Zerodha PM Interview — Dormant Kite Trader Re-Activation

It tests whether you can re-activate dormant Kite traders inside Zerodha's hard constraints: no advertising spend and no aggressive nudges. The interviewer probes whether you scope the goal before solutioning, segment dormant users by why they actually went quiet rather than treating 1.6 crore users as one block, propose low-cost product and education levers like Varsity or Console insights instead of paid campaigns, and define a north-star metric plus a guardrail metric that protects trust. It also checks if you understand the SEBI F&O regulation shock that drove a lot of the recent dormancy.

Start by clarifying the goal and what re-activation actually means here, then break the dormant base into distinct groups by the reason they stopped trading, then pick one defensible group to go after first. Propose levers that cost almost nothing and do not spam, sequence a small measurable experiment, and state the one metric you would move plus the guardrail you would not let slip. Close by naming what you would do if the first experiment failed. Avoid leading with a feature or a campaign.

The biggest is proposing paid acquisition, performance marketing, paid referral, or discount-led win-back, all of which contradict Zerodha's no-advertising model. The second is recommending aggressive push notifications or repeated reminders, which breaks the no-spam line. The third is treating every dormant user as the same person instead of separating regulation-driven F&O dropouts from scared-off first-timers or one-time tax-season visitors. Many also jump straight to a feature with no goal scoping and no success metric, or propose a metric with no denominator or attribution.

The dynamic is close to a real Zerodha loop, which candidates describe as calm and conversational rather than aggressive. The persona, Meera, behaves like a senior product lead who pushes on every assumption and never praises an answer mid-round. The difference is consistency and instrumentation: it probes every claim for a baseline and attribution the same way every time, never gives outcome feedback during the session, and produces a transcript-backed scorecard afterward. A human interviewer varies; this one holds the bar steady.

Your transcript is evaluated against role-specific dimensions: how well you segment the dormant base, whether you reason inside the no-ads and no-spam constraint, the rigor of your metric definition including denominator and guardrail, your grasp of the Indian broking and SEBI F&O context, and how you sequence a low-cost experiment. Each dimension has observable signals drawn from what Zerodha actually evaluates. You get a scorecard that quotes the specific moments your reasoning held or slipped, rather than a single number.

Do not pitch. Spend the opening on diagnosis: ask what counts as dormant here, how leadership measures success, and what is known about why users went quiet. Then state your read of the distinct dormant groups out loud so the interviewer can react. Signal early that you understand the no-advertising and no-spam constraint without being told. This is the moment the interviewer decides whether you diagnose before you prescribe, which is the strongest early signal in this round.

Treat it as a design boundary, not a handicap. Reach for levers that are product-native and trust-building: Varsity education for users who lost confidence after a drawdown, Console portfolio insights that give a returning user a reason to log in, opt-in low-frequency digests, and removing re-KYC friction for users who genuinely want back. Frame every lever as something a returning user would thank you for, not something pushed at them. If you would not want the SMS yourself, do not propose it.

A strong answer names three or four distinct dormant groups with a different root cause for each, picks one to go after first with a stated reason, proposes one or two low-cost product or education levers tied to that group, and defines a north-star metric like returning-trader weekly active with a clear denominator plus a guardrail metric such as unsubscribe or trust signals. It sequences a small measurable test, states how it would attribute the lift, and says what it would do if the test failed. It never reaches for spend.

Zerodha grew to over 1.6 crore users with no advertising, relying on word of mouth, the Varsity education platform, and customer trust. The founder has said depending on advertising is like an addiction that captures your growth. Roughly 30 percent of investors came through referrals from existing users even after referral incentives were removed. It matters in this round because any re-activation idea that needs an ad budget or a paid push is, by definition, off the table, and proposing one signals you do not understand the company.

SEBI tightened futures and options with true-to-label charges, larger lot sizes, and weekly expiry rationalisation, which cut speculative trading volumes across Indian broking. A large slice of recently dormant users are former active F&O traders who pulled back because the economics and structure of their trading changed, not because they disliked the app. Recognizing this lets you segment correctly and avoid proposing re-activation tactics that assume people simply forgot about the platform.

It is primarily strategy at a mid-level altitude: framing the problem, segmenting users, choosing a defensible target, and reasoning under hard constraints. Execution shows up in how concretely you sequence a first experiment and define its metric and attribution, but you are not asked to write a spec or design a screen. The interviewer cares more about whether your plan is coherent, low-cost, trust-safe, and measurable than about feature polish.