Is that multiple monthly or annual?

Suppose a site sells for $24,000 and earns $1,000 a month, which is $12,000 a year. On annual profit that is a 2× sale. On monthly profit it is a 24× sale. Nothing about the transaction changed between those two sentences. One number is the price divided by a year of earnings; the other is the price divided by a month of them.

Both conventions are in live use. Neither is wrong. The failure is comparative: a buyer who reads "24×" against a mental benchmark of annual multiples concludes the seller has lost their mind, and a seller who reads our median above against a monthly benchmark concludes their site is nearly worthless. Both walk away from a deal that was priced normally.

Do not trust the label; do the division. Take the price and the profit figure the listing states, and see which period reproduces the multiple the listing quotes:

• Divide the price by the annual profit. If that lands on the quoted multiple, it is an annual multiple.
• Divide the price by the monthly profit — or multiply the annual result by twelve. If that lands on the quoted multiple, it is a monthly one.
• If neither lands, the multiple is not computed from the two figures on the page. That is worth knowing before anything else: something has been annualized from a good month, or the multiple is quoting revenue while the page shows profit, or the price is an ask rather than a sale.

The rough magnitudes give you a prior while you check. Our medians on annual profit are low single-digit numbers — you can read the current ones, with the sample size behind each, in our live figures by asset type. So a quoted multiple in the twenties or thirties is very likely a monthly one, and a low single-digit figure very likely an annual one. Use that to decide what to verify, never as the answer itself — our own data holds real sales at both extremes, and a multiple can be enormous simply because the profit figure underneath it is tiny.

Every multiple we publish is on trailing-twelve-month profit. We chose one convention and hold it everywhere — the medians on this site, the calculator's ranges, the comps export and the paid report — because a dataset that mixed periods would be unusable and quietly so.

Listings do not oblige. Where a page states profit for a month rather than a year, our ingest rule has, since the first version of the parser, multiplied it by twelve and recorded that the figure was annualized from a monthly one, so a row's basis stays inspectable rather than becoming an assumption. That rule exists because monthly figures show up often enough on real listings to need one.

Annualizing a single month is an approximation, and it is worth naming as one: a month is not a twelfth of a seasonal year, and a listing showing its best month is a different claim from a listing showing a trailing year. Where a page gives a full year, we take the year.

• Benchmarking across conventions. Comparing a multiple you read in one place against a benchmark computed in the other is a twelvefold error dressed as market research. Before any two multiples are compared, both must be on the same period and the same basis.
• Anchoring a negotiation on the wrong unit. "Comparable sites go for 3×" is a very different opening position depending on the period, and the party who has not checked is the one who concedes.
• Building a model on a mixed set. If you are assembling your own comps in a spreadsheet, one row entered on the wrong period does not simply add noise — it lands twelve times away from the others and drags a mean with it. Medians are more robust than means for exactly this reason, which is why we publish medians and show the sample size beside them.