Why this role exists

Collections is one of the highest-leverage parts of a lending business. A 100 bps reduction in roll rate at DPD30 is worth more than most growth experiments. The non-obvious things are still on the floor: tighter early-bucket nudges, smarter agent dispositions, settlement logic that doesn't leak margin, voice-bot strategies that actually move the needle, and the right thing to do when a 21-year-old borrower can't pay.

This PM owns that surface. We're not looking for someone to run the existing roadmap. We're looking for someone who'll find the next 20% of recovery — and the next playbook none of us have written yet.

What you'll own

You'll be accountable for outcomes, not feature ship dates. The big ones:

- Recovery and roll rate across DPD buckets (DPD7 through DPD180+), particularly the early buckets where unit economics live or die.
- Cost-to-collect — every rupee recovered should cost us less over time. You'll trade off automation, agent capacity, and channel mix.
- Agent productivity and experience — the collections agents are your power users. You'll sit with them, watch them work, and rebuild their tooling around what actually slows them down.
- Borrower experience in distress — collections is where our brand is most fragile. The product you build has to recover money and leave borrowers willing to come back when their next paycheck lands. This is harder than it sounds.

You'll work cross-functionally with Risk, Data Science, Collections Ops, Engineering, Compliance, and Customer Support — none of whom report to you, all of whom you'll need to move.

What success looks like at 6 months

- You have a written, defensible point of view on where the biggest recovery upside lives — backed by your own data work, not someone else's deck.
- At least one experiment in market that meaningfully moved a roll-rate or recovery metric.




- You've shadowed collections agents enough that you can describe their day in detail. You've fixed at least one thing they were quietly suffering through.
- Risk and Ops trust you. They tell you things they wouldn't have told the previous PM.
- You've made at least one decision your manager disagreed with at first, and were right.

What we're looking for

We're filtering hard for a small set of behaviors. Pedigree, brand-name companies, and lending experience are not on the list.

You take problems, not tasks. When you see a vague goal like "reduce DPD30 roll rate," your instinct is to go figure out what's actually driving it — by reading the data, calling borrowers, sitting with agents — not to ask for a PRD template. You write your own brief.

You operate without a playbook. You've worked in a place where the right next step wasn't obvious and no one was going to hand you one. You built it. You're more comfortable in ambiguity than most.

You ship, then learn. Your default is a v0 in two weeks, not a perfect spec in three months. You'd rather have shipped two flawed experiments than written one beautiful PRD. You know the difference between rigor and theater.

You have strong data instinct. You don't need to write SQL — but you can stare at a cohort table and tell us which numbers are lying. You know what "looks weird" looks like. You ask analysts the right questions and you don't accept hand-wavy answers.

You care about users you'll never meet.



Collections users are people in financial stress. The valuable PMs we've worked with treat that with seriousness, not pity, and not detachment. You should be able to design flows that recover money without making a 22-year-old feel like they failed at being an adult.

You write clearly. Your thinking shows up on the page. We will ask you for a writing sample.

What we are explicitly not looking for

- Lending or fintech experience. Helpful, not required. We will train you on credit, DPD, regulation, and collections strategy faster than you can train yourself on judgment.
- An MBA or a top-tier-consulting background. Neither helps nor hurts.
- Someone who can "manage stakeholders." You'll have to, but if that's the headline skill on your resume, this isn't your role.
- Big-company PMs whose biggest win was getting a feature through review. This is a small-team, high-ownership seat.

How we'll evaluate you

Four conversations, designed to test the things above:
1. A real problem from the collections roadmap. We'll give you a one-paragraph problem statement and 48 hours. You'll come back with a brief — your framing, the data you'd want, your top three bets, what you'd ship in the first two weeks. We're testing for clarity of thought, not the "right" answer.
2. A working session. We'll sit with you and pressure-test the brief. We want to see how you update when you're wrong.
3. A craft conversation. Walk us through a product decision you got right that no one would have noticed if you hadn't made it. And one you got wrong.
4. References we choose, including someone you've worked alongside (not just a manager). We'll ask them about ownership and judgment, not personality.

We don't do brain teasers, estimation puzzles, or "design Instagram for dogs."

📌 Product Manager (Karnataka)
🏢 mPokket
📍 Karnataka