INSIGHTS · LAUNCH

What has to be true before a lending product launches

Most lending launches that struggle do not fail on the product. They fail on the parts around it: funding, credit policy, decisioning, collections and the economics that tie them together.

·

A new lending product usually starts as a customer proposition: a better way to finance a car, a purchase, a working capital gap. That is the right place to start, and it is also the part of the launch that tends to get the most attention. The parts that decide whether the product survives its first year get less of it. They are less visible, they sit across several teams, and most of them only show their weaknesses once real loans are on the book.

Before the first loan is issued, we look for a short list of things to be true. None of them needs to be perfect. All of them need to exist, have an owner, and be connected to each other.

Funding that matches the product

Capital is not just an amount. It comes with a tenor, a cost, eligibility criteria, concentration limits and reporting obligations. A product with long tenors funded by short-term money, or a product whose target customers fall outside the eligibility criteria of the facility, has a structural problem before it has a single customer. The funding conversation should shape the product design, not follow it.

It also helps to know early what a funder will want to see after launch: which data, how often, and in what format. Building that reporting in from day one is far cheaper than reconstructing it from operational systems later.

A credit policy that someone has written down

Every lender has a credit policy. Not every lender has one that is written, versioned and understood by the people who apply it. The policy should say who the product is for, who it is not for, what evidence is required, which limits apply and who can approve exceptions. Without that, the first months of lending become a series of individual judgements, and the portfolio that results is hard to explain to anyone, including the team that built it.

Decisioning that can be changed safely

The first version of any scorecard or rule set will be wrong in places. That is expected. What matters is whether it can be changed quickly, whether changes are tested before they go live, and whether every decision can be traced back to the inputs and the rules that produced it. A decision engine that only one engineer can modify is a launch risk, however good its first version is.

Collections designed before the first loan

Collections is often treated as something to set up once arrears appear. By then the important choices have already been made: what the loan agreement allows, what contact data was captured at origination, which payment methods are available, how reminders are sent. A lender that designs its early collections journey alongside the origination journey recovers more and spends less doing it.

Unit economics that include the uncomfortable lines

A launch model usually captures pricing, cost of funds and an expected loss rate. The lines that are easier to leave out are acquisition cost by channel, servicing and collections cost per loan, fraud losses, and the effect of early repayments on revenue. Those are often the lines that decide whether the product makes money at scale. They should be in the model before launch, even if the first estimates are rough, so that the business knows what to measure from the first week.

A short pre-launch checklist

  • Funding terms, eligibility and reporting are agreed, and the product fits inside them.
  • The credit policy is written, approved and owned by a named person.
  • Decision rules can be changed, tested and audited without a release cycle.
  • The early collections journey is designed and the loan agreement supports it.
  • The economic model includes acquisition, servicing, collections and fraud costs.
  • The first management report is defined, so that the team knows what to look at on day one.

The point of all this

None of this slows a launch down when it is done early. It replaces the expensive kind of learning, where problems surface in the portfolio months later, with the cheap kind, where they surface in a workshop before the first loan is issued. The product still matters most to the customer. Everything around it is what lets the business keep offering it.

MORE INSIGHTS

SCALE

Scaling a loan book without losing credit quality

Read the note

DECISIONING

AI in credit decisioning: where it earns its place

Read the note

Working through this right now?

Tell us what you are launching, scaling or rebuilding.

Discuss a growth opportunity

or write to [email protected]