Online real estate information sites are routinely used by homeowners, home buyers, real estate agents, and developers as gauges of a home’s market value.
Some of these sites use automated valuation models (AVMs) to provide information on estimated market value, usually for homes currently on the market. The information is available on many home search websites across North America.
AVMs are often proprietary, but generally these use algorithmic calculations that take into account characteristics such as square footage of both home and property, number of bedrooms and bathrooms, footprint of the structure, and property tax information and prior sales prices for nearby homes. The factors are weighted differently according to the specific geographical location.
Consumers often rely heavily on AVMs – sometimes too heavily. According to industry executives, some have median error rates of 8 percent. Of course, the accuracy of the information varies according to the individual website.
In fact, AVM error rates vary widely, and in some places they far exceed the national median. In large urban areas, as well as in desirable suburban and waterfront locales, AVM estimates may be off by tens of thousands of dollars in either direction. Sometimes online valuations are higher than actual on-the-ground selling prices, and sometimes they are significantly lower.
There are several reasons why AVMs can be off. For one thing, an algorithm can’t determine the actual physical condition of a residence. Sometimes homes have specific characteristics that add to or detract from value, such as a poorly located bathroom, a tiny bedroom, an unusual layout, or an obstructed view. Moreover, AVMs don’t take into account title issues, such as concern over surveys or boundaries.
Many agree that AVMs can be a good place to start when you’re in the research stage of house hunting. However, once you’re in search mode, forewarned is forearmed: ask your real estate agent for more detailed comparatives.