What is a good unit turn time in multifamily?

Ask five operators what a good turn time is and you will get 3 days, 5 days, 7 days, "depends on the market," and one person who does not measure it.

They are all answering a different question, because almost nobody agrees on when the clock starts.

The clock starts in four different places, depending on who you ask

Before comparing your number to anyone else's, pin down what you are counting. In practice we see four definitions in use:

Move-out to move-in ready. The honest one. It includes the days nobody scheduled anything.

Scope approval to punch complete. The maintenance-flattering one. It starts after the delay has already happened.

Move-out to leased. The asset-manager one. It mixes maintenance performance with leasing performance, so a slow turn and a soft market look identical.

Whatever the PMS reports. Which is usually one of the above, chosen by whoever configured it, and rarely documented.

A property reporting 4.5 days on the second definition can be sitting at 11 days on the first. Both numbers are real. Only one of them explains the vacancy loss on the P&L.

Ayla ran 340 units and lost $601,807 in a year

Benchmarks are abstract, so here is a real property.

Ayla Apartments is a 340-unit community in Tampa, Class C, built in 1966, managed by ResProp. Techs traveled between work orders by golf cart. When they needed a new work order they drove back to the leasing office to get one. Paper from start to finish.

The make-ready vacancy loss for a single year was $601,807.

That is the number that matters, not the turn-time average. Nobody at Ayla was failing to work hard. They were losing at least an hour of labor a day to travel, and the units sat while that happened.

After GetDone, average make-ready time dropped 51% and vacancy loss dropped 55%.

The average hides the properties that are actually bleeding

A portfolio average is close to useless for management, because turn time is not normally distributed. You have a long tail.

Sort your properties by average turn descending and look at the top five. That is where the money is. A portfolio at 6.2 days average might be four properties at 4 days and one at 19, and the one at 19 is the entire variance.

The same applies inside a property. Ask for the distribution, not the mean:

The 90th percentile is the useful management number. The median tells you whether the process works. The 90th tells you what happens when it does not.

A vacant day costs about $50, and the arithmetic is worth doing yourself

At $1,500 a month, a vacant day is $50 in rent you will never collect. That is the whole mechanism, and it is why turn time is the one maintenance metric that lands directly in NOI rather than in a labor line.

Two days off an average turn is $100 back, per turn. Multiply by the turns you actually did last year, not by a per-100-units figure somebody handed you. Your turnover rate is the variable that makes those extrapolations wrong.

We worked through the full cost model, including the parts that are not rent, in what a vacant day actually costs your portfolio.

Four things move turn time, and only one of them is effort

When a turn runs long, the cause is almost never that somebody was lazy.

First, the handoff. Paint finishes and nobody tells carpet. The unit sits for two days between trades because the schedule lives in one person's head.

Second, materials. The tech arrives, the part is wrong, and the unit waits for a delivery nobody tracked.

Third, scope discovery. Move-out inspection was cursory, so real damage surfaces on day four and resets the schedule.

Fourth, travel. On a spread-out campus this is not trivial. Ayla lost an hour a day per tech to it.

Notice that three of the four are information problems, not labor problems. That is why a whiteboard stops working as soon as the person who maintains it takes a week off.

What a defensible turn-time program looks like

Pick one definition and write it down. Move-out to move-in ready is the one we would defend, because it is the one the P&L agrees with.

Timestamp every stage rather than the endpoints. Move-out, scope, paint, carpet, punch, clean, inspect, ready. Without stage timestamps you know a turn took 11 days and nothing about where it went.

Report the 90th percentile alongside the median, per property.

Review the longest turn each month with the supervisor. Not to assign blame. The outlier usually reveals a broken handoff that is also quietly costing you two days on every other unit.

Tie it to dollars in every report. "Turn time improved 0.8 days" gets nodded at. "We recovered about $40 a unit per turn, on 190 turns" gets funded.

The benchmark matters less than knowing your own number

If you want a single number: under 7 days from move-out to move-in ready, measured with stage timestamps, on a stabilized conventional asset. Under 5 is genuinely good. Over 10 on a Class A property means something structural is broken.

But the benchmark is the least useful part of this. Ayla's problem was never that 8 days was worse than 5. It was that $601,807 was leaving the building every year and nobody could point at where.

What definition does your PMS actually use?