Nine weeks on a pre-foreclosure, then the servicer approved a modification

The cure rate is the assumption your whole model rests on, and the public numbers show how big it is. ATTOM put first-quarter 2026 filings at 118,727, with starts at 82,631 and completed bank repossessions at 14,020. Starts were up 20 percent year over year and repossessions up 45 percent. Compare the two figures inside the same quarter and roughly one start in six ends in the bank taking the house. The other five resolve somewhere upstream through reinstatement, modification, a sale to a retail buyer, a sale to an investor, or a filing that gets withdrawn. Timing lags mean those are not the same loans, so treat the ratio as a scale indicator rather than a conversion rate.

So a pre-foreclosure list is a list of people whose most likely outcome is keeping the house or selling it to someone other than you. If you underwrite your marketing spend on the assumption that a contract equals a closing, your cost per acquisition is wrong by a multiple, and the fix is to price your mail against closings rather than against signed contracts.

Two things vary by state and should be confirmed locally before you build process around them. Reinstatement and redemption rights, meaning how late in the process an owner can bring the loan current, differ by state and between judicial and non-judicial foreclosure, and so does how much of the default record is published and where. A real estate attorney licensed in your state should be the one telling you what a signed contract does when a borrower cures.

The filing pipeline has now run more than twelve straight months of annual increases, so the lead supply is expanding. Volume upstream is not the same as deeply discounted property reaching you.

1,350 for the lesson that most defaults never make it to a sale is cheap. I paid more than that in holding costs learning something dumber. The part I want to steal is asking about the modification application on the first call. I have made maybe forty of those calls and it never once occurred to me that the owner might already be halfway to a solution that has nothing to do with me.

Run your own numbers forward and the picture gets less painful. 300 letters at roughly a dollar all in with the data is 300 a month. Six conversations, one contract, zero closings this cycle. If one in six contracts closes and one in six calls contracts, you need something like 36 conversations for a closing, so 1,800 pieces of mail for a 12k assignment fee at your spread. That is a workable ratio and a terrible cash flow pattern, because the 1,800 pieces go out over five or six months and the fee arrives once. What killed you was not the spend, it was that the spend was too small to produce a second file while the first one was dying. Either mail more per cycle or accept multi-month gaps and fund them.

Your 450 inspection was the only line item I would keep as-is. A soft roof on a 289 comp changes the number by a lot more than 450.

The thing you could not have seen is the servicer file. A modification in underwriting leaves no trace in the county record, so title comes back clean and the notice of default still sits there looking like a live opportunity right up to the day it gets rescinded. Some investors record a memorandum of their purchase contract to put the world on notice of an equitable interest, and whether that is permitted and what it actually accomplishes varies by state, so ask a local attorney before you do it. It also makes some sellers hostile, which matters when your leverage in the deal is entirely that the person likes you. I would rather ask the owner for the name of their single point of contact at the servicer and the date of their last written communication, and then ask again a week before closing.

Nine weeks and 1,350 is the cheapest version of this story I have read here. Mine involved a dumpster.