From the founder

3D Asset Mapping Cost Seg Engine > Top Down Estimator Studies

An estimator guesses the whole asset. We verify everything public first, then estimate only the small part that is left.

Bola Akinsanya · Founder, Unlevered· July 14, 2026

When I got my own cost segregation study back, the thing that bothered me most was not the price. It was that there was no math. Six thousand six hundred dollars for a stack of paper I could not trace, and the entire method underneath it was a template. Someone had taken a spreadsheet, dropped my purchase price in, and split it into buckets by percentage. That is the whole method. Land is some percent, the building is some percent, the short-life property is some percent, and the percentages came from a library of other houses, not from mine.

I want to explain why that approach is the weak version, because it is the version most people are paying for.

An estimator looks at every property and makes one core decision: how the dollars are allocated. It assumes your house behaves like the average of the houses it has seen. And for the boring parts of a building, that assumption is fine. The walls, the framing, the shell, the rough plumbing, the electrical rough-in. Those really do move together, house to house, and a percentage gets you close.

But the moment you have remodeled, the average stops describing you. My doors are not the builder’s doors. My appliances cost more than the plumbing that was originally put in the house. The furniture I stocked my rentals with holds far more usable value than a template would ever assign to it, because a template does not know it is there. It was not in the MLS photo when the house sold, so as far as an estimator is concerned it does not exist. Every one of those decisions is a real dollar sitting in a real short-life category, and the estimate rounds all of it off.

The moment you have remodeled, the average stops describing you.

Actualizing is the opposite move. Instead of starting from a percentage, you start from the property record and you rebuild the asset. If your house was built legally, its records are public. The parcel, the assessor’s land and building split, the permits, the dimensions, the scope of every remodel that pulled a permit. We pull all of it. Then we lay the photos and the receipts on top, and we rebuild the house as a data map, room by room, down to the square footage. Every component goes in its own box. Then, and only then, do we assign value, and the value is not a guess. It is logic, rebuilding the house from what we actually know about it.

This is where the uncelebrated part lives, and it is querying. Database architecture is not a glamorous sentence. But if you understand it, you can take a house, break it into every component it contains, put each one in its place down to the square foot, and then apply hard, repeatable logic to decide where every dollar belongs. The value drivers are exactly the things a template throws away. The furniture you brought in. The light fixture that replaced the original one from the listing photo. The finish upgrade nobody wrote down. Those are all things we can extrapolate, categorize, and then make a firm, defensible decision about.

The value drivers are exactly the things a template throws away.

Underneath that sits a component library that took the better part of nine months to harden. A socket for every component and composite that could reasonably exist in a building. And then, for each asset type, each recovery period, each judgment call, we wrote and rewrote the queries by hand, over and over, until the basis they produced could stand up to an audit. That is the part you cannot buy off the shelf, and it is the part that actually differentiates the work.

Here is the example that made all of this concrete for me. A study crossed my desk where the preparer had assigned a set of movable structures to the land. A detached gym. A sauna. The gym did not even have electricity run to it. The sauna was a sauna. Both units could physically be picked up and moved with the owner. But because the preparer estimated instead of rebuilding, they folded the gym into the house and the sauna into the house and moved on. Together those two units cost about a hundred thousand dollars. Treated correctly, as relocatable personal property, they are five-year assets, not part of a thirty-nine-year shell and not buried in nondepreciable land. That is not a rounding error. That is a six-figure component sitting in the wrong bucket because nobody built the house from the ground up.

Now I want to be honest about the other direction, because this is where the internet version of this story usually cheats.

We ran our rebuild against a national template on the same 650-square-foot unit, placed in service in a 40 percent bonus year. Same property, two methods. The template landed on a depreciable basis of about $176,977 and split it roughly 18 percent five-year, 12 percent fifteen-year, 70 percent shell. Our rebuild landed on about $194,344 and split it 17, 11, and 71. Look at those splits. They are nearly identical.

On a single small unit with no dramatic remodel, the template and the rebuild converge, and the honest headline is that our year-one deduction came out to about $28,838 against their $27,441. A real difference, but a modest one.

So why do it the hard way for a result that close?

Two reasons. The first is the roughly $17,000 of additional depreciable basis, which came almost entirely from allocating less to land, because the rebuild could actually see what belonged to the building. The second matters more. Every dollar in our version is a sourced component with a citation and a photo behind it, not a scaled estimate reconciled to fit a basis. Same neighborhood of an answer. A completely different thing to defend. When the small units converge, sourcing is the whole edge. And when the property is not small and simple, like the one with the misfiled gym and sauna, the gap stops being modest and starts being the difference between a right answer and a wrong one.

Same neighborhood of an answer. A completely different thing to defend.

That is the line I care about. This is cost segregation, not categorization. Categorization sorts a house into buckets that already existed. Cost segregation rebuilds the house and lets the buckets fall out of the evidence. One is an estimate dressed up as a study. The other is the study. As this strategy moves from something a few people whisper about to something high earners expect their advisors to know, the difference between those two is going to matter a great deal, because getting it approximately right and getting it actually right are not the same, and only one of them holds up when someone asks to see the math.

A note on all of this. A cost segregation study accelerates deductions, it does not create cash, and a deduction only helps if you can actually use it under the passive activity rules. The year-one figures above assume a 40 percent bonus year and a single small unit; your facts, your bonus percentage, and your state’s treatment will differ, and year two is always smaller. The two reviewed studies are described with identifying details removed. This is a founder’s explanation of our method, not tax advice. Talk to your own tax professional. See our full disclaimer.