.png)
.jpeg)
“Narrator drastically improved the time it takes our analysts to answer questions. And with the Activity Schema it’s super easy for our team to manage the data powering all of that reporting.”
Like a lot of large companies, PBS uses many different microservices. This can be great for engineering autonomy, but makes it super difficult for analysts to answer questions that require data from multiple systems.
Rusty led the effort to integrate all of that data into one data warehouse so all data was unified. However, making sense of the data to answer strategic questions was a tedious and time-consuming process that only allowed them to handle a few requests at a time.
“Analysis before Narrator was very one off, slow, and required a lot of manual work because there was no process and tooling for it. We didn't have any sort of precedent for getting that data and putting it together” says Rusty.
When he first learned about Narrator he saw a huge opportunity to level up their analytics. “Every reporting solution I used — you need to anticipate what queries you’re going to run in advance. If you haven’t anticipated correctly, or if your existing models can’t handle the question, then now it takes forever to answer a single request. With Narrator, you set up these simple building blocks and assemble them in the UI to answer infinite questions. You can also grow your reporting solution as you go instead of having to think of all the queries you have to answer upfront like you have to with most of the tools in the modern data stack.”
Narrator works on top of the Activity Schema data model, which models customer actions instead of facts and dimensions like a traditional star schema. It took PBS’ data team less than a day and a half to set up their Activity Schema from scratch. From Rusty: “What makes Narrator so simple is you don’t have to think about joins or how things relate when you’re setting up the activities. You only have to think about each activity by itself and just map the raw data to that single activity concept using simple SQL.”
It was also an easy choice in comparison to dbt: “I’m a huge fan of dbt and expected to be using it a lot more. In dbt’s world of production models and nested dependencies, any change upstream can cause a cascade of changes. With Narrator, it’s all separate and orthogonal. All datasets used for analytics are based on the Activity Schema, so there’s only one layer of dependency. The easy setup and management of our Activity Schema made Narrator a no-brainer.”
At PBS, they’re now encouraged to ask an infinite number of follow-up questions, creating a culture that feeds curiosity — that liberates the business to ask, iterate, and innovate.
“Narrator allows us to focus on the strategic questions we can answer from the data that we have. Whether X feature or Y action led to more subscriptions or donations for example” says Rusty.
“When it takes only a few minutes to get what you need, we want every question to lead to 2 or 3 more. Now everyone is asking questions and using data to make better decisions.”
