The development of internet sites dedicated to real estate property in the United States is one of the most striking developments of the U.S. real estate market over the past 10 years. Today, the leaders of the sector, Zillow and Trulia, allow millions of owners, tenants and investors access centrally to a lot of information in relation to their property and real estate product search: location, pictures, taxes, rent estimated and of course estimate of the value of a property.
At Zillow, the estimation of the value of real property is called the “Zestimate”. According to the website, more than 110 million of real estate at the moment have a Zestimate. Also impressive either, what is hidden behind this statistic and how is it calculated?
The “Zestimate” is above all an algorithm that incorporates a variety of real estate data public and grouped by Zillow itself. This algorithm is not made public by Zillow indicating however use the following parameters:
‘Physical’ statistics: location, area, number of rooms and bathrooms, etc.
The Tax assessment of the property: value set and used by the counties for property taxes (note this value has nothing to do with the market value of the property)
Transaction history: when it is known
Why Zillow estimates are not always reliable?
1. The data which has Zillow is sometimes incomplete. And Zillow acknowledges this himself. Zillow has a grid outlining the accuracy of the data used to calculate the Zestimate according to areas. As shown below, on a scale of 1 to 4, Zillow does not always have complete information to determine its Zestimates. When nearly half of these are in the range of +/-12% of the actual sale price, the difference in nominal can quickly be raise 20 000 USD or many over…
2. The comparable used are not always relevant. Comparative analysis of near similar properties of the target product, if relevant, the used parameters must be extremely precise and realistic. However if it does not find enough comparable for example, Zillow sometimes expands the area comparable in far too much way. Now we all know, the quality of a neighborhood can evolve from one street to the other, especially in the United States.
3. Zillow integrates hardly products of foreclosures, foreclosures, auctions, etc. If Zillow has data on these products, they do not necessarily correspond to the reality on the ground. For example, a price indicated by Zillow may reflect the price of the auction and not the price at which it will be actually purchased. Similarly, products that are purchased in bulk or via more complex schemas will not be represented reliably on Zillow.
4. The estimate of the value of real estate cannot be reduced to a single algorithm. Zillow does not visit the 110 million of goods for which he produced a zestimate. It can therefore be integrated factors such as the actual condition of the property, interior finishes, the addition of a room more, work, etc. Zillow cannot replace the crucial role for that of a Realtor that will produce an estimate also based on his experience. Finally, the value of a property also depends on the market and offer and existing demand on the box or the target product.
How to use Zillow?
Despite the limitations of its algorithm, Zillow and its competitors, offer remarkable tools, especially on the scale of a country as large as the United States. The usefulness of Zillow rest whole when it comes to find the characteristics of a product or a box to obtain information of synthetic and centralized database.
But, even though Zillow will continue to improve its algorithm, its model has limitations and rely only on these data during an acquisition is not recommended. As with the purchase of a car, real estate requires a “technical control” and look experienced professionals.