How we calculate a neighborhood’s numbers

Where the data comes from

Two sources, and no others. We do not buy a proprietary index and we do not model anything we cannot show you the inputs for.

The U.S. Census Bureau — how many homes

Housing counts come from Census block groups. A neighborhood on our register is a named area made up of one or more whole block groups, and its housing figures are the sum across its members.

RentCast — what sold, and for how much

Sales activity comes from RentCast, per block group, over a trailing twelve months. That gives us a count of homes sold and a median price for each block group we then aggregate.

The arithmetic, figure by figure

Single-family homes (the mailable universe)

The sum of single-family detached units across the neighborhood’s block groups. This is the number that matters most, because it is what we actually mail — every one of them, every month.

The honest caveat: the Census single-family column is not populated for every block group. When some members report it and others do not, the sum is single-family only and therefore undercounts the neighborhood. When none of them report it, we fall back to total housing units and the brief says “homes” rather than “single-family homes” — because calling a housing-unit count a single-family count would be a fabricated statistic, not a rounding choice.

Homes sold, last 12 months

The sum of RentCast’s trailing-twelve-month sold counts across the neighborhood’s block groups. A whole number of real transactions, not an annualised projection from a shorter window.

Turnover rate

Homes sold in the last twelve months divided by the mailable home count, as a percentage. Nothing more.

This used to be wrong, and it is worth saying how. We previously averaged the per-block-group turnover rates RentCast reports. An unweighted mean of rates across groups of very different sizes is not the rate of the whole: a 50-home block group at 10% and a 3,000-home block group at 3% average to 6.5%, while the real combined rate is about 3.1%. Small, churny blocks dragged every neighborhood’s figure up, and on one neighborhood it produced a turnover rate that implied more sales than the sale count printed two inches above it. Deriving turnover from the two figures we already publish means homes, sales and turnover now reconcile with each other, because they are the same numbers.

Median list price

RentCast’s median list price for each of the neighborhood’s block groups, combined by taking their mean.

It is a list price, and the briefs now say so. Until recently they labelled this figure “median sale”, which it is not. Homes do not always sell for what they list for, and an agent knows that better than we do — but the label should not have been ours to blur.

The other caveat: combining block-group medians by their mean is not a true weighted median. In a neighborhood whose block groups are similar in size and price, the difference is small. Where they differ sharply, it will read high. Treat the figure as an indication of scale, not as an appraisal.

What one close is worth

The median list price multiplied by an agent-side commission rate of 3% — one side of a conventional total. Founding agents pay no fee at closing, so what a brief shows as going to the agent is the whole of it. Since the input is a list price, this figure inherits everything said about it above.

How a neighborhood ranks against its market

The comparison on each brief — largest, fastest-moving, most expensive — is a straight count of the other neighborhoods we have graded in that market, using the figures above. Ties share the better rank. Where a neighborhood is missing an underlying figure, it is left out of that ranking rather than imputed, so a denominator on one line may differ from the denominator on the next.

What these numbers are not

They are not a prediction. Nothing here forecasts who will sell or how many listings you will take. The signals we use to shape a card — a sale nearby, long tenure, high equity — are real events with a reason you can say out loud, and they are not a score. Anyone selling you a probability that a specific homeowner will list is selling you a number they made up.

They are not an earnings claim. A neighborhood’s commission pool is what the whole neighborhood transacts in a year across every agent working it. It is a measure of the opportunity in a place, not of your share of it.

They are estimates from listing data. Every brief says so. Where a figure is missing we render nothing rather than a placeholder — a blank is honest and a zero is not.

They describe a neighborhood, not a house. Nothing we publish about a neighborhood is an opinion about any individual property or owner.

How often it updates

Census housing counts change on the Census Bureau’s schedule. RentCast figures roll forward as their trailing window moves. The comparison between neighborhoods is recomputed hourly from whatever the underlying figures currently say.

Citing these numbers

You are welcome to. If you quote a figure from a neighborhood brief, please link to that brief and name the date you read it — the underlying windows move, and a figure without a date is unfalsifiable for the next person. If a number here looks wrong to you, tell us: the turnover error above was found by someone checking our arithmetic against our own page, which is exactly the intended use of this page.

Browse every neighborhood we have graded →