Case Study: Real-Time Insights Help Propel 10X Growth at E-Learning Provider Seesaw
January 28, 2022
Seesaw Learning Inc. provides a leading online student learning platform used by more than 10 million K-12 teachers, students and family members in the U.S. every month. The San Francisco company has grown steadily since its founding in 2013, with its hosted service in use in 75% of American schools and in another 150 countries.
Of course, when COVID-19 hit in early 2020 and forced schools to abruptly switch to full-time remote learning, the need for Seesaw’s platform skyrocketed. So did Seesaw’s growth, according to co-founder and Chief Product Officer, Carl Sjogreen.
“Due to remote learning, most of our key metrics grew by 10X,” Sjogreen said in a video interview with SiliconANGLE’s theCUBE as part of the AWS Startup Showcase in September 2021.
Wealth of Data, Little Observability
Those metrics included the data generated by existing and new Seesaw users as they interacted with the service. Storing all of that data was not a problem. Seesaw was able to scale up its main database, an Amazon DynamoDB cloud-based service optimized for large datasets. Seesaw’s database holds multiple billions of records.
However, making any of that data ready for analytics, and then sharing those insights in a timely fashion, was a different matter. And teachers and principals were clamoring for data such as which students were having trouble finishing lessons on time. While Seesaw employees were begging for internal application usage data to improve the customer experience.
“We had so much data that any question you asked immediately brought up five others,” Sjogreen said.
In other words, Seesaw desperately needed 360-degree real-time observability of its operations. And that was only possible if both internal and external users could drill down into the freshest data possible in order to get the answers they needed.
However, Seesaw’s DynamoDB database stored the data in its own NoSQL format that made it easy to build applications, just not analytical ones. And the batch-oriented analytical tools that Seesaw was using, such as Amazon Athena, were not up to the task.
“A lot of our data infrastructure was custom built and cobbled together over the years,” Sjogreen said. “We had a very disorganized data infrastructure that, as we’ve grown, was getting in the way of helping our sales and marketing and support and customer success teams really service our customers in the way that we wanted to.”
To fix this, Seesaw looked into building a traditional data warehouse paired with a set of ETL pipelines on top of DynamoDB, but concluded that it would take too much engineering work and not satisfy its high-performance needs.
“Plug-and-Play” Deployment
Seesaw turned to Rockset, deploying our real-time analytics database in early 2021 on top of DynamoDB. Like DynamoDB, Rockset can ingest, store and query massive amounts of data. For real-time analytics, the cloud-native Rockset improves upon DynamoDB by being able to simultaneously ingest massive data streams, indexing that data so it is available for queries within two seconds, and then enabling a high number of concurrent SQL queries. Results, even for complex queries, would be returned in milliseconds.
Rockset works well with a wide variety of data sources, including streams from databases and data lakes including MongoDB, PostgreSQL, Apache Kafka, Amazon S3, GCS (Google Cloud Service), MySQL, and of course DynamoDB.
“Rockset comes with all batteries included, including real-time data connectors with Amazon DynamoDB,” said Venkat Venkataramani, Rockset CEO. “You can just point Rockset at any of your Dynamo tables, even though it’s a NoSQL store, and Rockset will in real-time replicate the data and automatically convert it into fast SQL tables for you to do analytics on.”
As a result, Seesaw was able to deploy Rockset in no time.
“One of the key advantages of Rockset was that it was basically plug and play for our Dynamo instance,” Sjogreen said. “We were able, within hours, to start querying that data in ways that we hadn’t before.”
Actionable Insights for Educators and Seesaw Employees Alike
Once Seesaw began using Rockset to generate analytics and make product usage data available through SQL queries, it became increasingly challenging to move data out of Rockset and into business systems like Salesforce for their sales team to use. To work around this problem, the data team was forced to make various API calls and upload CSVs.
This manual process was extremely time-consuming, taking two days of developer time just to add a single field within Salesforce. Seesaw needed something simpler so that their data team could focus on high-priority tasks.
This led Seesaw to Hightouch. Hightouch is a reverse ETL solution that syncs data from various data sources to specific target destinations. Using Hightouch to sync the Rockset-powered insights directly to Salesforce, Seesaw’s sales and marketing teams can now view product usage data directly within Salesforce, enabling them to identify product qualified leads (PQLs), users with low engagement, and potential new customers. Instead of taking days to send data from Rockset to Salesforce, Seesaw now syncs data in minutes.
This is extremely valuable to Seesaw. As a product-led-growth company, Seesaw offers a bottom-up sales model where teachers can use the service for free, and will only approach school districts when a critical mass of teachers and students have adopted the service. By using Rockset to run queries in real-time on their massive data stores and Hightouch to sync Rockset analytics data to Salesforce, Seesaw can arm its sales representatives with up-to-date insights about how widely and deeply used Seesaw is specifically within a district when reaching out to its IT officials.
Rockset and Hightouch also work in parallel to help Seesaw’s product team narrow down the most-needed features. This led to the creation of a student progress dashboard. After less than six months, Seesaw decided to move all of its analytics to Rockset, while maintaining DynamoDB as its database of record and using Hightouch to operationalize the data.
- Customer usage data is generated in Seesaw and stored in DynamoDB.
- Rockset’s native DynamoDB connector automatically ingests and indexes all data within seconds, without ETL, to enable sub-second SQL queries.
- Query results are pushed seamlessly into Salesforce using Hightouch, a reverse ETL tool, to arm the sales team with up-to-date insights about their customers.
- Query results are also pushed to Retool to help the product and leadership teams visualize their analytics.
Seesaw Depends on Rockset and Hightouch To Accelerate Growth
As schools reopen, Rockset serves a key role in providing Seesaw key insights into how educators and students are coping with the return to classrooms. Hightouch helps Seesaw’s customer-facing teams leverage this information to increase growth and product adoption.
“The more we can understand how they [schools] are using Seesaw, where they are struggling with it as part of that transition, the more we can be a good partner to them and help them get the most value out of Seesaw in this new world that we’re living in.”
For Rockset, working with Seesaw has given us valuable insights into how to improve our service. Our Role-Based Access Controls (RBAC) security feature was originally a request by Seesaw.
For Rockset’s CEO Venkataramani, working with Seesaw has had its own personal benefits — bringing him closer to his two elementary-aged children who are both users of Seesaw.
“When I told them that Seesaw was considering using Rocket for analytics, they were thrilled, they were overjoyed because finally they understood what I do for a living,” Venkataramani said. And by working so closely with Seesaw, “for the first time I actually understood what [my] kids do at school.”
That’s observability.
Interested in learning more about this story? Watch the webinar below.
How E-Learning Platform, Seesaw, Scaled 10x During Shutdown with Rockset & Hightouch.
Rockset is the real-time analytics database in the cloud for modern data teams. Get faster analytics on fresher data, at lower costs, by exploiting indexing over brute-force scanning.