Airbnb Market Scores: How to Read One Without Getting Fooled

A market score is several different metrics collapsed into a single number, weighted according to somebody's judgment about what matters. The metrics are data. The weighting is an opinion, and it's usually not published. That gap is where scores mislead people.

What you're actually looking at

Any composite score has two layers. Underneath are measurable things: occupancy, average daily rate, revenue, how much supply there is and how fast it's growing. On top is a decision about how much each of those counts.

Change the weighting and the ranking changes. A score that leans on revenue growth will favor markets that ran hot last year. One that leans on supply constraint will favor markets that are hard to build in. Both are defensible, and they will hand you different answers about the same two cities.

This isn't a flaw in any particular product. It's what a composite is. But it changes the question worth asking. Not "is this score accurate," which has no answer. Accurate at what.

Three questions worth asking any score

What went into it, and in what proportion?

If the weighting isn't published, you're being asked to accept someone's investment thesis without seeing it. That may well be fine. It might be a good thesis. But you should know that's the transaction.

Where a methodology is available, read it once properly. You're looking for whether the inputs match what you care about. A score built for a buy-and-hold investor is answering a different question than one built for someone doing arbitrage on a twelve-month lease.

What time window does it cover?

A score computed over trailing twelve months and a score computed over trailing twenty-four will disagree in any market that had an unusual year, which since 2020 is most of them. Neither is wrong. They're describing different lengths of history.

The failure mode is comparing two markets using scores calculated over different windows, or reading a score published eight months ago as though it describes now.

Is it scoring the market, or scoring properties like yours?

This is the one that bites hardest. A market-level score averages over every property type in the area. If you're buying a studio in a market whose numbers are carried by four-bedroom houses, the headline score is describing someone else's investment.

That gap between the whole-market view and the slice of it you'd actually own is large enough to reverse a decision. It's worth being clear about which of those two views you're reading before you let a number influence anything.

What no score can see

A score ranks markets against each other. It cannot tell you whether you should buy, because the variables that settle that question aren't market variables at all. They're yours.

It doesn't know what you'd pay, or on what terms, and a market that works at one entry price fails at another with identical revenue. It doesn't know whether you'll run the property yourself or hand it to a manager taking a fifth of the gross. It doesn't know whether you're good at this, and you should not skip past that one: the same property under two operators produces materially different numbers, and no amount of market data narrows that gap.

Every one of those sits downstream of the score and upstream of your actual return. A market score is a statement about a place. An investment is a statement about a place, a price and a person, and only one of those three is being measured.

Where a written read helps, and where it doesn't

AI-generated market analysis is now standard across this category. AirDNA launched an AI underwriting product in August 2026, and we build an AI consultant into Market Analysis ourselves. So it's worth saying plainly what that changes and what it doesn't.

What it changes: a paragraph can tell you why, and a number can't. It can say occupancy is holding but rates are softening, which is a different situation from both falling, even where the two produce a similar score. Our own read works this way, reasoning through a market the way an experienced investor would rather than restating the chart.

What it doesn't change: it's still working from the same underlying inputs, and it still can't see your purchase price. Treat a written analysis as something you can argue with. Ask what would have to be true for the read to be wrong. That's a more useful posture than taking it as a verdict with better manners.

How to use a score properly

Use it to shorten a list, never to end one. A score is good at telling you which fifteen markets aren't worth your next four hours. It is not good at choosing between the two finalists, because by that point the differences that matter are ones it never measured.

Then do the actual work on the shortlist: supply trend, seasonality shape, regulation, and your real numbers at your real entry price. That's the full market analysis process, and no score substitutes for it.

Takeaways

• A score's inputs are data; its weighting is an opinion, and the opinion is usually undisclosed.
• Ask what the score optimizes for before asking whether it's accurate. Different questions.
• Check the time window, and never compare scores computed over different ones.
• A market-level score averages over property types you may not be buying.
• No score knows your purchase price, your financing or your operating ability, which is where most of your return is decided.

Run your own numbers on the shortlist

A score can hand you five markets worth looking at. What it can't do is tell you what a specific property in one of them returns at the price you'd pay.

The Revenue Calculator is free and takes no account to use, so it's a reasonable next step once a score has narrowed the field. Estimates are estimates, but these ones are about your property rather than about an average.