The apartment unit count sweet spot, according to an operator who built a 5,000-unit Southeast portfolio and then sold it, is 100 to 250 units. Below roughly 75 units you can’t afford a property manager on site. Above 250, the property stops being a bigger version of the same business and becomes a different one.
Michael Crow of Soreste Capital Group learned the ceiling the expensive way: a 500-unit Atlanta property on 38 acres, bought off market with seller financing, that consumed a significant amount of his own capital and ended in a total loss after 18 months. His conclusion — “250 units times two is not the same as 500 units” — is the argument this article works through.
Below: where the floor and ceiling actually sit, what breaks at scale, how Crow used terms instead of price to buy off market, how he picks markets, and why his build-for-rent land positions never penciled.
Key takeaways
- Around 75 units is the practical floor for affording a full-time on-site property manager; below that, the operating math doesn’t support the position.
- Crow’s best returns came from 100–250 unit properties. Today he’d target 75–150 units with about 200 as a top end.
- A 500-unit property on 38–40 acres is not two 250-unit deals. Crow lost his entire investment on one and says he will not do that size again.
- When buying off market, tell the seller they can set the price or the terms, but not both — then use seller notes, holdbacks and adjustments to price identified problems instead of arguing over price.
- Politics is the risk investors forget: rent control and eviction restrictions can make an otherwise sound property unprofitable, which is why Crow built in the Southeast rather than New York.
From the Real Estate Pros Show
This article draws on an interview with Michael Crow of Soreste Capital Group / Dunross Capital Partners on the Real Estate Pros Show, hosted by Joseph Meacham.
The 75-Unit Floor and the 250-Unit Ceiling
The floor is set by staffing. “You don’t get a lot of efficiencies unless you get up to around 75 units,” Crow says. “Then you can afford a property manager on site. If you get below that, you really can’t.” That single threshold explains why so many 30- and 50-unit deals underperform their models: the owner either pays for management the property can’t support, or does it themselves and eats the time.
The ceiling is set by control. Across a 5,000-unit Southeast portfolio, Crow’s best realized returns came from properties in the 100 to 250 unit band. Anything larger, in his words, meant “we had our work cut out for us.”
What he would buy today is tighter than what he bought then: 75 to 150 units, with roughly 200 as a top end. The narrowing reflects a preference for control over scale — one asset a single strong manager can genuinely run.
Property age matters as much as unit count. His filter is vintage that is old enough that there’s value to add, but not old enough to have fundamental problems. That’s a real distinction in underwriting. Cosmetic and unit-turn value-add is a business you can execute on a schedule. Structural, plumbing-stack or envelope problems are a capital sink that no rent bump covers.
Layer the rest of his criteria on top: a good market, sound fundamentals, good schools. Nothing exotic. The point is that unit count is the first filter, not the last, and getting it wrong makes every other good decision harder.
Why 500 Units Is Not Two 250-Unit Deals
Crow decided to do bigger deals. His reputation meant sellers brought him product before it hit the market, and one of them brought a 500-unit Atlanta property on roughly 38 acres that needed work. Single owner, willing to seller-finance the whole thing. “I’ll make it easy,” the seller told him.
He put a significant amount of his own capital into it and brought in a couple of partners. The problems started immediately — not with the financing, but with the physical and operational reality. “Right from the start, we started to find out that there’s a big difference in trying to manage 40 acres of apartments and 500 units,” he says. The timing didn’t help; they were coming out of COVID.
Eighteen months of putting money in followed. Then the decision every operator eventually faces: fix it or stop. “I got to make a decision. Either we’re really going to fix this and something’s going to change, or we need to take our loss and move on.” He took the loss and lost the entire investment.
The lesson is structural, not situational. Two 250-unit properties give you two managers who each know every tenant, two rent rolls, two maintenance teams sized to a walkable footprint. One 500-unit site on 40 acres gives you what Crow calls “some big city that you have to manage.” Vacancy hides in it. Deferred maintenance hides in it. Nobody knows every resident.
He is unequivocal about repeating it: “I will not do those again.”
My lesson was, 250 units times two is not the same as 500 units in a deal. I’d rather have two separate properties with a manager that knows every tenant, not some big city that you have to manage. And I will not do those again.
— Michael Crow, Soreste Capital Group / Dunross Capital Partners
Deal Structuring: Price or Terms, But Not Both
Crow bought every property in his portfolio off market, and his lever was almost never price. His line to sellers: you can set the price or the terms, but not both.
“You tell me the price, I’ll tell you what my terms are.” That reframing does two useful things. It stops the negotiation from becoming a single-number standoff, and it moves the conversation onto ground where a well-capitalized, credible buyer has an edge — structure.
The tools he used inside that structure:
- Seller notes — the seller carries paper, which lowers the equity check and often gets a higher headline price agreed to
- Holdbacks — cash withheld at closing against a specific, identified condition or performance issue
- Adjustments — pricing a known problem into the deal terms rather than relitigating value
The principle underneath all three: when you find a problem in diligence, don’t argue about what the building is worth. Identify the problem, put a number on it, and propose a mechanism that protects you if it turns out worse than expected. “You can identify a problem, have a reasonable approach to it, and with a seller sit down and work out a solution that works for both sides,” he says. “I think that makes deals.”
His own critique of that strength is worth more than the technique. “Your greatest strength is usually your greatest weakness.” He made a lot of deals and underweighted operations — and specifically paid a price for not moving quickly enough on renovating units. Structuring gets you in. It does not turn units.
Choosing Markets and Reading the Cycle Instead of Timing It
Crow’s thesis was formed in 2015–2017: buy Southeast apartments. He executed it to 5,000 units, bought aggressively during COVID while others panicked, and started selling as rates climbed. Since then he has been waiting for prices to stabilize rather than guessing at a bottom.
“People try to time the bottom. You really can’t do that in real estate. Everybody thinks they can and nobody can until you look backwards.” What he does instead is watch whether the numbers still work and act accordingly — foot on the gas or foot on the brake.
His current read on financing: stop waiting. “We’re not going back to two and three percent interest rates on the long-term money.” The adjustment is to underwrite at today’s cost of capital, require actual cash flow, and let prices come to that math.
Market selection is population-driven, not price-driven. South Carolina, Florida, Georgia and Texas grow, and growth creates demand for the product you own. “If you start investing in markets where there’s no growth, it’s really difficult to make money long term. You might get an opportunity to buy something right and flip it,” but the long hold doesn’t work.
The risk he thinks investors most often ignore is political. “The biggest risk in real estate that people forget about is politics.” Rent control and eviction restrictions can make a well-bought property unprofitable regardless of how you underwrote it, so he stays out of jurisdictions likely to impose them.
One current dislocation: Texas apartments are, in his words, “kind of a mess right now” — high-growth markets where some operators underperformed, which is exactly where he’d look.
Where the Build-for-Rent Math Stopped Working
Crow assembled land for build-for-rent development in Huntsville, Alabama and on the coast of Georgia. After several years, he couldn’t get the numbers to work. He sold the land at a loss.
The reasoning behind selling rather than holding is the part worth copying. “Sometimes you got to take your losses quickly and move on to something else, because you don’t think the fundamentals are going to change anytime soon. And I don’t think they will.” Holding land is only justified by a view that the math improves. If construction costs keep climbing, the entitlement isn’t the bottleneck — the cost stack is.
His current position is blunt: “You can’t get the build-for-rent market doesn’t really pencil out right now.” Costs continue to rise and builders are scaling back.
That same observation is why he expects existing apartments to make another run. New supply doesn’t pencil, builders are pulling back, the country has a housing problem that isn’t resolving, and the absorption issues that hurt apartment operators over the past few years have largely worked through. Existing units become the only product that can meet demand at a price renters will pay.
The trade-off is that this is now widely understood. Crow’s own caution applies: “When it’s well-known that it’s strong and it’s obvious, it’s just hard to find good deals.” He says the same about industrial — well-located industrial is excellent and nearly impossible to buy right. Which puts the emphasis back on off-market sourcing and structure rather than on the thesis itself.
Staffing, Third-Party Management and What AI Actually Replaced
At 5,000 units, Soreste was fully vertically integrated — its own construction teams and its own property management. Crow is candid that this wasn’t strategy so much as necessity: “We really didn’t find that other people were doing all that great of jobs. We kind of got into it because we had to.”
When he shrank the portfolio and moved to third-party management, the results were uneven, and the variable wasn’t the firm. “Some are good, some are not. And it’s, again, people more than the firm. If they had the right person in the property, they were probably fine. And if they didn’t, then we struggled.”
That is the practical case for the unit count band. A property small enough for one strong on-site manager to know every tenant is a property whose performance you can diagnose. When you sign a third-party management agreement, you are hiring an individual, not a brand — so underwrite the individual and put a replacement mechanism in the contract.
On the analytical side, the headcount change has been dramatic. Crow previously ran a team of six analysts plus offshore staff. Today, one person using AI does the modeling and analysis, and he reviews more deals than before.
He resists framing that as a return story, and his framing is better: “If it helps you avoid mistakes, that’s a pretty good return on investment. Sometimes you’re worrying about maximizing your return, and other times you’re saying if you can avoid the losses, that’s just as good as making money on something else.” Use it to surface risks you’d otherwise miss — then price them into your terms.
Frequently asked questions
How many units does an apartment property need before an onsite property manager pencils out?
Roughly 75 units. Michael Crow’s rule is that you don’t get real operating efficiencies until you reach about 75 units, at which point the property can support a full-time manager on site. Below that, the revenue base doesn’t cover the position.
That threshold shapes acquisition strategy more than most investors account for. A 45-unit building either carries management cost it can’t afford or gets run part-time by an owner or a shared regional manager, and both outcomes show up in collections and turn times.
What actually goes wrong when you jump from 250-unit deals to 500-unit deals?
Physical sprawl and loss of tenant-level knowledge. Crow’s 500-unit Atlanta property sat on roughly 38 acres, and he says the difference between managing 250 units and managing 40 acres of apartments showed up from the start. No single manager can know every resident on a site that size, so vacancy, delinquency and deferred maintenance are harder to see and slower to fix.
In his case that gap, combined with a post-COVID operating environment, meant 18 months of injecting capital before he chose to take the loss and exit rather than fund a turnaround.
Is it better to buy one large apartment property or two smaller ones in the same market?
Two smaller ones, in Crow’s view. He’d rather own two separate 250-unit properties, each with a manager who knows every tenant, than one 500-unit site. The staffing model is more accountable, problems surface faster, and you retain the option to sell one asset without unwinding the whole position.
The counterargument — one loan, one insurance policy, one set of vendors — is real but smaller than the operating penalty he experienced.
How do you use terms instead of price to get an off-market apartment deal done?
Tell the seller they can set the price or the terms, but not both. That’s Crow’s opening line on off-market deals, and it moves the negotiation away from a single number toward structure — where a credible buyer has more room to work.
From there the mechanics are seller notes, holdbacks and adjustments. When diligence turns up a problem, put a dollar figure on it and propose a mechanism that covers it rather than reopening the purchase price. That approach gets deals signed that a straight price fight would kill.
Why doesn’t build-to-rent development pencil in many markets right now?
Construction costs. Crow held land for build-for-rent in Huntsville, Alabama and coastal Georgia and could not make the numbers work after several years of trying, so he sold at a loss on the view that costs would keep rising and the fundamentals would not change soon.
The knock-on effect is the investment case for existing product. With new construction not penciling and builders scaling back, existing apartments become the supply that meets demand — which is part of why he expects them to make another run.
The bottom line
Set your unit count band before you look at a single deal — Crow’s is 75 to 150 with 200 at the top — and treat anything above it as a different asset class requiring a different staffing plan, not a bigger version of what already works for you.
