Two hosts in the same beach town, similar three-bedroom listings, both with cleaners they'd describe as "pretty good." The first pays hourly and has slowly become an amateur auditor — checking arrival times against door codes, wondering why Tuesday's clean took ninety minutes longer than Saturday's. The second pays a flat rate per turnover and has a different worry: the cleans are fast, sometimes suspiciously fast, and she's started finding things a rushed person misses. A stray hair on the bathroom tile. A coffee mug put away dirty.
Neither host is imagining things. They're each living inside a different incentive structure, and each structure is producing exactly the behavior it rewards. The question of how to pay Airbnb cleaners — flat rate or hourly — turns out to be one of the most studied questions in labor economics. The research has a clear answer. It also has a clear warning, and most hosts only ever hear the answer.
The question underneath the question
When you choose between hourly and per-turnover pay, you're really choosing what you're buying. Hourly pay buys time: the cleaner sells you their afternoon, and what happens during it is loosely coupled to what they earn. Flat-rate pay buys an outcome: a guest-ready property, however long that takes.
That distinction sounds academic until you notice how it changes the economics of every minute. Under hourly pay, a slow clean earns more than a fast one — not because anyone is scheming, but because the meter runs either way. Under flat-rate pay, every wasted minute comes out of the cleaner's effective wage, so speed becomes their problem instead of yours. Economists call this an alignment of incentives, and there's a famous natural experiment showing just how much it matters.
What happened when a glass company changed how it paid
In the mid-1990s, Safelite Glass Corporation — the windshield-replacement company — switched its installers from hourly wages to piece rates: a fixed amount per windshield installed. Because the company changed pay structures while everything else stayed constant, the economist Edward Lazear was able to study the shift with unusual precision. His 2000 paper, "Performance Pay and Productivity," became a classic.
The result: output per worker rose by roughly 44 percent. Even more interesting was where that gain came from. Only about half of it was existing workers speeding up. The other half came from sorting — under piece rates, the fastest, most capable installers earned more and stayed, while workers who preferred to coast under hourly pay drifted away. The pay structure didn't just change behavior. It changed who showed up.
Both halves of that finding matter for a host. Flat-rate turnover pay doesn't only make your current cleaner more efficient; over time, it selects for the kind of cleaner who thrives on efficiency. The pros who can turn a three-bedroom in three focused hours seek out flat-rate work, because their skill converts directly into earnings. Hourly pay, by contrast, quietly repels them — it caps what their speed is worth.
The catch: you get exactly what you measure
Here's the part of the research that rarely makes it into hosting forums. Around the same time as Lazear's study, the economists Bengt Holmström and Paul Milgrom formalized what's known as the multitasking problem: when a job has several dimensions and you pay for only the measurable one, effort migrates toward what's measured and away from what isn't.
A turnover has an easily measured dimension — done or not done, by when — and a hard-to-measure one: how well. Pay purely per turnover with no quality check, and the theory predicts precisely what our second host observed. Speed rises, because speed is what pays. Care erodes, because care is invisible. The cleaner isn't cutting corners out of laziness; they're responding rationally to a contract that prices thoroughness at zero.
Safelite knew this. The overlooked detail of that famous experiment is its quality safeguard: installers whose windshields failed had to redo the job on their own time, unpaid. Defects became expensive to the worker, not just the company — and quality held steady even as speed soared. The piece rate and the quality check weren't two separate policies. They were one system, and neither works alone.
What this means for turnover pay
Put the two findings together and the answer for hosts is straightforward: pay a flat rate per turnover, and make quality observable.
Flat-rate is right for reasons beyond speed. It matches how professional cleaners already think about their work — as a route of jobs, not a stack of hours. It makes your costs predictable per booking, which matters when your income arrives per booking. It removes the low-grade surveillance dynamic of hourly pay, where every long clean invites suspicion and every question lands as an accusation. And through Lazear's sorting effect, it gradually staffs your properties with the people flat rates attract: fast, experienced, self-managing.
But flat-rate without verification is half a system. If completion is the only thing anyone ever checks, completion is what you'll get — on time, every time, and gradually hollower. The fix isn't hovering or re-inspecting every clean yourself. It's making the quality dimension visible the way Safelite did: a defined standard, evidence that it was met, and the shared understanding that misses get corrected by the person who missed them.
In practice, that looks like a room-by-room checklist the cleaner confirms against, photos of the states that matter most — beds staged, bathroom surfaces, the supply closet — and the occasional spot-check. The photos aren't about distrust. They're the mechanism that lets you pay for outcomes honestly, because now the outcome includes quality, not just completion.
Setting the rate itself
The research also implies a few principles for the number you pick, even though the right figure varies wildly by market.
Price the job, not the clock. Scope the rate to the actual work — bedrooms, bathrooms, laundry loads, whether restocking is included — so a bigger job pays more without renegotiation. A rate that ignores scope invites exactly the corner-cutting you're trying to prevent.
Pay above the local floor. Sorting cuts both ways: if your rate is the lowest in town, the fastest cleaners take other routes and you inherit whoever's left. The premium you pay for a reliable professional is almost always smaller than the cost of one bad review — or one scrambled same-day replacement.
And pay fast. A flat rate that arrives weeks later loses much of its motivational force. The tighter the loop between finished job and money received, the more the incentive actually functions as one.
One more note: the cleaning fee your guest pays is a pricing decision, not a payroll one. Plenty of successful hosts pay their cleaner more than the fee on small stays and less on long ones. Untangling the two frees you to set each number for its own job.
A pay structure is a communication structure
What the Safelite story ultimately teaches is that compensation is a message. Hourly pay says: I'm buying your time, and I'll be watching how you spend it. Flat-rate pay with no quality loop says: finish fast, nothing else registers. Flat-rate pay with a visible standard says the thing you actually mean: I trust you to work at your own pace, the standard is defined, and meeting it is what gets rewarded. Cleaners hear these messages clearly, even when hosts don't realize they're sending them.
The hard part was never choosing flat-rate — most experienced hosts land there eventually. The hard part is building the verification half of the system without turning yourself into a full-time inspector. That's the gap Stayput was built for: it texts your cleaner the job for each turnover, walks them through your checklist, and asks for photo confirmation of the rooms and supplies that matter — so quality becomes observable in the normal flow of work, not through awkward follow-up calls. The flat rate rewards speed; the photo loop protects care. If you're running turnovers on trust and hoping, you can see how it works at stayput.lumenlabs.works.