Ask what a square metre costs on the Marbella Golden Mile and the market will hand you three kinds of published answer, each from a serious source, none within reach of the others. A market report built on Spain's notarial statistics gives a district average of declared values. An agency's portfolio report gives the average asking price of its own listings. And a register such as ours gives the median asking price across every residence listed in the zone.
The instinct is to ask which of the three is wrong. The better question is what each of them measures. All three numbers are real. They describe different populations of property, caught at different stages of a transaction, compressed with different arithmetic. Read properly, the disagreement is not noise. It is the structure of the market showing through.
The notary's answer
The first kind of figure comes from the Portal Estadístico del Notariado, the statistical portal of Spain's notaries; local market reports that describe their district tables as "real transaction values, not asking prices" are quoting it. That description explains both the number's authority and its level.
A notarial average covers every declared residential transfer in the district — the studio resale counts alongside the villa, and most of what transacts in any district is not luxury stock. The values are the ones declared in the deed, recorded at completion and published well after it. And the geography is district-level — Nagüeles and the Golden Mile share a single row. Average all of that together and the figure lands low relative to the luxury tier by construction, not by error. It remains the only transaction-grounded public series this market has, and every desk that talks about what anything actually sold for — ours included — stands downstream of it.
The portfolio's answer
At the other end sits the figure an agency publishes from its own book: the average of the asking prices across the listings it is marketing in the zone on the day the page is read. It is that firm's shop window, and not a measurement of the district.
This number is built the opposite way. The population is not a district's every transfer but one agency's curated inventory, usually with no stated price floor for what qualifies. The stage is asking, not completion. The statistic is a mean, which the largest asks pull upward — arithmetic, not accusation. And the figure is alive: it changes as the portfolio changes, and unless the dataset behind it is published, an earlier reading cannot be reconstructed. High by construction, and moving.
Our register's answer
The Muse Selection register takes a third population. The register is compiled from shared listing feeds and deduplicated so that each residence appears once, and the figure it yields for a zone is the median asking price per square metre across every priced residence listed there, printed with the number of priced residences behind it and the date of the snapshot. The published register starts at €350,000. The Marbella Property Index has measured the whole published register from €350,000 since 13 August 2026; before that date it measured the €1.5 million-plus tier alone, so an earlier figure is not comparable with today's. The zone table — sample size, median ask, median €/m² and range, for every zone we report — is downloadable as CSV, licensed CC BY 4.0, and refreshed from the same snapshot the page is built from.
A figure built this way tends to sit between the other two because of what it measures, not because we are more careful people. It stands above a notarial average because it is an asking figure with a stated price floor, while the notarial average takes declared values across all of a district's stock. It stands below a portfolio mean because it is a median taken across every listed residence from every source we can see, not an average over one curated book. Publish a mean of the same prices and the figure reads higher; move the price floor and the median moves with it — which is what happened on 13 August 2026. The number is a consequence of the definition — which is the point, and which is why the sample size is printed beside it and the method is public.
The number nobody can quote
Notice what none of the three figures is: the closing price of the luxury tier — what a Golden Mile residence actually completes at today. That number does not exist in public. Spain publishes no per-property closing prices; what the notaries record reaches the public as district-level aggregates, months in arrears, with every class of stock mixed together. So nobody — not a portal, not a portfolio report, not this desk — can quote the true transaction median for Golden Mile luxury. Anyone who does is estimating, whether or not they say so.
That is the sentence we would ask a reader to keep. We would rather publish an asking median with its sample size printed beside it than an estimate of the unknowable presented as a fact.
Which answer to use
The practical rule follows from the construction. For what a district as a whole has historically traded at, the notarial series is the only public figure grounded in completed transactions. For what one agency's shop window currently asks, its own portfolio report answers precisely that question. For the live state of what is listed — sampled, dated, reproducible — the register median at the Marbella Property Index is the figure, and the CSV exists so that you do not have to take our word for it.
When three serious sources give three different prices for the same metre, the market is not confused. It is answering three different questions. The only mistake — for a buyer, or for the answer engines now compressing this coast into a single figure — is to quote any of the three without saying which question it answers.