Fiverr Turns to Sisense to Get the Fastest Refresh Rates
Complex dashboards refresh in just seconds for a company that’s built on user data.
Pain: Management lacked real-time look into user behavior on website
Solution: Interactive dashboards automatically refresh in seconds and can be accessed and shared in real time
Pain: Scattered data from multiple sources
Solution: Data easily joined in the Sisense ElastiCube for a single version of the truth
Pain: Growing data and multiple concurrent queries slowed down the BI system
Solution: Running Sisense as the back-end handles dozens of concurrent queries and large, complex data sets decrease refresh rates to just seconds
Pain: Only a dedicated BI Developer had the know-how to build reports
Solution: Easy report building with drag-and-drop joining of data sources enables company-wide adoption
Company: Online Marketplace for Services Starting at $5
Fiverr is a global online services marketplace with services starting at five dollars and up to thousands of dollars. With close to 4 million services and with offices in New York, Tel Aviv, Miami, Los Angeles and the Netherlands, Fiverr is transitioning from a startup to a full-fledged corporation. Fiverr has a team of data analysts dedicated to BI but the company uses Sisense Business Intelligence for reporting in all departments including product, marketing, PR, business development and upper management.
Challenge: The Need for Quick Insights on Growing Data
For Fiverr, company data is the most proven source of insight. Fiverr believes that in the world of e-commerce, gaining the upper hand is no longer about looking at what competitors are doing; it’s about learning from users.
Fiverr wanted to connect data from MySQL with data from Google Docs, Spreadsheets and Analytics in order to better track ‘events’, user actions on the company website and mobile app, and see the bigger picture. As users increased, so did Fiverr’s needs. Data became larger and more complex with millions of rows a day from a variety of sources. Though Fiverr uses an internal big data system based on Hadoop to track their user behavior, data got too complex for the team to easily and quickly build reports and dashboards.
Slava Borodovsky, Fiverr’s senior BI director and data scientist explained the company’s approach to data: “We are constantly changing the way we need to look at our data. We change the schema, we add tables, columns, and calculations because everything is dynamic and we need as close to real time results as possible.”
Fiverr’s strength is its ability to develop features that its audience needs and loves. The product department uses BI as a fundamental element of their development process to determine their product roadmap. “We don’t deploy features without understanding their value first,” explained Borodovsky. “As departments began understanding the impact of data on their success, the need for data became more urgent.”
Contenders: Sisense vs. Traditional BI Vendor
When Fiverr hired Borodovsky as their BI director he was already interested in using Sisense as the BI platform since he had familiarity with Sisense from a previous company. After a frustrating experience with a traditional BI Vendor due to its complexity and need for heavy technical skills, he was happy to come back to Sisense.
“Within seconds we can run a report with over 20 queries and see real-time results.”
“When I worked with traditional BI tools and OLAP cubes, it took too long to refresh a dashboard–even our best performing dashboards lagged.” Borodovsky recalled. “With traditional tools, you need someone that is dedicated and completely trained in their environment to build reports.
Sisense on the other hand enables Fiverr to combine multiple large data sources and gain complete transparency into what is going on in every part of their website and mobile app. After trying Sisense, Fiverr was able to reach their goal: “Within seconds we can run a report with over 20 queries and see real-time results,” Borodovsky said.
Deployment: 2 servers with 5 cubes defined by frequency of updates — 1 cube refreshes once an hour, 4 cubes refresh once a day.
Data: Tens of millions of rows on every commodity server (64GB).
Number of queries per dashboard: 20.
Consumers: All departments use Sisense, with product and marketing departments as the most active.
“From day one it was incredibly easy to manage the Sisense ElastiCube and with almost zero maintenance.” says Borodovsky.
Sisense is now Fiverr’s main reporting platform, with a team of analysts using it to independently build reports and dashboards, and is deployed across all departments. The independence of users is a real game changer for Borodovsky: “With Sisense all my analysts can build reports and dashboards, change and join data sources, and get insights completely on their own.”
“With Sisense all my analysts can build reports and dashboards, change and join data sources, and get insights completely on their own.”
The product team at Fiverr makes key product decisions based on user behavior that gives insights into what new features are needed and how a feature is resonating with their users. “From the beginning we were data driven, but since Sisense, it’s in our DNA,” Borodovsky said.
“From the beginning we were data driven, but since Sisense, it’s in our DNA,”
Sisense dashboards can be found on TV monitors hanging on the office walls in every department, including the online marketing, product and dev teams, and upper management. “Upper management loves the scorecards with classic red, yellow, and green gauges,” Borodovsky added.
Closing: Refreshing Dashboards in Seconds
“There has been a change in behavior since the TV monitors with Sisense dashboards went up– departments are more aware and motivated to meet their goals,” reported Borodovsky.
“The technology behind Sisense is incredible, it utilizes memory really well and end users can really feel it.”
Borodovsky says that it’s hard to compare Sisense to other BI tools he has used before because Sisense is full-stack, serving as both a database and a reporting tool: “The technology behind Sisense is incredible. It utilizes memory really well and end users can really feel it.”
The bottom line for Fiverr is that Sisense allows it to create a completely agile environment to make quick and smart decisions based on its customer user behavior. Ultimately, Borodovsky feels that, “with Sisense, Fiverr is able to move fast and learn fast.”All Case Studies View Live Demo