Valuing a property is a professional discipline with defined methods, most of which do not apply to ordinary houses.
Comparable sales
Recent transactions of similar properties adjusted for differences.
Which is the dominant residential method.
Income approach
Value derived from rental income.
Which applies to investment property.
Cost approach
Land value plus rebuilding cost less depreciation.
Which is used for unusual properties and insurance.
Automated valuation
Statistical models producing estimates from sale data.
Which are approximations with wide error ranges.
What a valuer actually does
Inspects the property, identifies recent comparable sales, adjusts for differences in size, condition and location, and forms an opinion.
Which is a professional judgement supported by evidence rather than a calculation.
Two competent valuers can reasonably reach different figures for the same property, which is why lenders commission their own.
Why estate agent valuations differ
Marketing appraisals aimed at winning an instruction.
Which are not the same exercise.
Down valuations
Lender valuations below the agreed price.
Which can collapse a purchase.
Condition and improvements
Not all spending adding equivalent value.
Which surprises people who have renovated.
Automated estimates
Useful as a starting point and no more.
Why buyers and sellers disagree so persistently
A seller values their own property including improvements they made and the effort they put in; a valuer values it against what similar properties actually sold for.
Which are two different exercises producing two different numbers.
Improvements typically return less than they cost, sometimes considerably less, and that is one of the most consistent findings in residential property.
Location premiums
Schools, transport and amenities capitalised into prices.
Which is measurable and substantial.
Surveys against valuations
Condition reports serving a different purpose.
Which buyers frequently conflate.
Market conditions
Valuations reflecting a moment rather than a fixed worth.
Challenging a valuation
Providing comparable evidence to the lender.
What buyers should take from this
Asking prices are opinions, sold prices are evidence, and the two frequently differ substantially in either direction.
Which makes sold price data the thing worth researching before making an offer.
Land registry and equivalent records publish actual transaction prices in most countries, usually free.
Surveys worth paying for
Condition surveys separate from lender valuations.
Which identify problems a valuation does not look for.
Leasehold and tenure
Lease length materially affecting value.
Which catches buyers unfamiliar with the structure.
New build premiums
Prices frequently not sustained on resale.
Which is a well documented pattern.
A general note
Property law and practice differ substantially by country.
Why it is worth knowing how these things work
Most of the systems that shape ordinary life are invisible by design. Nobody explains why a parcel took an odd route, why a film left a streaming service, why a wait in an emergency department extended, or why the price of a flight changed between two searches.
The absence of explanation is rarely deliberate concealment. It is that the people running these systems are solving their own problems, and the reasoning behind their decisions is obvious to them and completely opaque from outside.
Understanding the mechanism does not always change what you can do about it. It does remove a category of low-grade frustration that comes from assuming something is arbitrary, unfair or aimed at you personally when it is usually none of those things.
A note on sources
Where regulation is involved, national regulators publish the actual rules and they are generally clearer than press coverage of them. Where an industry is involved, trade publications aimed at people working in it are considerably more informative than consumer coverage.
Practices described here vary substantially between countries, and anything with legal, financial or medical consequences warrants checking against the rules that apply where you are.
The pattern that recurs
Across almost all of these systems, the same three things turn out to be true. The behaviour that looks irrational from outside is optimising for something the observer cannot see. The cost that seems unexplained is usually concentrated in one specific stage of the process. And the information that would resolve the confusion is generally published somewhere and read by nobody.
That last point is the most useful one. Regulators, operators and industry bodies publish an enormous amount of explanatory material that answers exactly the questions people complain about not being able to get answers to. It is dry, it is not promoted, and it is free.
What to do with any of this
Very little of it changes a decision on its own. What it changes is the ability to ask a better question: of a company, of a service, of a regulator, or of yourself before assuming that something went wrong.
A final caveat
Everything above describes general practice. Individual companies, jurisdictions and circumstances differ, sometimes substantially, and the rules change more often than summaries of them get updated.
Where a decision matters, the official source for your own country is worth the ten minutes it takes to check.