By Dmitry Tsyplakov, Machine Learning Expert/ Digital Product Manager.

The key to a successful real estate investment lies in an accurate and quick property appraisal. At the same time, the estimates are both at the time of purchase and the possible dynamics of prices in the future.

At the moment, investors use three main ways to obtain this information:

  1. Hire a professional appraiser.
  2. Request a comparative analysis of similar ads from your real estate agent.
  3. Use an automated valuation model (AVM).

The first two differ only in cost and depth of analysis, but the latter mainly differs in the approach to valuation process.

How does an automated assessment work?

It is not difficult to guess that automated assessment is due to the work of artificial intelligence, which analyzes a huge amount of information, is able to compare hundreds of parameters in seconds and give a result. It has no “human factor” and “ blurred eyes”. Which is undoubtedly an advantage over the job of an appraiser or real estate agent. Typically, artificial intelligence is able to assess various factors that influence price formation. For example, the number of bedrooms and bathrooms in the house, the proximity of a school and shopping centers, or the presence of a pool, backyard garden. In addition, it analyzes the repeat sales index, the sales history of a particular property and similar offers. This method of real estate appraisal initially worked only with residential properties, but the interest of investors, lenders, as well as the development of a system of algorithms for training artificial intelligence, allow today to evaluate commercial properties as well. You also need a WordPress theme that does some basic things. You can check the top rated themes from Showcase IDX.

AVM market.

Many companies operating in the real estate market include in their services the work of evaluating properties using an automated model.

Zillow Zestimate .

The company Zillow one of the first introduced into their work this service, despite skepticism on the part of real estate agents. Nevertheless, the company’s employees believed in the idea, and today they are among the most popular on the market.

Data for analysis is taken from open sources such as reports from counties and tax inspectors, information from various listing services and brokerage firms. Zestimate publishes an estimate for over 100 million homes nationwide based on property characteristics, market data, OTC data and analyzing trends in the real estate market, including seasonal changes in demand. Thus, the price of an object may not be the most accurate, but it is an excellent break-off point for determining the final cost of a particular home.

Trulia AVM.

The company Trulia is part of the same group of companies as  Zillow. However, they use a completely different algorithm for how artificial intelligence works. This allows them to show price changes for a particular property over time, as well as compare prices for properties in the same neighborhood, city, county, or state. To form the final value of the house, the company provides data such as a crime map, information about local educational institutions, travel time to the selected unit, as well as suburban traffic and public transport.

Realtor.com AVM.

For the most accurate determination of the value of a house, Realtor uses three main data sources, each of which has its own unique AVM algorithm: Collateral Analytics , CoreLogic and Quantarium .

Redfin Estimate.

According to the statistics of the company Redfin , the frequency of errors in determining the value of real estate objects on their website is only 2.92%. They base their analysis on data from listing services, to which they have direct access, and data from real estate agents who describe their properties, in addition, information on recently sold houses can affect the final value of a house.

CoreLogic RealAVM.

RealAVM is one of CoreLogic ‘s products. To evaluate a property, this program uses public records of the characteristics and history of the house, information about recent similar closed transactions. Typically, the services of this company are used by lenders, brokers, real estate agents and consumers to compile for themselves primary information about the value of homes.


Initially, the company was focused on working with agents, brokers and especially with investors. But they recently launched a new project, ComeHome , which is a direct competitor to Zillow and Realtor.com. Their competitive advantage is the fact that the company uses several automated valuation models at once, which significantly increases the accuracy of determining the value of a house.

Realtor Property Resource (RPR) RVM.

This is a product of the National Association of Realtors . Accordingly, the basis for calculating the assessment of a particular object is the analysis of the country’s largest real estate database. Is it worth clarifying that only members of the association can use this assessment? Nevertheless, according to the association, the RPR program is able to facilitate the work of a realtor and help to quickly respond to customer inquiries.


Machine versus man.

Despite the fact that artificial intelligence is developing by leaps and bounds, capable of processing large layers of information with lightning speed, today it cannot be compared with the in-depth analysis of a professional real estate appraiser. Since there are factors that significantly affect the final price of a house, but cannot be digitized by a machine. Nevertheless, the interest in this topic from experts, investors, developers is so great that it is possible that in the very near future he will be able to completely replace a person in this routine work. In the meantime, this is a great helper in order to determine the initial cost of a house, find out the price range for such offers and know what to start from in finding the perfect property.


Leave a reply

Please enter your comment!
Please enter your name here