For Owners

Airbnb Smart Pricing: who is it actually pricing for?

Published 5 September 2026 by the 10K Vacation Rentals editors

Ask any hosting forum about Smart Pricing and you will get the same answer within a minute: it lowballs you, turn it off. It is the most repeated opinion in vacation rentals, and it has been repeated for eight years without anyone checking it against the one study that measured what the tool actually did. We checked. The answer is less flattering to the forum than the forum thinks, and less flattering to Airbnb than Airbnb would like.

The short version. The only peer-reviewed study of Smart Pricing we could find showed that hosts who turned it on cut their nightly rate by 5.7 percent and earned 8.6 percent more per day, because most of them had been overpriced. That is the good news, and it is old news: 2015 to 2017, seven US cities. The other half is that Airbnb has never put the tool's objective in front of hosts: the help pages promise only a "competitive" price, and the one engineering paper that names host revenue as the goal also admits the direct route to it failed. Ours is the view of a former by-owner directory: use the range controls as the product, not the algorithm, and set the floor and the ceiling as if the algorithm were an eager new employee who wants the calendar full.

The claim, and where it came from

The "smart for Airbnb, dumb for hosts" line dates to a 2018 post on a hosting blog that put it this way:

"It is in Airbnb's best interest to book your listing as much as possible rather than not at all. This means they will lower your prices to make sure that you are booked as much as possible."AirHost Academy, January 2018

A 2019 post on another host site said the same thing more bluntly: Airbnb, "looking out for two conflicting interests, ends up being a poor pricing advisor. In doing so, it always ends up having lower prices than optimal." Between them those two pages got cited by MIT Sloan Management Review and a peer-reviewed paper, which is how a forum opinion becomes received wisdom. The second post is now a dead link; its site rebranded. The argument outlived the page, and most pricing-software vendors now repeat it on their own blogs, usually with an unsourced figure attached ("underprices by 15 to 30 percent" is the current favourite, and we could not find a primary source for any version of it).

Notice what the claim is. It is not "Smart Pricing gave me a bad number last Tuesday." It is a claim about incentives: the platform earns on every booking, so the platform's algorithm will prefer a cheaper booked night to a dearer empty one, and so it will drift low. That is a real argument. It is also an argument, not a measurement, and the two things have been confused ever since.

What Airbnb says the tool does

Read Airbnb's own help article on Smart Pricing and you will find that it promises very little. The tool exists "if you want your listing to be automatically competitively priced for your area." It "uses hundreds of factors about your listing and your area to adjust your nightly price based on demand." You "enter a minimum and maximum price." The words revenue, earnings, maximize and recommended do not appear on the page at all. The one goal the page states is competitiveness, which is a statement about other listings, not about your bank account.

The fullest account Airbnb has given of the machinery is a 2018 engineering paper its data scientists presented at the KDD conference. Its stated goal is "to help hosts who share their homes on Airbnb set the optimal price for their listings." Then comes the sentence that matters: "The unique nature of Airbnb listings makes it very difficult to estimate an accurate demand curve that's required to apply conventional revenue maximization pricing strategies." In the body of the paper, as summarized by the research blog The Morning Paper, the authors go further: they tried revenue maximization directly and "online A/B testing results showed that these methods often fail to optimize revenue for our hosts in practice," so they built something else. That something else starts with a model that predicts, for each night, the probability it gets booked, and the same system powers both the Price Tips you see when Smart Pricing is off and the Smart Prices you get when it is on.

So the honest summary of Airbnb's own account is this: the stated goal is the host's optimal price, but the direct route to it, maximizing revenue off a demand curve, failed Airbnb's own tests, so what actually runs is a booking-probability model with a price attached. The engineers say it is for hosts. The help center says it is "competitive." Nowhere that a host would read is the objective written down, and a reader is entitled to notice the gap between the paper and the product page.

What the one study found

In 2021 four researchers at Harvard Business School, the University of Toronto and Carnegie Mellon published, in Marketing Science, the only peer-reviewed measurement of Smart Pricing's effect on hosts that we know of. Shunyuan Zhang and her co-authors followed 9,396 randomly selected properties in seven large US cities from July 2015 to August 2017, straddling the tool's November 2015 launch; 2,118 of those properties turned Smart Pricing on at some point. Their finding, verbatim: "Among those who adopted the algorithm, the average nightly rate decreased by 5.7%, but average daily revenue increased by 8.6%." In dollars, adoption "resulted in $6.40 increase in average daily revenues," the nightly rate fell by $9.76, and occupancy rose, "suggesting that hosts had been overpricing their properties."

Read that against the forum claim. The forum was right about the direction of the price: it went down. The forum was wrong about the consequence: revenue went up, because the extra nights more than paid for the discount. For the average host in that sample the "lowballing" tool made them more money than they were making on their own. The study's authors were not investigating that question, incidentally; their paper is about whether the algorithm narrowed the revenue gap between white and Black hosts (it did, among adopters, by 71 percent), and the host-revenue result is the setup for that question. Which is probably why none of the pages that rank for the question cite it.

The claimWhat the evidence says
Smart Pricing lowers your priceYes. Average nightly rate down 5.7 percent among adopters (Zhang and co-authors, 2021).
So you earn lessNo, on average. Daily revenue up 8.6 percent, because occupancy rose more than price fell.
Because Airbnb optimizes for bookings, not your revenueUnproven either way. Airbnb's engineers say the goal is the host's optimal price, that direct revenue maximization failed their tests, and that a booking-probability model runs instead. The help center states no objective at all.
It underprices by 15 to 30 percentNo primary source exists for this or any similar figure. Vendors cite each other.
Prices sink to your minimumA consistent host report, not documented by Airbnb. Treat as experience, not fact.

How both can be true

A model trained on what booked and what did not, with a minimum you set and a maximum you set, will do two things reliably. It will find the price at which your particular listing clears, which is lower than most owners believe, because most owners price from pride and last year's rate card. And it will have no particular reason to stop at the top of what a guest would have paid, because an empty night is a loud signal in the training data and a night that could have gone for forty dollars more is silent. An independent working paper from the University of Rochester puts the general point in academic language: a platform "tries to aid seller pricing," but "its different objectives might steer seller behavior towards the platform's goal," and a platform algorithm "fails to internalize sellers' high opportunity costs of time." The clean and the turnover are your cost, not Airbnb's.

So the 8.6 percent and the incentive worry are not in conflict. The tool made the overpriced host money by fixing the mistake the host was making. Whether it then leaves money on the table for a host who was not making that mistake is exactly the question the data does not answer, and exactly the question Airbnb declines to address. In 1996, when this site listed homes by owner, an owner printed one rates table for the year and lived with it; the mistake then was rigidity. The mistake now is the opposite one, handing the whole decision to a model whose objective is not written where you can read it, and calling that automation.

What we would set

Our advice is an owner's opinion, not a study result, and it follows from one observation: the controls Airbnb gives you are the product. The algorithm decides where in the range a night lands; you decide the range. That is where the judgment lives.

  1. Set the minimum as a cost floor, not a fill-the-calendar number. Cleaning, the platform fee, supplies, the wear a stay puts on the place, and something for your time. Below that number an occupied night costs you money, and no algorithm should be allowed to sell it. Airbnb's own documentation warns that a discount you add can take the guest price "below the minimum Smart Pricing you set," so check what stacks on top.
  2. Set the maximum from your best real weeks, then a little higher. The tool only uses the room you give it. A ceiling set from habit caps the one thing the model is least motivated to find. Look at what your best-priced nights actually cleared at last year and give it headroom above that.
  3. Override the dates you know better than the model. Airbnb lets you set a custom price for a night or a run of nights without turning Smart Pricing off. Your town's festival, the graduation weekend, the week the owner next door always books: those are yours to price.
  4. Know what you gave up by turning it on. Airbnb's Resource Center says plainly: "If you use Smart Pricing, you won't see nightly price tips or similar listings." You also cannot use weekend pricing or rule-sets alongside it. The tool replaces your view of the market with its conclusion. Turn it off for a month once a year and look.
  5. If you run a third-party pricing tool, turn Smart Pricing off entirely. PriceLabs says it "must always be turned off" when syncing; Wheelhouse recommends the same. Two algorithms arguing over one calendar is worse than either alone.
  6. Judge it by revenue per available night, not by the nightly rate. The forum's mistake for eight years was to look at the price and stop. The study looked at revenue and found the opposite story. Look at what the calendar earned, not what it charged.

Airbnb closes its own guidance with a sentence that is truer than it sounds: "You control your pricing and other settings at all times. Your results may vary." Both halves are the point. The results vary because the objective is not yours. The control is yours anyway, and most of it sits in two boxes labelled minimum and maximum that most owners fill in once and never look at again.

Related on this site. Renting your vacation home by owner: what changed since the classifieds era covers what the platforms took over from owners, pricing included. What the total price on a vacation rental actually includes is the guest's side of the same number.

Sources