Okta’s cloud-based platform enables users to securely connect to any technology—from APIs to products to applications—from any device. Founded in 2009, the San Francisco-based firm serves more than 2 million users and thousands of customers, including Experian, Adobe, Nasdaq and 20th Century Fox.
Okta’s team of around 10 data analysts and data engineers supported an organization of more than 1,000. The team worked with two main sources of data: 1) back-end data about Okta’s product and users, and 2) data from Salesforce, which three-quarters of staff interacted with regularly, including the sales, customer success and marketing teams. A cloud-centric company, Okta stored available data in a Redshift warehouse, but much of its data remained siloed in business systems and was thus out of reach for analysis. With the company’s data engineers having to custom code data pipelines, there was a bottleneck in making data accessible in order to drive decision-making and innovation.
We always feel resource-constrained, and hand-building basic, repetitive ETL scripts is time-consuming and the least favorite activity of most data engineers.
“We always feel resource-constrained, and hand-building basic, repetitive ETL scripts is time-consuming and the least favorite activity of most data engineers,” says Cathy Tanimura, Okta’s Senior Director of Analytics and Big Data. Between coding, quality assurance, release and maintenance, the process of fulfilling a data pipeline request could take 2 to 3 weeks for a source they had integrated with previously. For example, to create a Salesforce data pipeline, engineers would have to write the code, QA the columns and definitions, then schedule it for the once-weekly release. “When you really need a column for a report, it feels like forever,” Tanimura says.
For a new data source, the process sometimes took several months. “Between figuring out the API, going back and forth with the vendor, the API not working as expected, and asking for modifications, it was a very painful process,” Tanimura says. “Most vendors aren’t expecting you to pull all of your data into your own database, and when they change something without notifying you, it breaks.”
The result was that Okta’s data analysts, who were embedded across business teams, had to wait a long time to get their hands on useful data. Many potentially fruitful sources of data—including hundreds of Salesforce tables—were out-of-reach altogether. Analysts and business teams wanted to access that data, but it could take months to reach the top of the priority list for time-constrained engineers. Each new ETL script added complexity to the system and further prolonged fulfillment times. At the same time, engineers burdened with fulfilling data requests had less time to focus on critical projects.
Okta’s data team needed a way to expand the availability of data without hiring a lot of additional data engineers.
Before learning about Etleap, Tanimura didn’t know there were cloud-based tools that could ease the process of building and maintaining data pipelines. “It opened my eyes,” she says. She researched a couple of other vendors before selecting Etleap, because the tool exclusively focused on cloud ETL and enabled custom transformations. Etleap loaded data into Redshift from tools Okta was using, such as Salesforce, Marketo and Jira.
This feels like a no-brainer: This is going to make our analysts more productive and let our engineers focus on other things.
When Tanimura made the case to executives, she emphasized cost-savings and immediate pay-off. “This is a bargain,” she told company leaders. “This feels like a no-brainer: This is going to make our analysts more productive and let our engineers focus on other things.”
Tanimura was initially concerned about whether a startup like Etleap could meet the stringent security requirements of a public company that provides security solutions to some of the world’s most high-profile corporations. But Etleap passed Okta’s extensive security review, which included resistance to attacks and regulatory compliance.
After getting Etleap up and running, analysts no longer needed to approach engineers to access a data source. Instead, they could use Etleap to create a new pipeline or modify an existing one without having to write any code. “The interface is very easy to use, and analysts are totally self-sufficient,” Tanimura says. Okta now turns to Etleap first when looking to bring new data sources into Redshift.
Thanks to Etleap, Okta’s data science team has been able to expand its productivity without making new hires. “Hiring data engineers is difficult and time-consuming. Etleap makes us look good because we’re providing a lot of value without throwing a lot of bodies at it.”
Hiring data engineers is difficult and time-consuming. Etleap makes us look good because we’re providing a lot of value without throwing a lot of bodies at it.
Instead of waiting weeks or months to access data—or never getting it at all—analysts can now get it within hours. That agility and independence has enabled analysts embedded with business teams to undertake projects they otherwise wouldn’t have, solving problems and spawning innovation. For example, a member of Okta’s project management office can independently analyze Jira tickets and report to executives. An analyst on the customer success team is able to do reporting based on customer sentiment surveys, easily refreshing analysis at regular intervals. Okta’s marketing team can analyze data from Marketo, generating insights for demand generation.
Meanwhile, Okta’s data engineers have been able to prioritize higher-value projects, including building a new streaming infrastructure and beefing up resilience management. “I’ve been really happy with Etleap,” Tanimura says.