How we value a property

Every figure on this site is an estimate*. That word is not a disclaimer bolted on at the end — it is the accurate description of what a number produced without opening the floorboards can be. The useful question is therefore not how confident we are. It is what the figure was built from, which parts of it are known and which are assumed, and where it could be wrong. This page answers that, in the order the deal page presents it.

Where a property comes from before anybody values it

Every week our system prices residential property across 207 postcode districts in London and the Home Counties and ranks what it finds. That produces candidates, not deals. A candidate becomes something we send only after a person has looked at it, put a value on it, costed what can be costed and written down why it looks cheap — including what is wrong with it. If nothing is worth sending, we send nothing.

We start from what has actually sold

An asking price is an opinion, and often the opinion of whoever most wanted the instruction. So the evidence we lead with is completed sales: HM Land Registry Price Paid Data, which records what somebody actually paid when the sale was registered.

Each deal page carries the comparables table itself rather than a summary of it: the address, the property type, the month and year it sold, the floor area in square feet where one is on record, the price per square foot, and the price. You can read the evidence and disagree with the conclusion, which is the point of publishing it.

Two exclusions are worth knowing about. Repossessions and transfers between companies are left out, because neither is an ordinary sale between a willing buyer and a willing seller, and including them would drag the picture down. And Land Registry data is a record of the past: a sale appears once it has been registered, some time after it completed. It tells you where the market has been, not what somebody will pay next month. The sold prices we show contain HM Land Registry data © Crown copyright, used under the Open Government Licence.

Price per square foot, only where the area is on record

Price per square foot is the most useful single way of comparing two similar properties, because it takes size out of the comparison. It only works where the floor areas are actually known. Ours come from the EPC register, published under the same Open Government Licence, and the figure shown is simply the sold price divided by that recorded area.

Where no energy certificate matches a comparable, the square-foot and the price-per-square-foot cells are blank. They are not filled in with an approximation. A blank cell is information: it tells you that nobody has published a measurement for that property, and that any comparison you make with it is being made on price alone.

The cost lines we model

A value on its own is not a deal. What turns one into the other is the cost of getting from the price on the listing to a finished property somebody else buys. Where a worked appraisal has been published, the deal sheet sets those costs out in four sections, and every line is on the page rather than summarised into a total.

• Acquisition — the purchase price, stamp duty, a separate stamp duty saving line where one applies, and the legal cost of buying.
• The project — the lease extension where there is one, the build or refurbishment, professional fees, service charge, the cost of holding the property while the work is done, and a contingency.
• Finance — how long the project is assumed to take in months, how much of it is assumed to be borrowed, and the interest that follows, with the rate that was assumed shown beside it.
• The exit — the sale price we think it achieves, the selling agent's fees and the legal cost of selling.

Those produce the two totals — the total project cost and the cash actually deployed — and from them the returns: gross margin*, total return*, return on cash* and annualised return*. Gross margin is what is left of the modelled sale price after the modelled costs. It is gross because it is before tax and before anything the sheet has not managed to cost, and it is not a prediction of what you would make. Any line we estimated rather than obtained is tagged estimated on the page, so the difference between a quoted number and an assumed one is visible without asking.

Where no appraisal has been published for a property, the page says so plainly and the headline figures stand as what they are: gross, before purchase costs, works, finance and the cost of selling.

Why "not yet assessed" beats a guess

When a cost has not been established, the deal sheet does not leave the line blank and does not put a nil in it. It says "Not yet assessed". Three other things happen at the same time: the missing line is named in a warning pill at the very top of the page, before any figure at all; the sheet is flagged as incomplete; and a callout states that the margin and the return above it are stated before that cost.

This matters more than it sounds, because on these properties the two lines most likely to be unknown are usually the two largest — the cost of extending a lease and the cost of the works. A blank line in a ledger reads as nil to almost everybody, and a nil in either of those places would turn a marginal deal into a good-looking one on screen.

A guess would be worse still. A guessed number looks exactly like a known one: it sits in the same column, in the same typeface, and every figure below it inherits it silently. Nobody can audit it, because there is nothing to audit — no quote, no valuation, no source. Saying "we do not know this yet" is less satisfying and considerably more useful. It tells you precisely where to spend your own time and money: a builder through the property, or a valuer on the lease. That is a cheap thing to be told and an expensive thing to find out later.

How long it has really been on the market

The day count shown on a listing is not always the day count of the sale. It resets when a property is withdrawn and re-listed, or moved to another agent. The figure we publish is the true continuous marketing spell across any agent, and where the two disagree the page says so and explains why — a counter that has been reset is a signal in itself.

Alongside it the page shows when marketing began, how many times the property has been listed, how many agents have had it, and any reduction with both the amount and the percentage. Where there have been several changes of asking price, they are shown as a table with the dates, so a slow drift down looks different from a single decisive cut.

What the figures do when the assumptions move

Two sections exist because a single number invites more confidence than it deserves. Where they have been published for a property, the deal page shows what the gross margin* and the annualised return* do at purchase prices below the asking price, and what they do if the eventual sale comes in below our estimate. Both are tables of estimates*, not floors: the market can move further than any row on them, and every other assumption in the sheet moves with it.

What this is not

It is not a survey. Nobody from here has been on the roof, lifted a floorboard, or tested the electrics. It is not a valuation by a chartered surveyor and it is not a mortgage valuation, neither of which we can or do produce. It is not a promise about what a property will sell for, and it is not advice to buy anything. Property values can go down as well as up.

What it is: our own working, published in full, on a property we thought was worth somebody's attention. We show the comparables so you can test the value, the cost lines so you can test the deal, and the gaps so you know what is still unpriced. Your surveyor, your builder and your solicitor are the people who will confirm or demolish it, and every one of these pages is written on the assumption that you will ask them.

*Estimates based on comparable sales and local data; not a valuation and not guaranteed. Always do your own due diligence.