Personal-loan marketplace

An applicant could finish a loan form and still reach a lender whose policy did not fit. The product needed to make lender fit part of the journey without asking the borrower to understand each bank’s requirements.

I chose a common first application rather than a separate journey for each lender. Income band, bureau score, employment type and geography could be captured once and compared with partner eligibility. The alternative was to move policy differences into the borrower’s task: choose a lender first, discover its requirements and repeat work if the fit was wrong.

I established lender-wise MIS and a daily operations view so application progress, partner service levels, payout reconciliation and disputes could be discussed together. Weekly reviews used stage-level drop-offs to decide what needed investigation. A submission acknowledgement, an underwriting decision and a partner payout answered different questions and could not substitute for one another.

The documented delivery connected a shared application and eligibility routing to lender onboarding, reporting and commercial follow-through. Lenders retained the credit decision; the product team retained responsibility for understanding the handoff. Available programme figures do not isolate this design or an individual review change, so no release-specific uplift is calculated.

The lesson I carry forward is concrete: a common application helps only if someone can explain whether the case reached its lender, what stopped it and what the partner owes next. That follow-through belonged alongside the product, rather than becoming an unowned queue after launch.

The shared form could not make the lenders interchangeable. Eligibility, the KYC handoff, service commitments and rejection feedback remained lender-specific. I ran those partner conversations alongside engineering’s work on the application, with growth connecting the journey to CashKaro rewards surfaces. That joined the customer-facing entry point to a destination the team could actually operate.

This was a choice about where to handle complexity, not a claim that the platform made a credit decision. Lenders retained review and approval. My responsibility covered the application, routing, partner commercials, operations and P&L; the product boundary extended to understanding what happened after submission.

That operating model was more demanding than launching a form and treating the lender handoff as completion. It required engineering, growth, lender teams and operations to keep a case understandable after it left the interface. The record describes the review cadence; it does not include a shareable experiment showing a particular hypothesis, intervention and isolated effect.