The number at the top of a listing has a strange power over people. Buyers treat it as the fair price, sellers treat it as a promise, and arguments start when reality lands somewhere else. That number is an automated estimate, produced by a model that has never set foot in the house. It is a useful starting point and a poor stopping point. Understanding how it is built, and what it cannot see, changes how much weight you should give it. The estimate is a guess dressed up as a fact, and the gap between the two costs people money.
Start with how the figure is actually made. An automated valuation model takes public records and recent sales data and runs them through a formula. It looks at square footage, the number of beds and baths, the lot size, the year built, and what similar nearby homes recently sold for. From all of that it produces a single dollar figure that looks confident and precise. But everything feeding it comes from data that already exists on paper. The model is pattern matching against the past, not judging the home in front of it. That distinction is the whole story.
The companies behind these estimates are honest about their limits, if you read the fine print. They publish their own median error rates, and those numbers are worth knowing. For homes actively on the market, the typical error often sits in the low single digits, roughly two to three percent. For homes not currently listed, where there is less fresh data, the error runs higher, frequently around six to seven percent or more. A median error means half of homes miss by less and half miss by more. On a four hundred thousand dollar home, even a small percentage is tens of thousands of dollars in either direction.
The deeper issue is everything the model cannot see. It does not know that the kitchen was gutted and redone last year. It does not know the roof is failing or that the basement floods in heavy rain. It cannot tell a tired interior from a renovated one, or a quiet street from one next to a busy road. Two homes with identical records can be worth very different amounts in real life. The model treats them as twins because on paper they are. The features that move a buyer most are often the ones it has no way to measure.
This matters differently depending on which side of the deal you are on. A seller who anchors to a high estimate may price the home above what buyers will pay and watch it sit. A buyer who trusts a low estimate may walk away from a home that is actually fairly priced. Both are letting a formula make a judgment that belongs to a person. The estimate did not tour the home, read the neighborhood, or account for what is happening in the market this week. Treating it as the final word hands real decisions to a tool that was never built to make them.
So what is the number actually good for? It is a fine first glance, a way to get oriented before you do the real work. Use it to understand the rough range a home might fall in, then set it aside. The work that matters is a comparative market analysis from an agent who knows the area, or a full appraisal when real money is on the line. Those look at condition, recent local sales, and the things a model misses by design. The estimate starts the conversation. It should never end it.
It also helps to know that the estimate moves, sometimes for reasons that have nothing to do with the home. When new sales post in the area, the model recalculates and the figure shifts. A home can gain or lose thousands on paper in a week without a single nail being touched inside it. That movement reflects the data, not the house. Lenders understand this, which is why they order a real appraisal before approving a loan instead of trusting an online figure. The estimate is free and instant, and it is priced accordingly. The appraisal costs money and time because it answers a question the model can only guess at.
The healthiest way to read one of these figures is with respect and suspicion at the same time. Respect, because it is built on real data and gets you in the right neighborhood quickly. Suspicion, because it is blind to condition, timing, and the human side of a sale. The people who get burned are the ones who mistake the estimate for an answer. The people who do well treat it as a question worth checking. A home is worth what a ready buyer will pay for it, and no formula has met that buyer yet.
