a usage drop.
If the drop correlates with a competitor's campaign launch, that reframes the whole question from 'what's broken' to 'how do we respond competitively,' which is a different conversation with leadership than a technical or operational fix. I'd present findings as a ranked set of hypotheses with supporting and contradicting evidence for each, rather than a single confident answer, unless the data is genuinely conclusive — a 15% drop over three weeks is significant enough that I wouldn't want to recommend action based on a guess dressed up as an analysis."
Follow-up questions
- The drop turns out to be concentrated entirely in one merchant category. What data would you pull next to understand why that category specifically?
- Leadership wants a recommendation within 48 hours, but your investigation isn't fully conclusive yet. What do you present?
Question 2: Is the Cashback Campaign Worth It?
Marketing ran a cashback campaign last month that increased transaction volume by 22% during the campaign period. They want to make it a recurring monthly campaign. How would you determine whether this is actually a good decision for the business?
Why interviewers ask this
This tests whether a candidate looks past a headline growth number to unit economics and counterfactual thinking — did the campaign create new value, or just pull forward transactions that would have happened anyway, at a cost.
Example strong answer
"A 22% volume increase sounds like a clear win, but the real question is whether that volume was incremental — genuinely new activity the business wouldn't have had otherwise — or whether it was cannibalized from before or after the campaign window, and at what cost per incremental transaction.
First, I'd check the weeks immediately before and after the campaign for a dip relative to trend — if volume dropped noticeably right after the campaign ended, that suggests some of the 'increase' was users timing their existing spend to capture the cashback, not new spend. I'd also compare against a control cohort if one exists — users or regions not exposed to the campaign — to estimate what volume would have looked like without it.
Then the cost side: cashback campaigns have a direct payout cost per transaction, plus often a lower effective take-rate during the promo period if merchants are also getting reduced fees as part of the push. I'd calculate cost per incremental transaction, not cost per total transaction, using the counterfactual estimate from the control comparison — that's the number that actually tells you if this is worth repeating.
I'd also look at what kind of transactions the campaign drove — if it disproportionately pulled in small, low-value transactions or already-loyal users who would have transacted anyway, the incremental value is much lower than the headline 22% suggests. Ideally I'd want to see whether the campaign brought in new or dormant users specifically, since that's the segment where a cashback push has the clearest case for genuine incremental value.
My recommendation wouldn't be a flat yes or no — it'd likely be 'yes, but targeted differently': for example, restricting the cashback to dormant-user reactivation or new-user acquisition, where the incremental case is strongest, rather than a blanket campaign that pays out to users who'd have transacted regardless."
Follow-up questions
- The control comparison shows only 8 of the 22 points were genuinely incremental. How does that change your recommendation?
- Marketing argues the campaign also has a brand-awareness value that's hard to quantify. How do you factor that into your recommendation?
Question 3: Sizing the Tier 2/3 Lending Opportunity
Leadership wants a rough estimate of the addressable market for PhonePe's lending marketplace in tier 2 and tier 3 Indian cities. You won't have perfect data. How would you approach this?
Why interviewers ask this
This tests structured market-sizing reasoning under real data constraints — PhonePe wants to see a candidate build a defensible estimate from available proxies, and be explicit about assumptions rather than presenting a single confident number.
Example strong answer
"I'd build this top-down from PhonePe's own data first, since that's the most reliable input available, rather than starting from a generic national market-sizing exercise. I'd start with the number of active PhonePe users in tier 2/3 cities — data we already have — then apply a credit-eligibility filter based on transaction history depth and consistency, since lending eligibility typically requires a minimum track record of financial activity.
From there, I'd estimate what fraction of eligible users are likely to want and qualify for a lending product, using proxies like existing informal or formal borrowing behavior visible in transaction patterns — recurring EMI-like payments to other platforms, for instance — as a signal of credit appetite and creditworthiness, while being clear this is an approximation, not a precise measurement.
I'd triangulate that bottom-up estimate against external reference points — published RBI or industry data on tier 2/3 credit penetration and average loan ticket sizes for similar digital lending products, where available — to sanity-check whether my internal-data-based number is in a plausible range, rather than relying on either source alone.
I'd present the result as a range, not a point estimate — say, 'X to Y crore in addressable annual loan originations, driven primarily by assumptions on eligibility conversion rate' — and I'd explicitly flag which assumption the range is most sensitive to, since that tells leadership what to validate next rather than treating the number as settled. If the eligibility-to-actual-uptake conversion assumption is the biggest swing factor, that's the thing worth running a small pilot to actually measure, rather than debating further in the abstract."
Follow-up questions
- Your range comes out very wide — 3x between the low and high end. How do you tighten it, or do you present it as-is?
- How would you adjust this estimate for the fact that tier 2/3 cities have historically lower digital lending adoption than tier 1?
Question 4: Responding to a Competitor Fee Cut
A major competitor just announced they're cutting merchant transaction fees to zero for the next six months. Leadership asks you to analyze the impact and recommend a response. What's your approach?
Why interviewers ask this
This tests competitive-response reasoning — whether a candidate reacts reflexively (match the price cut) or actually analyzes which merchant segments are price-sensitive versus sticky, and what a sustainable response looks like versus a short-term reaction.
Example strong answer
"My first instinct wouldn't be to recommend matching the cut immediately — that's the reflexive response, and it's expensive and hard to reverse once merchants get used to it. I'd want to understand our actual exposure first: which merchant segments are most likely to switch based on fees alone, versus which are sticky for other reasons — integration depth, settlement reliability, support quality, or simply switching-cost inertia.
I'd pull data on merchant fee sensitivity if we have any prior signal — merchants who've previously churned or reduced usage after fee changes, or merchants in categories with thin margins where even small fee differences matter a lot, like grocery or fuel, versus categories where transaction fees are a rounding error relative to margin, like high-ticket retail.
I'd also estimate the actual revenue exposure — how much of our merchant fee revenue comes from the segment most likely to be poached, not just total merchant fee revenue, since a blanket response sized to defend all revenue when only a fraction is actually at risk would be overreacting and costly.
My recommendation would likely be segmented rather than a blanket match: matching or beating the fee cut specifically for the price-sensitive, thin-margin segments most exposed to switching, while holding pricing for segments where our stickiness comes from something other than fees — and using the resources saved from not doing a blanket cut to invest in the non-price differentiators, like faster settlement or better merchant support, that make switching less attractive regardless of fees.
I'd frame this to leadership with an explicit time-bound: since the competitor's cut is stated as six months, I'd propose treating our response as provisional too, reassessing based on actual observed merchant movement in the first 4-6 weeks rather than locking into a permanent pricing change based on a competitor's temporary promotional move."
Follow-up questions
- Six weeks in, data shows minimal merchant movement even in the price-sensitive segment. Does that change your view of the original analysis?
- How would you factor in the risk that the competitor makes the zero-fee period permanent rather than temporary?
Question 5: Fixing the Merchant Onboarding Drop-Off
PhonePe's merchant onboarding funnel shows a significant drop-off between "application started" and "application completed" — roughly 40% of merchants who start never finish. How would you diagnose and address this?
Why interviewers ask this
This tests funnel-diagnosis skills and whether a candidate distinguishes between a UX/process problem and a fundamental eligibility or motivation problem — each requires a very different fix.
Example strong answer
"A 40% drop-off is large enough that I'd want to break the funnel into its individual steps rather than treating 'onboarding' as one stage, because the fix for someone dropping off at document upload is completely different from someone dropping off at business verification.
I'd pull step-by-step completion rates across the funnel to find where the biggest single drop happens — if it's concentrated at one specific step, like document upload, that points toward a UX or technical friction problem: is the upload failing on certain devices, is the required document list unclear, is the file size limit too restrictive for how merchants are actually photographing documents. If the drop-off is spread evenly across every step instead, that suggests something more fundamental, like the wrong merchant segment being funneled into onboarding in the first place, or the value proposition not being clear enough to sustain motivation through a multi-step process.
I'd also segment by merchant type and acquisition channel — small individual merchants and larger registered businesses likely have very different onboarding friction points, and merchants acquired through a field sales team versus a self-serve app flow may behave completely differently, since a field agent can walk someone through friction that a self-serve flow can't.
For merchants who dropped off, I'd want a small follow-up outreach — even a short survey or a sample of support-team calls — to get qualitative reasons directly, since funnel data tells you where people drop off but not always why, especially for issues like confusion or distrust that don't show up cleanly in step-completion metrics.
Once I have a specific, data-backed hypothesis — say, document upload failing disproportionately on lower-end Android devices — I'd propose a targeted fix and measure its impact on that specific step's completion rate before assuming it fixes the whole funnel, since onboarding funnels often have more than one meaningful leak point."
Follow-up questions
- The data shows drop-off is spread evenly across every step, not concentrated anywhere. What does that change about your investigation?
- How would you prioritize fixing this funnel against other competing analytics priorities, given it's a 40% drop-off but each individual step loss looks smaller?
Preparation tip
PhonePe's business analyst interviews consistently reward candidates who resist the first plausible explanation and instead lay out competing hypotheses with what evidence would confirm or rule out each one. Answers that jump straight to a recommendation without first structuring the diagnosis tend to read as guesswork, even when the final recommendation happens to be right.