Ad-hoc Reports and Complicated Code Result In Low Analytics Adoption
As a data focused company, Teachable struggled to build a truly self-service analytics and BI stack. Their previous analytics tool, Looker, required knowledge of a complex business logic layer called LookML, which prevented most of the company from running their own reports.
“A few real power users in the company could do their own work in Looker or directly in SQL,” says Peter Jaffe, Head of Data at Teachable, “but the complexity of Looker was too high a bar for most users to get past. They could use the platform to consume reports prepared by the data team, but virtually no one was able to generate or even edit their own dashboards.”
Even straightforward, predictable requests had to go through the data team. As a result, two full time employees were spending the majority of their time answering ad-hoc requests, time to insight took weeks, and adoption of the tool was low.
“We have made a lot of improvements over the past two years and rebuilt the data warehouse from the ground up. But it was clear that we would not come close to reaping the benefits of the new architecture if we continued to put it behind the wall of Looker,” says Jaffe.
Governed Data Access for All Delivers Insights in Seconds
After evaluating a few different analytics tools, the team at Teachable chose Sigma because of its ease of use, especially among non-technical users.
“Two main things attracted us to Sigma when we were selecting a platform: One was the ease of use for ‘non-data’ people, and the other was the operational model in which end users can add and connect to datasets on their own. This was a huge selling point for me: Any user in the company can connect to and join any table they have access to in the data warehouse, without need for the data team to set anything up. This is the true ideal of democratized data,” explains Jaffe.
Connecting Sigma to Redshift, Teachable’s cloud data warehouse, and assigning permissions to teams was finished within the first two days.
Almost immediately, teams across the organization were able to build their own dashboards and directly ask questions of their data using Sigma’s spreadsheet-like UI.
One interesting use case comes from the product team. Now armed with Sigma, product managers have begun prepping dashboards before product releases so that when new updates are rolled out, the team can immediately see how these features are performing.
“We have a new product called Coaching where people can sell blocks of their in-person time through the platform. The PM for that release posted a PDF of a dashboard she had setup and scheduled with Sigma. Everybody was excited that something like that could be posted automatically. It gave everyone a lot of immediate insight into how the product was doing that otherwise would not have been seen,” reveals Jaffe.
Another team now benefiting from Sigma is the fraud team. The risk of bad actors trying to exploit the platform is an ever present risk. To counter this, Teachable’s fraud team pulls lists of schools to investigate every day. But with the explosion of new online courses, it’s been a struggle for the fraud team to keep up with demand. Sigma has enabled the team to quickly adapt to the new volume of accounts, act on hunches quickly, and limit the amount of risk for the company.
“Before they had Looker and dashboards created for them from the data team, but they had no flexibility. They now have a lot more flexibility to follow their own hunches.”
Head of Data at Teachable
“Before they had Looker and dashboards created for them from the data team, but they had no flexibility,” Jaffe explains. “They now have a lot more flexibility to follow their own hunches. We have been experiencing explosive growth and new schools have increased. The fraud team’s ability to tweak their own dashboards has been pretty important.”
Product Feedback Cycles Now Happen in Hours, Not Weeks, and 5X Increase In Analytics Adoption
The most obvious benefit to adoption of Sigma has been the reduction in ad-hoc requests, which has freed the data team up to do more impactful work and unlock greater value for the company. “We have reduced two full time employees doing virtually nothing but answering ad-hoc data requests to half of one full time employee’s time spent on ad-hoc requests,” claims Jaffe.
“I had a goal of reducing ad-hoc data requests by 70%, and we have easily passed that. Business users are able to make their own data pulls through Sigma, and the requests we do get are for much more complex data modeling and analysis. As a result, the role of data analysts at Teachable has been transformed from running rote SQL jobs to something much more thoughtful, and more valuable to the company.”
The role of data analysts at Teachable has been transformed from running rote SQL jobs to something much more thoughtful, and more valuable to the company.
Because teams across departments are able to quickly adopt Sigma into their workflow and build their own dashboards, data-driven decision making continues to grow across the company. Jaffe revealed that 5X as many people have created top 20 dashboards (by usage over the past 90 days) in Sigma since migrating from Looker.
By building custom tailored dashboards around a specific set of new features, the product team is now able to rapidly surface insights that shape product development and strategy. The new product feedback cycle has been reduced from 2-4 weeks to hours, as product managers are now able to design their own performance dashboards.
Reflecting on the impact Sigma has had on not just the data team, but the entire company, Jaffe said this: “End users tell me, unprompted, that they love Sigma. I have used a variety of BI platforms at a number of companies in the past, and I have never seen user adoption like this.”