Automated home value estimates have been a popular tool for real
estate search websites, and while these tools may satisfy clients’ needs for
quick information, in many cases the information they provide is inaccurate.
Although many agents and brokers are aware of the limitations of these
models, many consumers are not, and improper use of these tools can encourage
mistakes.Automated valuation models (AVM) are designed to predict a home’s price based on comparing it with similar properties in the area. They do this using county property record data from thousands of offices around the country, comparing several attributes such as square footage, the number of bedrooms and bathrooms, and other property features. This data cannot take into consideration the specifics of the neighborhood or the details of the properties (lot size, improvements, quality of the maintenance…). It is my experience that while they can be a good first indication, and sometimes accurate, they miss the mark in roughly 30% to 40% of cases. Where all homes are similar, like in a condominium complex or a given tract of identical homes, they may be quite accurate. In places where each home is unique, in nature or in location, they can be quite off (think: Los Altos Hills, or Palo Alto to cite a few).
As always with statistical data, a percentage of the evaluations will be
accurate – but be careful it is not by mistake. Also, figures do not know the market conditions, which can only be understood by a
professional dealing with local sales 24/7.
I do not mean to be critical, but I just want to remind clients that
these valuations should just be a first indication, not the final opinion of value. Another way to look at it is, as we like to
say in the business, "home valuation is not an exact science".Read more details and the fullRE Insider interview.
Thanks for reading,
Francis
Silicon Valley real estate specialist
Detailed, local trends etc...
Current mortgage rates
