Why Zillow's Zestimate Is Often Wrong for Utah Homes

If you’ve pulled up your address on Zillow and thought the number looked off, you’re probably not wrong. Utah is a non-disclosure state, and that one fact does more to explain Zestimate’s local misses than anything about the algorithm itself.

What “non-disclosure state” actually means

In most states, when a home sells, the final sale price becomes part of the public record. County assessors get it, title companies report it, and it flows into every pricing model that touches real estate. Utah doesn’t work that way. The deed transfer gets recorded, but the actual sale price isn’t required to be part of that public document.

A national valuation model built on public records simply doesn’t get the same raw material here that it gets in Ohio or Florida. It has to lean on whatever it can piece together instead: assessor estimates, MLS data where it’s shared, tax records that lag reality by months. None of that is a substitute for knowing what the house three doors down actually closed for last month.

Why this hits Washington County harder than most places

Zillow has been upfront for years that its own published error rates run noticeably higher for homes that aren’t currently listed, exactly the situation a non-disclosure state creates by default, since off-market comps are the ones a model has the least real data on.

St. George compounds it. A meaningful share of the housing stock here is custom or semi-custom: casitas, larger desert lots, view premiums on anything backing up to red rock. Those properties don’t fit neatly into an automated comp model built for rows of near-identical suburban tract homes. Add in how fast some pockets of Washington County are turning over right now, and a model working from stale or incomplete data can drift from reality quickly.

What actually gets you a reliable number

None of this means every online estimate is useless. It means the data source matters more here than it does in a disclosure state. A number that’s built from recent local closings, active and pending listings, and the specifics of your actual property is going to hold up better than one built primarily from public records that never had your county’s real numbers to begin with.

That’s the gap our estimate is built to close. It’s calibrated specifically to Washington County, not the country. It won’t replace a full comparative market analysis from a local agent, but it’ll get you a lot closer than a nationwide model that’s never had access to what actually closed on your street.