Automated valuation model (AVM)

What is an automated valuation model?

An automated valuation model estimates the price of a property based on data. These models often use a lot of similar data that traditional valuations use. Data includes public records, prior sales prices, sales prices of similar homes. Automated valuation models are often used to provide fast estimates at a large scale.

What is an AVM?

AVM is the commonly used acronym for automated valuation model.

Why do automated valuations models matter?

AVMs are influential in modern real estate pricing. Real estate buyers and sellers often rely on data from an automated valuation model. Often they see this data before consulting a person. This is because sites like Zillow often display price estimates based on an AVM. And most home buyers and sellers start their research on real estate portals. As a result, real estate professionals must understand how these AVMs work. Agents often have to advise clients on AVM accuracy and influence.

‍

In addition, if the calculation method of an AVM changes, it can have a meaningful impact on home pricing.

‍

It’s important to note that AVMs not only estimate prices but  also influences them. AVM pricing can anchor buyer and seller expectations.

How accurate are automated valuation models?

Major automated valuation models claim accuracy within 2% of the final sales price 50% of the time. On one hand, this is an impressive accomplishment of modern technology. On the other hand, this might simply be cherry picking the 50% of properties with easier estimates. The remaining 50% of estimates may be way off.

What are some popular automated valuation models?

The most well-known automated valuation model is the Zillow Zestimate. But Redfin, Trulia, and Realtor.com all have their own automated valuation models.

What are some criticisms on automated valuation models?

Most real estate agents have several key criticisms of automated valuation models:

• Limited accountability: AVM estimates have no penalty for being wrong. The accuracy is within 2% of pricing 50% of the time. But misses on the other 50% only hurt agents trying to provide better pricing information.
• Data inconsistency: AVMs rely on data. If they have weaker access to data in certain markets, that can skew prices. An AVM with great New York data may provide terrible estimates in Minnesota.
• Lack of human judgement: AVMs might be accurate based on housing data and statistics. But more-nuanced components like renovations and marketing impact can be overlooked.  overlapped. This can favor homes in poor condition. They may need more work but will look similar on paper as nicer homes.

What are some examples of automated valuation model software?

CoreLogic, Attom Data Solutions, and HouseCanary all offer automated valuation model software. These services help companies market their own AVMs to customers. These services also help customer request AVMs on many properties at once. They also typically provide more data and insights than real estate portal AVMs do.