Why Do Property Valuations Disagree? Zoopla vs Land Registry Data

By Sean R. · Updated 18 August 2026

Put the same address into Zoopla, Rightmove and a Land Registry-based valuation tool and you can easily get three different numbers, sometimes tens of thousands of pounds apart. That's not a bug in any one of them. They're built to answer slightly different questions, from different data, using different methods.

They use different underlying data

Zoopla and Rightmove's automated valuation models (AVMs) blend HM Land Registry sold prices with their own listing data, asking prices, how long properties sit on the market, and price changes during a listing. That listing data is proprietary and updates constantly, which is part of why their estimates can shift month to month even when nothing has sold nearby.

A tool built purely on Land Registry data (like this one) only uses actual completed sales, the price a property legally changed hands for, recorded by HM Land Registry. It's slower to reflect very recent market shifts, but every number it's built from is a real, verified transaction, not an asking price or an estimate someone else made.

They weight recency and location differently

How far back a comparable sale can be and still count, how tightly it needs to match your property's type and size, and how much weight a sale gets based on distance, every valuation model makes different calls on all of this. Widen the radius or the time window and the number moves. There's no single "correct" set of rules; it's a genuine methodology choice, and reasonable tools can land on different answers from the same underlying market.

Confidence bands matter more than the headline number

A valuation with 90+ comparable sales within half a mile from the last six months deserves far more trust than one built from five sales three years ago at the edge of a 15-mile radius, even if both spit out a specific-looking number. Any tool that gives you a confidence score or a range, rather than a single suspiciously precise figure, is telling you something important: how much to trust the estimate, not just what it is.

Which one should you trust?

Honestly, none of them, for anything that matters financially. Automated valuations of any kind are a starting point, not a substitute for a RICS survey or a formal mortgage valuation when you're buying, selling, or remortgaging. See our guide on RICS surveys vs online valuations for when you need the real thing.

What automated tools are good for is a fast, free sense-check: is this roughly in the right ballpark, and how has this area moved over time? For that, it's worth checking more than one and looking at the confidence/comp count behind each, not just the headline figure.

A worked example

Take a 3-bedroom semi in a Watford postcode sector. Zoopla's estimate: £415,000, built from 40+ recent listings including several still-on-market asking prices. Rightmove: £398,000, weighted more heavily toward recent price reductions in the area. A Land Registry-based tool (like this one): £389,000, built from 12 actual completed sales in the same sector over the past 18 months, HPI-adjusted to today. None of these is "wrong", they're answering the same question with different inputs and different recency assumptions. The £26,000 gap between the highest and lowest isn't noise; it's the visible effect of asking-price optimism (Zoopla/Rightmove) versus completed-transaction reality (Land Registry).

Frequently asked questions

Which number should I use when negotiating a price?

Lean toward the Land Registry-based figure, since it reflects what buyers have paid, not what sellers are currently asking. Use the AVM range as a sense-check, and always look at the individual comparable sales behind it, not just the headline number.

Why did my valuation change since last month?

Two possible reasons: a new comparable sale completed nearby and entered the dataset, or the HPI adjustment shifted slightly as more recent regional data became available. Small month-to-month movement is normal and doesn't mean your property's value genuinely changed overnight.

Do these tools ever agree?

Yes, in areas with a high volume of recent, tightly comparable sales, the gap between AVMs and Land Registry-based tools narrows considerably, since there's less room for differing assumptions to diverge. Sparse markets (rural areas, unusual property types) show the widest gaps.

Should I be worried if my valuation seems low compared to what I've seen elsewhere?

Not necessarily. A Land Registry-based figure is deliberately conservative because it only counts completed transactions, not asking prices or reduced listings that never sold. If you're selling, a slightly lower data-driven baseline is a more honest starting point for pricing than an inflated asking-price-derived estimate that risks sitting unsold for months.

In summary

• Different tools use different underlying data, asking prices and listing behaviour (Zoopla, Rightmove) versus completed sales (Land Registry-based tools).
• A gap of tens of thousands of pounds between tools is normal, not a sign one of them is broken.
• The confidence score or comparable-sale count behind a figure matters more than the headline number itself.
• Use automated tools as a starting point for negotiation or curiosity, never as a substitute for a RICS survey or mortgage valuation when money is changing hands.

If you're trying to decide which figure to trust for your own address, the practical move is to check more than one tool and look specifically at how each arrived at its number, how many comparables, how recent, how close by. A tool that shows its working is worth more than one that only shows a headline figure, regardless of which brand it carries.

None of this means you should distrust automated valuations altogether, it means reading them the way you'd read any statistical estimate: alongside its method, its inputs, and its confidence, not in isolation.

For the full method this site uses to reach its own number, see our guide on how UK property valuations work.