Explaining a KYC Submit-Rate Rise From 25% to 35% Among New Business Signups

A fintech platform for businesses tracks its KYC submit rate: the share of new signups that go on to submit Know Your Customer (KYC) verification documents, defined as KYC submits divided by signups. A time series of this metric shows it rising from 25% to 35%. The report gives only the two levels. It does not say over how long the rise happened or whether it was a step or a gradual climb, so those are clarifying questions below. ### Clarifying Questions - Over what period did the rate move from 25% to 35%, and was it a sudden step or a gradual climb? - Is the rate computed by signup cohort (submits among the accounts that signed up in a period) or by calendar period (submits in a week divided by signups in the same week)? - Does a submission count at any time after signup, or only within a fixed window? - Did the absolute numbers of signups and of KYC submits each rise, fall or stay flat? - Is the rise treated as good news to be explained, or as a suspected problem? ### Part 1 — Core hypothesis and the data to check it What is your core hypothesis for the increase, and what data would you pull to verify that it is real and to locate where it comes from? ```hint Two moving parts A ratio can rise because its numerator grows or because its denominator shrinks. Look at both series before reading the ratio. ``` ```hint Who versus how often Ask whether the mix of who signs up changed, or whether the same kinds of signups now submit more often. ``` #### What This Part Should Cover - A clear primary hypothesis and how it would be confirmed or rejected - Checks of the metric definition, logging and cohort maturity - A decomposition into numerator versus denominator and mix shift versus within-segment change - The specific tables, events and cuts to pull ### Part 2 — Finding the cause after the obvious checks pass Assume you have looked at the data: the increase is real, there is no logging problem, and no segment shows an obvious anomaly. How do you keep looking for the cause? The interviewer pointed the discussion toward two groups of causes: changes the company itself may have made that led more users to submit KYC, and external factors outside the company's control that could make users more willing to submit. ```hint Line up the dates Put the start and shape of the rise next to a dated list of everything that changed in the signup and verification funnel. ``` ```hint Who would an outside event touch For an external cause, ask which customers it would affect more than others, and whether those customers moved more. ``` #### What This Part Should Cover - Internal levers across the funnel, each with a way to test it - External drivers, each with a testable signature - A method for attributing the change, using timing, dose-response and comparison groups - Whether the rise turns into verified, transacting customers ### What a Strong Answer Covers - Clarifies the metric definition and time frame before hypothesizing - Separates measurement artifacts, mix shift and genuine behavior change - A structured, prioritized hypothesis tree with a test for each branch - Quasi-experimental reasoning (timing, comparison groups) rather than correlation alone - Ties the conclusion to business value, such as verification approvals and transaction activity, not submissions alone ### Follow-up Questions - Suppose the number of signups fell while the number of KYC submits stayed about flat. How does your conclusion change? - How would you tell whether a new KYC flow created additional submissions or only brought forward submissions that would have happened later? - If the submit rate rose but the share of submissions that pass verification fell, what would you investigate? - How would you set up monitoring so that the next move of this size is explained within days rather than weeks?