What is the development of Zillow

Zillow grows in the recession with the help of self-service analytics

Tableau: Why did you consider Tableau for your analytics?
Torry Johnson: For us analysts, Tableau Desktop is a wonderful tool with which we can manipulate our large amounts of data from which we want to gain new insights. It's quick and easy to use. And it's much better than digging through raw data. And we're seeing more and more acceptance, more euphoria, from various employees in the company who are enthusiastic about Tableau Server. This is good for us because we save a lot of time by having to generate fewer reports for them and they spend more time actually using the available data.
Steve Brownell:I generate a lot of information that others then share with our boards of directors and employees across the company so they can make smart decisions based on that data.
Torry Johnson:Ultimately, of course, we want our users to be able to find a new home through Zillow. This is probably the biggest purchase of a lifetime and it's also a very personal one. We are finding ways to simplify this process. In this way we alleviate the stress of decision-making and turn the whole thing into an entertaining, even exciting moment in your life.

Tableau: How can Tableau Desktop and Tableau Server help with this?
Steve Brownell:With the rental dashboard in Tableau Server, we can call up individual states and send the providers there our current rental offers. At the same time, we can identify and correct quality problems by considering both dimensions. We have the opportunity to take a close look at where there are new opportunities or problems with data quality and to prioritize individual regions in order to obtain new offers.
Torry Johnson:Many of the data sets we work with are very large, and Tableau gives us the ability to evaluate this data and quickly see where something interesting is going on - whether there is an outlier in a certain area or if there is a certain characteristic somewhere occurs that may direct us to a larger business or a problem with a product.

Tableau: How big is your data set?
Torry Johnson:We have data on every home in the United States. That is over 100 million apartments and houses - so over 100 million lines only in this database.
Steve Brownell:We are dealing with many different metadata dimensions in our contacts here: Did the customers contact brokers via Zillow? Did you do it on a mobile app? Was it about a sale object? Was it a rental?
Torry Johnson: We also have a mortgage product where lenders send tens of millions of loan offers to borrowers every month. That's another hundreds of millions of lines. The mere fact that we can now properly manage this data in Tableau helps us a lot.

Tableau: How does Tableau help you evaluate this data?
Torry Johnson:With the large data sets, we look for parameters such as distributions and histograms or for something that shows us a new way of possibly generating higher sales or creating something new with which we can make the house or apartment search more pleasant for our customers .
Steve Brownell:Our product managers now have the ability to filter all that data into Tableau Server so they can do the exact analysis they need. We simply provide them with the data to get the information they need.
Torry Johnson:The real estate market is geographic and because Tableau has powerful mapping capabilities, the software is particularly useful to us. In addition to line and bar charts, we can provide additional context so that businesses can search for patterns in specific geographic regions that would be difficult to narrow down without a map. There aren't that many other products that have a built-in card interface with mouse actions.

Tableau: Can you give examples of the patterns Tableau showed you?
Steve Brownell:We noticed geographic discrepancies with the dashboard in Tableau Server, with some customers marking one unit as an apartment and others as an apartment. At first we thought we were dealing with a strange mix of housing styles, but then we found that there were simply geographical differences in terminology. Once we saw that on a map, we knew how to fix this problem.
Torry Johnson:Despite hundreds of millions of loan offers, some users are still not getting enough offers for exactly the loan they want. We'll look at the distribution of this data across different loan types and identify those users who have particularly low credit ratings and are therefore not getting as many loan offers. Or if someone is looking for an investment property then with the current situation in the real estate market it is a little more difficult to find a lender who is willing to provide the loan. So we can look at these different parameters and what kind of response the users are getting. And we'll see if they have a good experience looking for a loan on Zillow or not.
Steve Brownell:Our economists have published quite a bit with Tableau Public. The negative self-assessment, for example, was really cool. For example, if a home is flooded, owners can zoom in on that area by county or zip code and see how often the area is flooded on average. This makes the data personal and not just a national snapshot in a newspaper article to be taken for granted. The whole thing can be enlarged and you can see how far it actually affects you compared to the average American. We firmly believe that this is much more effective.

Tableau: Are you getting good feedback online in Tableau Public?
Steve Brownell: We have had a very positive response from the Tableau Public products that we have published on the Internet. Users enjoy being able to look around their immediate surroundings and find out to what extent the current market situation affects them. We provide many types of key data on the development of the market and home value. People are interested in the extent to which the trend in a small market segment compares to a neighboring one.

Tableau: How much has the use of Tableau spread at Zillow?
Torry Johnson: We started with Tableau Desktop about three years ago. Back then, it was about five users on the analytics team who examined our large datasets and answered questions from various people across the company, and back then it was more the analysts who should dig deeper.
Steve Brownell: Then, as we spread the word, the software started to become more widespread; we put it in email reports, and we used it in presentations at our meetings, and people kept asking where we got our graphs from.
Torry Johnson: Tableau is now being used across the company.
Tableau:How has Tableau Server affected the way Zillow employees work?
Steve Brownell: Tableau enables users to actively access data. You don't have to wait for it to be presented to you; it allows our small team of analysts to respond better to more users because we don't have to answer each question individually, but rather provide a framework that users can use to answer their own questions.
Steve Brownell: It's just a fantastic way of interacting with data, rather than just receiving it passively, and it gives us the ability to take another level of insight that is only possible because not just data is presented to you, but you can generate something of its own based on a platform.
Tableau: How has Tableau Server affected the way the analyst team works?
Torry Johnson: About a year and a half ago, we switched to Tableau Server so we could get more reports to employees across the company.
Steve Brownell: This gives us the opportunity to free up a lot of resources that we can use to work faster and create more platforms instead of trying to find answers. In my opinion, it has made me more efficient as an analyst because I can serve more people and we are faster overall become.