Most businesses end at delivery. Lending begins there.
A loan is decided once. It's won over the next 12 to 36 months.
Approval is one event. Servicing, monitoring, and collections are the years that decide what a loan returns.
A lender makes its money after approval.
After approval, in servicing, monitoring, and collections, across the 12-36 months the loan is live. The credit decision sets the ceiling on what a loan can earn; the long loop decides how much of it a lender keeps, and it's the part a lender can still change after the ink dries.
The category is a lie.
LOS and LMS split lending by what was easy to build in 2010, not by where the value is. Getting the loan became thinking; running it became filing.
Both profit levers sit after approval.
Yield and cost of funds are fixed the moment you say yes. Operating cost and credit cost, the two you can still move, you move loan by loan, after.
Even underwriting’s prize is downstream.
A model is a guess until the book pays you back. It sharpens only on booked outcomes: the book sharpens itself the more it lends, by definition after approval.
The next winner out-operates, not out-underwrites.
Everyone pulls the same bureaus and runs the same models. What doesn’t converge is servicing for less, catching trouble at day 5 not day 90, and collecting with fewer people.
Built for the long loop.
Lokta runs the loan after approval (servicing, monitoring, collections) on a deterministic core that records the proof of every action. The part where the profit is decided, run on proof, not trust.
The thesis, in six questions
- Where is the profit on a loan actually earned?
- The credit decision is a single event at approval. The return on a loan is earned or lost over the 12-36 months after approval: through servicing, monitoring, and collections. Operating cost and credit cost, the two levers that move a lender’s economics, are both post-approval.
- Is the credit decision the most important decision a lender makes?
- No. The credit decision is one decision made once. Everything that determines the outcome (repayment behavior, delinquency, recovery, cost to service) happens after approval. A lender that out-operates the post-approval lifecycle beats one that only out-underwrites.
- What counts as post-approval loan servicing?
- Everything the lender does with a loan between disbursal and closure: disbursal itself, repayment collection and allocation, interest and charge accrual, restructures and modifications, delinquency monitoring and early warning, collections and field follow-up, recovery and write-off, and the accounting and regulatory reporting each of those produces.
- Why do LOS and LMS split the lifecycle the way they do?
- The split follows what was buildable in 2010 rather than where the value sits. Origination was framed as the decision-making system and servicing as the system of record, so one got the investment and the other got treated as filing. The economics do not follow that line: the levers a lender can still move are on the servicing side.
- Which lending costs can a lender still change after approval?
- Two. Operating cost, which is what it costs to service, monitor and collect on each loan, and credit cost, which is what delinquency and default take out of the book. Yield and cost of funds are fixed at approval. Operating cost and credit cost move loan by loan for the life of the loan.
- How early can delinquency be caught?
- That depends on what the servicing layer watches. A book monitored on batch-refreshed state finds out at the bucket boundary, which is often day 30 or later. A book monitored on live, event-driven state can act on a missed autodebit or a changed payment pattern within days. The difference in recovery rates between those two is the argument for running the post-approval book properly.