Tableau: What made you look to Tableau for support with analytics?
Torry: For us analysts, Tableau Desktop has been a great tool to work with these large data sets that we have to find insights. It's quick and it's easy. It's been a lot better than digging through raw data. And we're starting to get more and more adoption, more excitement around Tableau Server with different people within the company, which is nice for us because we spend less time building reports for them and they spend more time actually utilizing the data that we have available.
Steve: I generate a lot of information that other people use to try to inform our higher-level executives and people all throughout the organization so that they're enabled to make intelligent decisions with data.
Torry: Ultimately we want our users to be able to use Zillow to find a home, which is the biggest purchase they'll make in their life and it's a very personal purchase. By finding ways to help make this process easier for them, we make it not so stressful to make this purchase that should be kind of a fun and exciting event.
Tableau: How are Tableau Desktop and Tableau Server proving to be valuable?
Steve: Our rentals dashboard on Tableau Server enables us to dive into states and feed providers with our rental listings and we can look for quality issues both by cross cuts across both those dimensions. We can really dive into where we have opportunities or issues with data quality and prioritize different regions to go after to acquire new listings.
Torry: A lot of the data sets we work with are quite large, and Tableau gives us a way to dig into that data and kind of quickly see where it is that something interesting is happening—whether there's an outlier in a certain area, or if there's a particular metric that we're looking at that's going to point us towards a big opportunity or possibly a problem with a product.
Tableau: How big is your data set?
Torry: We have data on every home in the U.S., and that's over 100 million homes—so that's over 100 million rows just in that database.
Steve: We've got lots of different metadata dimensions across contacts—where consumers are contacting real estate agents through Zillow? Was it on a mobile app? Was it on a for-sale property? Was it a rental?
Torry: We also have a mortgage product where lenders provide tens of millions of loan quotes every month to users who are trying to find the best rate on a mortgage—that's hundreds of millions of rows. Just being able to crunch that data into something manageable is where Tableau really helps us.
Tableau: How does Tableau help you examine that data?
Torry: With large data sets, we're looking at a lot of things like distributions and histograms and we're looking for anything that might point out an opportunity for us to possibly generate more revenue, or create something that helps our customers in their experience shopping for homes.
Steve: We can allow the product managers to basically slice all that data in Tableau Server by setting up a bunch of filters so that they can do the analysis that they want, and we just provide the data for them to make the insights.
Torry: Real estate is very geographical and because Tableau has such strong mapping support, it's really super useful to us.
Rather than just line charts and bar charts, we can provide more context so the business owner can be looking for patterns in geography. This is hard to do without a map. There aren't a lot of other products out there that have the built-in, drag-and-drop map interface.
Tableau: What are some patterns Tableau has illuminated?
Steve: Using this Tableau Server dashboard, we found there's a geographical discrepancy where some people label a unit as a condo, and others label it as an apartment. So we thought that we had this weird mix of units, but it actually turned out to just be geographical terminology. Once we saw that on a map, we knew how to come at the problem a little bit better.
Torry: With hundreds of millions of loan quotes, some users don’t get a lot of quotes for the particular loan they're looking for. By looking at the distribution of this data and how it's spread out across different loan types, we can figure out users who have particular credit score problems may not get as many loan quotes, or maybe if you want an investment property it's a little harder to find a lender who's willing to lend in the current kind of real estate environment. So we can look at all these different characteristics and see what type of response users are getting, and if they're having a good experience shopping for their loan on Zillow or not.
Steve: Our economics team has published a lot of things using the Tableau Public product. One really cool one that they've done is a negative equity. So if your house is underwater, you can zoom by county or zip code and see exactly how much your county or zip code is underwater on average.
You can really personalize the data as opposed to just having a national snapshot number in an article; you can zoom in and see how it affects you more realistically than just the average American. That's a lot more powerful, we think.
Tableau: Are you getting a good response online with Tableau Public?
Steve: We’ve had very high engagement rates with the Tableau Public products we've put out on the Web. People really enjoy seeing how they can dial into their neighborhood and see how the market is affecting them. We've got all sorts of metrics around growth of the market and home values. People are really interested to see how their trend in a small-market area compares to other neighborhoods.
Tableau: How has the use of Tableau expanded at Zillow?
Torry: We started by using Tableau Desktop about three years ago with roughly five users primarily on the analytics team to explore these big data sets and answer questions for different people from the business. It was more just for our analysts to dig in and look for insights.
Steve: It had organic growth from there as we started to disseminate that information. We'd embed it in e-mail reports and use it in presentations when running a meeting. People always asked where we got our graphs.
Torry: Now it's really spread throughout the company as a tool that we all use.
Tableau: How has Tableau Server impacted the way people at Zillow work?
Steve: Tableau enables people to have more a "pull" relationship than a "push" relationship with the data. So our small analytics team can better serve more users because we don’t have to answer every question; instead we provide a framework for people to answer their own questions.
Steve: It's just a great way to be able to interact with data as opposed to just receive data. That really adds another level of insight that you can gain by being able to not just be presented with something, but to be able to create something based off of a platform that's given to you.
Tableau: How has Tableau Server impacted the way your analytics team works?
Torry: About a year and a half ago, we switched to using Tableau Server so we could deliver more reports to users within the company.
Steve: It frees up a lot of resources for us to be a little more nimble and create more platforms than answers. I feel like it increases my efficiency as an analyst because I can serve more people and helps us to be more nimble.