Zumper has created an online apartment search platform that aims to make renting an apartment as easy as booking a hotel. Zumper’s free online and mobile rental search platform is used millions of times a month to find rental apartments in the United States and Canada. The company’s paid service, Zumper Pro, lets landlords and brokers manage listings and screen potential tenants. Founded in 2012 and based in San Francisco, Zumper has raised nearly $40 million in venture capital from a variety of marquee firms.
As an online search platform, Zumper epitomizes a data-driven company. Part of that identity means that everyone at the company, not just the most skilled data analysts, needs to be able to find and use the full range of the company’s data, quickly and easily.
“We’ve pushed and trained our analysts to use SQL, ” says Woody Anderson, Chief Technology Officer of Zumper. “But our business experts rely on Excel spreadsheets, and SQL and Excel are not compatible. People are usually strong in one but not the other.”
We’ve pushed and trained our analysts to use SQL. But our business experts rely on Excel spreadsheets, and SQL and Excel are not compatible.
CTO at Zumper
At first, Zumper was using Amazon’s Redshift data warehouse, which requires SQL proficiency and is designed to be cost-effective for organizations running a high volume of queries. Zumper performs complex queries, but typically only four or five times daily.
“We weren’t running enough queries every day to make the cost/value function work for us with RedShift, ” says Anderson. “The cost to scale RedShift only made sense if we wanted to have a fully-grown analyst culture at Zumper—which is not our goal.”
Data analysis was taking too long, even for SQL pros like Anderson. “SQL is a nightmare to write by hand, ” he notes.
For their part, the spreadsheet analyses weren’t able to go deep enough. The company’s finance, product management, and other business teams needed more complete and immediate answers to their data-related queries than their spreadsheet tools could deliver. They also wanted better ways to visualize the results of their queries.
After consolidating their data onto the Snowflake Data Cloud, Zumper sought a business intelligence (BI) tool that would work for both their SQL and spreadsheet users. They chose Sigma.
Anderson and his team at Zumper asked Snowflake representatives to suggest a BI solution that optimized the Snowflake cloud data warehouse; they recommended Sigma. Business experts can explore the data directly, without SQL, coding, or BI tool expertise, while data professionals are freed from the time commitment of gathering requirements and creating reports for the business, instead focusing on the performance, governance, and architecture domains they are uniquely qualified to own.
“Sigma enables us to do what we already did, but way faster and better, and with fewer errors.”
CTO at Zumper
Zumper’s sales and marketing teams use Sigma to perform A/B tests to determine the behavior of visitors to their online platform, based on their traffic into and out of the site. They put traffic data in a log table, remove organic traffic, and point Sigma at the joins. The product teams use Sigma to look at customer conversion information, while the finance team uses Sigma for accurate and up-to-date reporting and forecasting.
Sigma has significantly streamlined — and vastly strengthened — Zumper’s analytics processes.
“What used to take me 17 steps—including rewriting underlying data—with Sigma takes at worst case three steps, and often only one step, ” says Anderson. “A query that might have taken 38 minutes to complete now takes only 38 seconds.”
Sigma also helps Zumper manage the evolving role of data at the company, which demands use of the freshest data possible while still keeping pace with web-speed responsiveness to its customers’ needs.
“Business users can take any additional data they find and merge it with the production data, without slowing down the launch of new features, ” says Anderson. “We see Sigma as a centralized tool with a good collaborative aspect. That, for us, is a win.”
Zumper has also benefited from working with Sigma and Snowflake together. “Sigma on Snowflake provides lots of value, ” says Anderson. “All the data is in one place in Snowflake, and I don’t worry about costs, even if our analysts aren’t running queries constantly. And information we’ve gained from people at Sigma has allowed us to use our Snowflake cluster more effectively.”
Anderson appreciates that Zumper didn’t have to choose between SQL and spreadsheets. “Sigma enables us to do what we already did, but way faster and better, and with fewer errors,” he says. “Everyone in the company has more trust in our data and how we use, present, and visualize it. That leads to more trust across our organization, as well as with our customers and partners.”
Zumper plans to continue expanding its use of Sigma throughout the company.
“If everyone at Zumper used Sigma, we’d be able to share information more efficiently, ascertain core information and make it reusable, spend time on the right insights, and be more likely to make better decisions, ” Anderson says. “A wider array of insights from a wider array of data can’t help but improve our bottom line.”