What a processor's follow-up actually costs you per file

Cost per loan is the number that decides a thin year, so it attracts a lot of bad arithmetic. Most of it starts by pointing at the total cost to originate, which is the wrong denominator for anything a piece of software does to a file.

Here is a denominator you can check against your own payroll instead.

What the follow-up costs

A processor carries somewhere around 15 to 20 files a month. Fully loaded, salary plus taxes and benefits, that processor costs a lender roughly $85,000 to $95,000 a year. Divide it out and you get somewhere near $400 of processing labor sitting inside every file, and the largest single piece of that is borrower-facing follow-through: requesting documents, opening what comes back, catching the wrong ones, asking again, filing the right ones.

That is the line Loandock works on, and it is the only line it should be measured against. Not the cost to originate, which includes commissions and overhead and a dozen things no software touches.

The more useful question is capacity, not cost. If the chase comes off a processor's desk, the ceiling on files per processor moves, and the next hire is the one you do not make at the next volume spike. What that looked like on live files, hour for hour, is on the results page.

Why the industry cost-to-originate number misleads you

The benchmark everyone quotes bundles commissions, corporate overhead, occupancy, licensing, third-party fees, and every salaried person who touches a loan. Software that works borrower follow-through touches almost none of that.

Measuring against it cuts both ways, and both ways are bad for you. A vendor can claim a share of savings it has no mechanism to produce. And a skeptical CFO can disprove the claim in one line of arithmetic, at which point every other number in the deck gets re-audited. If you are going to take a number into a budget meeting, take one your own payroll can back.

Where the hours actually go

Almost none of the follow-through cost is judgment. It is round trips.

The condition is issued. The request reaches the borrower whenever the processor gets to that file. The borrower uploads after work, because that is when people do paperwork. Nobody opens it until the next sweep. It is the wrong document, or the right document for the wrong period, and the correction request waits for the sweep after that. One condition, several days, and nobody did anything wrong.

Stack that across a full condition list and you have the gap between conditional approval and clear-to-close, most of it spent waiting rather than working.

What to measure instead

Do not ask what percentage came off your cost to originate. Ask three things you can watch on your own files.

How long it takes for a condition to reach the borrower after underwriting issues it. How many round trips a wrong document costs before the right one is on file. And how many files a processor can carry before the follow-up starts slipping past the same-day mark.

Those three move first, and they move where you can see them, in the LOS, without a spreadsheet built to make a vendor look good.

What Loandock actually does to that line

Loandock works the borrower-facing follow-through. It requests the documents, checks what comes back against what the condition asked for, rejects the wrong ones with the reason, files the right ones, and writes the result back to the LOS with the timestamp and the document attached.

Loandock marks conditions fulfilled; your underwriter clears them.

That is one line on your cost sheet, not the whole sheet, and it is the line your processors are carrying today. Judge it there.