Tubi Uses Data as a Competitive Advantage
Tubi is a premier internet TV network for connected TVs, mobile devices and other screens, that offers a large collection of free movies and TV shows. Unlike many of its competitors, there’s no subscription fees or costs – commercials keep the platform cost-free.
With thousands of hours of video content available from hundreds of content partners, Tubi has billions of rows of data on its customers and the shows and movies they love. For years, Tubi has relied on Sisense for Cloud Data Teams to be the business intelligence (BI) platform that unifies the entire company around common insights and makes data-driven decision-making possible.
A Culture of Embracing Data
With mega-powers like Hulu, Disney, and Comcast as competitors in the digital content industry, Tubi has had to find creative and innovative ways to compete. Tubi doesn’t make its own content – it sources shows and movies through deals with other providers – so smart management of data is its secret weapon to help stand out from the crowd.
“We’re a small company relative to others in our space with a lot more capital than us,” said Rameen Mahdavi, Data Scientist at Tubi. “Getting a competitive advantage any way we can is extremely important. For us, that’s the speed at which we reach decisions – however we can get quicker, more accurate decisions through data, we’ll take it.”
To that end, Tubi has ingrained data literacy into every department in the company – they’ve become a data engineering-focused company masquerading as an entertainment platform. A majority of those working with data at Tubi have at least a passing knowledge of SQL, enough to create basic queries and build their own dashboards using Sisense for Cloud Data Teams.
“We pretty much embrace everyone digging into our data. That’s just part of our culture – our CEO and many of our business leaders have PHDs or engineering degrees, so they’re technically savvy and want to use those skills to explore data to aid their day-to-day work,” says Mahdavi. “We don’t necessarily look for SQL knowledge when hiring at our company, but we develop those skills internally.”
Soon after starting at Tubi, everyone from product managers and engineers to business development leaders will learn SQL and acquire an understanding of the data tables within Sisense for Cloud Data Teams – they’ll regularly write a few lines of SQL commands to develop data sheets, which quickly translate into graphs and simple visualizations. That process can be repeated to build upwards of 20 or 30 charts on a single dashboard.
“We’re not asking these teams to be using supercomputers or build the world’s most advanced algorithms,” said Mahdavi. “We see Sisense for Cloud Data Teams as a business intelligence system that helps us evangelize and scale data across the whole company. And it is really great for that.”
A Single Source of Truth
Mahdavi sees BI as the key to their quick decision making. For example, his team tracks metrics like “average streaming time” to learn how long users spend on the platform, then displays those metrics company-wide to inform Tubi’s content and marketing strategies in real-time. Tubi depends on being able to display data in ways that help individuals communicate with each other, and Sisense for Cloud Data Teams provides them with an intuitive, modern platform to aggregate those data sources and visualize them intuitively.
“Sisense for Cloud Data Teams allows us to go into the data, manipulate it, structure it very quickly, and then display it out to our team within a matter of minutes,” said Mahdavi. “It’s much quicker than other platforms we’ve worked with – we could still communicate data with combinations of other tools, but it wouldn’t be as quick to deliver results as it is with Sisense for Cloud Data Teams.”
In addition to streaming time, Tubi also examines its users to see how many ad impressions they produce, how many start and stop videos, how many install or register for the platform, and how many switch from competitive content platforms to Tubi. They also store financial information – each time an ad is served, Tubi tracks the financial impact within Sisense for Cloud Data Teams and shares it with advertising customers and partners.
“Sisense for Cloud Data Teams is really the central database for our business – it’s truly where we store all of our intelligence,” said Mahdavi. “Every single thing we do as far as tracking and measuring, we’ve put it into Sisense for Cloud Data Teams to manage.”
“Sisense for Cloud Data Teams allows us to go into the data, manipulate it, structure it very quickly, and then display it out to our team within a matter of minutes. It’s much quicker than other platforms we’ve worked with – we could still communicate data with combinations of other tools, but it wouldn’t be as quick to deliver results as it is with Sisense for Cloud Data Teams.”
Rameen Mahdavi, Data Scientist at Tubi
Flexibility for All Types of Analysis
As a lead data scientist, Mahdavi is responsible for monitoring the data sources that flow through Sisense for Cloud Data Teams and overseeing the dashboards to better understand how his team members communicate. As the various teams dig deeper into data, Mahdavi and other data scientists review their work in Sisense for Cloud Data Teams to ensure accuracy and consistency throughout.
For those who don’t have SQL or coding knowledge, there’s a strong system in place for them to be able to consume data and even do some basic manipulation of filters via Sisense for Cloud Data Teams’s data discovery for business features. Tubi’s CEO, for instance, does write SQL occasionally to analyze data himself, but he also consumes data by reading the charts that the data team provides every day. This flexibility ensures that as Tubi continues to scale, Sisense for Cloud Data Teams will remain an ideal partner to support the business.
“We’ll likely scale soon to hundreds of employees and have more people who need to analyze data without knowing SQL,” said Mahdavi. “But once we scale to that size, we can have our engineers and data team start building the data in a way that supports the drag-and-drop functionality within Sisense for Cloud Data Teams. We can already see that once we need to adjust or pivot our strategy, we can do that easily on the same platform.”
Moving forward, Mahdavi expects to further build out the use of Sisense for Cloud Data Teams’s Python integration, which creates an integrated workflow with SQL, Python and R all in the same environment. While SQL is great for working with data tables, the integration of Python enhances those tables with new visual options for data scientists and others with advanced skill sets.
“Python has a huge library of visualization options – it’s really where all our developers coordinate with each other to do more sophisticated analysis,” said Mahdavi. “So to have that functionality available directly within Sisense for Cloud Data Teams makes our lives a lot easier.”