What's the smallest real version of this? Costs, insurance, software, all of it

Before the numbers, decide whether you want to own the project or fund one. Those are different products. If it's the funding side, minimums with a sponsor doing conversions typically start around $50k to $100k and the work is diligence rather than construction. Doing it yourself, the floor is much higher than a few hundred thousand once you count carry.

The smallest real conversions tend to be single story commercial buildings of roughly 5,000 to 12,000 sf becoming four to ten units. A former dentist office, a small retail bay, a light industrial shed. At that size you're still looking at a total project cost in the low millions in most markets, because conversion construction is generally quoted per unit in six figures. A few hundred thousand dollars can be the equity in a project like that if you can find a lender for the rest, and it isn't the whole cost.

On your specific questions. Pre purchase feasibility usually means paying an architect for a test fit and a code review, plus sometimes a structural or mechanical engineer walk. Budget in the low tens of thousands for a small building, and treat it as money you may spend and then walk away, because that's often exactly what happens. Full design fees later run as a percentage of construction cost, commonly high single digits to low teens for a conversion.

Insurance during construction is builder's risk, which covers the building and materials while work is underway, plus a vacant building or vacancy endorsement, because a standard landlord policy generally excludes losses in an unoccupied building. You also want general liability, and your lender will name itself on the policy. Availability and pricing vary by state and by building, so get a broker who has written a conversion before.

You generally don't need a contractor license to own a project, you hire a licensed general contractor. Whether an owner can act as their own builder depends on your state's licensing rules, and some states restrict it sharply, so check your state board.

One thing that catches people at this size is the appraisal. A converted building with no rent history in a market with few comparables can appraise below cost, which changes how much a lender will actually fund.

The "money you may spend and walk away" line is the part I hadn't priced. So realistically you look at five buildings, pay for feasibility on two, and buy one? That's a real cost of doing business that never shows up in a pro forma.

@wren yes, and that's true in land too. I budget dead deal costs as a line item. Survey, title work, a consultant walk, and then you don't buy it. Patience is cheaper than the wrong building.

On the licensing bit, worth separating two things. Your GC needs a license and so do the mechanical, electrical and plumbing subs, and in most places those trade licenses are the ones the inspector actually cares about. Whether you personally can pull an owner-builder permit varies a lot by state, and in some it comes with the condition that you can't sell for a year or two.

Lending floor is a real constraint at your size. Construction lenders often have a minimum loan around $2m to $5m and won't look at a $900k draw facility. Below that you're talking to community banks or private lenders, and private construction money prices well above bank debt with points on top. Ask a couple of local banks what their smallest construction loan is before you go building a model, and get their terms in writing.

Add lease-up cost to the list. Eight units all delivering the same month means marketing, screening, and someone answering the phone. If you self manage that's your time. If you hire it out, expect a leasing fee plus ongoing management, often in the 6 to 10% of collected rent range for small residential, and small buildings sit at the top of that range because there's no scale.

Also budget for the entitlement side even if you never need a rezoning. A zoning verification letter, a survey, and a couple of hours of a land use attorney's time to read the district and the parking requirement. Cheap relative to the mistake it prevents.

Software, since nobody's answered that yet. Costar is the full commercial data set and it's an enterprise contract, usually four figures a month and they won't quote a single user casually. Crexi and LoopNet are the free-to-cheap tier for listings, Reonomy and similar owner-lookup tools run maybe a couple hundred a month. County GIS and assessor sites cost nothing and are where I actually spend my time.

@tally I'd add that for construction tracking people use Buildertrend or CoConstruct at roughly $100 to $400 a month, and honestly on a six unit project a shared spreadsheet and a phone do most of it. RSMeans for cost data is a subscription too, though your GC's estimator has better local numbers than any book.

Reading @tender and @rena-guide together, the floor sounds like it's set by the lender rather than by the building. You can find a $400k dead retail bay all day. Nobody wants to lend $1.5m to convert it.

@kestrel that matches what I'm seeing while building my list. The small buildings that pencil tend to be bought by people who already have relationships at a local bank, or who bring most of the money themselves and refinance after. Which is another way of saying the first conversion isn't usually somebody's first deal.

And the six figure per unit construction number is the one I'd stress test hardest before you get attached to any of this. At eight units that's most of a million dollars of hard cost on a building you bought for four hundred grand. If your finished rents are ordinary suburban rents, that arithmetic just doesn't close, and no amount of subscription software finds you a building where it does.