What Is an Automated Valuation Model (AVM) in Real Estate and Should I Use One?

What is an automated valuation model (AVM)?

An automated valuation model is a type of machine learning algorithm used by various online tools and websites to estimate the value of a real estate property. An AVM uses a wide array of publicly-available and user-submitted data, such as property type, size, general location, and comparable sales data (when available) in order to provide an immediate value estimate, often available as quickly as a click of a button.

Different types of automated valuation models exist for different purposes. Commercial AVMs, such as Equifax’s Freddie Mac Home Value Explorer, provide property valuations for the residential mortgage and secondary/capital market industries to use throughout the mortgage loan lifecycle. Generally, these tools are only available to commercial mortgage lenders, investors, and appraisers.

A plethora of consumer-focused AVMs are available through sites like Zillow, Trulia, and HomeLight. The estimates that these online value estimators provide are ballpark figures based on the best available data — which means you should be careful when comparing these estimates to other valuations, such as a comparative market analysis (CMA) that an experienced real estate agent can provide, or a property appraisal which requires a licensed appraiser.

How do automated valuation models work?

Generally speaking, there are two ingredients to a value report produced by an AVM: a massive amount of data about as many real estate properties as possible, and a proprietary algorithm based on machine learning and regression analysis.

Types of data used in automated valuation models

In general, AVMs use two datasets to produce the most accurate property estimates: an initial dataset to calibrate the model, and an ongoing, expansive dataset to predict each subsequent property.

The calibration data set includes a representative sample of the types of properties that the model will be estimating and also brings in a reliable source of truth for the valuation, such as recently sold home prices. The source of truth data is instrumental in allowing the model to identify which of all of the other data points available for the property are most effectively combined together in order to make a reliable prediction.

The ongoing dataset pulls in as much information about the property as possible from public sources and proprietary sources, such as user-submitted surveys or website analytics. An algorithm that was trained with the calibrated dataset uses these data to predict the valuation as accurately as possible. This dataset is expansive and grows rapidly to include the most recent, relevant data for as many properties as possible.

Examples of property data used by an AVM:

• Property size (acreage)
• Home size (square footage)
• The number of rooms (bedrooms, bathrooms, etc.)
• General location (state, city, zip code, and sometimes even neighborhood)
• Home quality characteristics (air conditioning, pool, garage size, etc.)
• User-submitted information (ex. HomeLight’s simple home value quiz)
• Property price history
• Property tax valuation history
• Property historical sales information

How algorithms used in automated valuation models work

The algorithms that fuel an AVM use machine learning and regression techniques that take massive amounts of data in order to make accurate and reliable predictions. Regression and machine learning are buzzwords that you’ll hear often when talking about algorithms and mathematical models.

Regression is a mathematical technique that uses one set of data in order to predict another. When you have a calibration dataset, you can use both points of data (the data point you are trying to predict and the data points you are using to make the prediction) to create an equation or model. In this model, you can input only the second type of data, and it will output the prediction for the first data point that you are trying to predict.

Machine learning models go a step beyond regression by building a way for the algorithm to continually improve over time through a growing dataset. A machine learning algorithm can continually recognize patterns as new data comes in, and make changes to the algorithm to produce increasingly accurate and reliable predictions.

Why do different AVM tools produce different estimates for the same property?

You may have noticed that the same property can have different estimates depending on which AVM tool you use. Each of these tools takes both publicly-available and proprietary data and uses its own proprietary models to estimate values. Differences in the approaches to building the models, and the type of proprietary data, such as user-submitted data or website analytics, account for discrepancies in the value estimations.

Why do people use automated valuation models in real estate?

The primary allure of an AVM estimate is driven by two factors: cost and time. As opposed to other valuation methods, AVM estimates are instantaneous and typically free. There is significant power for an individual to take as little as two minutes, navigate to a website, provide a little bit of information, and immediately get a value estimate for their property. While there are other free methods to get a property valuation, such as a CMA, this is the only method that is both free and immediate.

There are multiple reasons for getting a quick and immediate value estimate, even knowing that this estimate may not be as accurate or reliable as more comprehensive measures. One of the most common uses is for homeowners considering selling who need an initial valuation to start planning. AVM estimates are helpful “If people are curious or just thinking about selling and wondering what their home would be worth,” says Bates, “or if they aren’t in tune with the market and they just need a general idea.”

Outside of preparing for selling a home, there are multiple other scenarios where a quick value estimate is helpful:

• Homeowners getting an estimate of their borrowing power through a home equity loan or cash-out refinance
• Individuals determining their asset allocation when building a diversified investment portfolio
• Individuals going through a divorce planning the division of assets
• Individuals looking to understand the tax impact of homeownership
• Homeowners looking to make adjustments or changes to home insurance
• Individuals and families focusing on asset allocation when estate planning

Drawbacks of automated valuation models in real estate

All of the most notable drawbacks of AVMs are products of its biggest weakness: lack of quality data. While it’s true that AVM tools have access to enormous datasets, there are multiple data points that are either too difficult to get in an automated fashion, or that may be inaccurate or wrong. Inaccurate and missing data results in estimates that may be representative of the general property but don’t take into account individual factors that may wildly affect actual values.

Bates cites several specific examples of missing data that cause AVM estimates to be inaccurate. “In non-disclosure states, such as Utah, [AVMs] lack actual data from the sale of a house, including the sales price.” Here’s a map of the multiple non-disclosure states.

“We’ve also seen these estimates significantly undervalue homes,” continues Bates, “because the public record often doesn’t include the basement square footage in its property information.”

There are multiple other limitations to AVM estimates that can affect their accuracy:

• The condition, including necessary repairs and overall cleanliness
• Recent upgrades or improvements, such as a new kitchen, roof, or bathroom
• Specific location factors
  • Its position on a busy street
  • Access to public transportation
  • School district
  • Zoning
  • Particular views or access to amenities such as a park or beach
• Major changes in the area or community, such as local laws or news coverage
• Announcements that cause hype, such as large employers coming to the area

Notwithstanding these limitations, AVM estimates can still be very helpful to use as a starting point or springboard when exploring the value of your home. As a rule of thumb, the more standard your home is to the other homes in your area, the more you’ll be able to trust an AVM estimate.

What’s the difference between AVM and other valuation methods?

As with many things in life, if something is both free and quick, it usually isn’t also the best or most accurate. Project managers are very familiar with the project management triangle, which takes time, cost, and quality, and posits that enhancing two of these points comes at the expense of the third. For example, you can have something that is both quick and high quality, but it will be expensive. Or, you can have something that is both high quality and low cost, but it will take time.

When quality and accuracy is required, such as in the cases of mortgage loan approvals or for other official purposes, it’s important — and often required — to get an appraisal. Estimating property value has historically been a manual process, and in many cases is as much of an art as science when considering unique factors that affect home values.

How can I request an estimate using an automated valuation model?

There are multiple sites that provide free online estimates. Some of these sites are more limited than others in terms of the data they have access to and the tools they use to collect additional data to be as accurate as possible.

HomeLight’s Home Value Estimator includes a short questionnaire with seven questions that can increase the accuracy of your home valuation estimate. These include simple information such as the property’s condition and the year it was built. We’ll pair your answers with housing market data from multiple trusted sources to get a real-world home value estimate in less than two minutes.

An experienced agent is a great option to determine your home’s value

An online home value estimate from HomeLight is a great starting point, but we recommend getting a full comparative market analysis (CMA) from a top real estate agent as a next step.

“There are no two homes exactly alike” explains Bates, “so getting an agent with access to the right data and knowledge of the local area is best.” An experienced local agent can factor in the elements that AVMs might miss in order to get you a more accurate understanding of your home’s value and help you sell it for the right price.

Use HomeLight’s Agent Match tool to connect with a top local agent in your area. The best agents can combine the art and science of home value estimates.

The Bottom Line

Automated valuation models are an important and helpful tool in a homeowner’s toolbelt to get a quick, ballpark estimate of a home’s value. The effectiveness and accuracy of an AVM depends on its access to data and the extent to which its machine learning model has learned to make better predictions. With increasing access to data and more time for machine learning algorithms to improve, AVM models will eventually become more accurate and useful than they already are today.

Find Your Perfect Real Estate Agent

We analyze over 27 million transactions and thousands of reviews to determine which agent is best for you based on your needs.

Find an Agent

But at the end of the day, there are factors that impact a home’s value that will likely never be incorporated into these models, no matter how much the data and algorithms improve. For official valuations, you should use a licensed appraiser. When you are ready to take action and sell your home, you should consult with an experienced local agent to find the best value for your home.

Header Image Source: (Roman Samborskyi / Shutterstock)