The decision to keep the deal rather than sell the lead is itself the data point. When you built the platform to test deal flow and the first real appointment pencils as a flip, you have confirmed two separate things work: the lead gen and your own acquisition judgment. That is worth separating cleanly in your head.
The assumption doing the most work in your numbers is the $28k rehab figure. On a house owned by someone in her 70s, deferred maintenance tends to compress into systems, roof, HVAC, plumbing, and electrical, and those categories do not always surface in a walkthrough. The spread between $89k acquisition and $148k ARV is $59k gross before holding, closing, and carrying costs. If rehab runs to $42k instead of $28k, your margin compresses by roughly $14k before you account for anything else. That is still a deal, but it is a tighter one than the headline numbers suggest.
The risk you did not mention is scope creep on the platform side. You built this to generate your own deal flow. Now that you are also the operator on the resulting deal, you are running two parallel workloads. The platform needs iteration, and a live rehab pulls attention. That tension is manageable, but it is real, and it tends to show up around week four or five of a project when decisions start compressing.
On the platform itself, the next question is repeatability. One appointment from one form tells you the channel works. It does not yet tell you the cost per qualified lead, the conversion rate from form submission to appointment, or whether the Whitehaven market density can sustain consistent volume. Those metrics are what separate a working test from a defensible platform asset.
What does your current lead volume look like behind that one appointment, and how many submissions did it take to get there?