Start Simply: Discovery Wins Big With Big Data

Share this post

Five years ago, Discovery Inc. refocused its corporate marketing division to be more data-oriented — a move that would create a critical shift for the function.

The initial goals were modest. Discovery created a media analytics division in the hopes of
improving its ROI and maximizing ratings outcomes for its portfolio of networks. In creating their own marketing data and technology infrastructure, the Media Strategy & Analytics (MSA) team uncovered ways to spend money more efficiently and effectively while lowering their conversion costs. The initiative was so successful that the MSA team ended up consulting more closely with finance and marketing on budget and strategy.

AdPredictive founder and CEO Dan Carroll noted that marketing, media and analytics teams are actively stepping up to lead conversations about big data and optimization within their organizations. With the vast volume of meaningful, structured data available to them, marketers are often the first within an organization to have the opportunity make an impact in the efficiency of the business.

To that end, Dan recently interviewed Seth Goren, SVP Media Strategy and Analytics, Discovery, about how data can transform thinking within a company and why it’s so important to start simply.

Q: Your company underwent a monumental restructuring about five years ago. What
spurred the change?

There were three main reasons for us to refocus our corporate marketing division.

The first was that we were no longer acting like a traditional media client — one that relies on an outside agency to craft a media plan, and then gets internal buy-in for that plan- in other words, functioning as a ‘traditional’ client. Now we have teams of people working in-house, monitoring various platforms and buying ads in real time.

Second was the need to apply data science to our decision-making, and to use predictive
analytics to answer questions we once thought unanswerable.

And the third was measurability — we wanted the ability to more accurately talk about attribution and the role marketing plays in real business outcomes.

Q: How has measurability affected your role with senior management?

It’s helped us communicate better with our finance and management teams, specifically. We’re now able to tell them what the expected results will be from a set of choices. We’re able to better articulate the ways in which certain marketing decisions did or didn’t drive business, and that gives us a lot more credibility. As such, we’re kind of an extension of our marketing and finance teams now.

Q: How does data factor into how you evaluate vendors and integrate new technologies?

Being an in-house team, our conversations with vendors are fundamentally different than, say an agency. It’s not about how vendors can help us with this one RFP for this one campaign. We ask how they can help us improve the larger decision-making infrastructure so it’s more empirical, more scientific. If you can’t add value in that respect, then it is a wasted call.

Q: Was it a difficult transitioning, learning to be more data-driven?

Unfortunately, the deeper you get into this world of data architecture and ad tech, the more painfully aware you become of your flaws. All you can do is to do your best to shine a light on those dark corners and work with the right vendors to attack those problems aggressively.

Q: What’s the inevitable conclusion of this data-driven marketing infrastructure?

Because we started building our infrastructure five years ago, we were disproportionately
focused on linear ratings outcomes. There’s been a paradigm shift in how people view TV, of course, and we’re now focused on predicting a wider range of outcomes, including streaming and subscriptions. We are aiming to know the full impact of the budget that goes toward traditional tune-in, SVOD, and AVOD.

Q: How were you able to convince other divisions in the company to buy into your team and its data practices?

Ultimately, the financial case ends up being quite compelling – spend less, more efficiently, and build your data and tech infrastructure in the process. We like to position ourselves as the objective source for distribution of marketing capital to targeted audiences. We can prove when you make X decision, you’ve reduced your chances of Y outcome Z percent. That’s a high standard of accountability.

Q: What’s the next step in this evolution?

Artificial intelligence, machine learning. Platform convergence is another area we’re looking at. We’re trying to connect these systems to improve how and when we sequence our messages and at what price.

Q: What advice do you give people who are just starting to implement ad tech and big data into their organizations?

Start with one problem. Solve it. Things will take off from there. Our initial problem, for example, was the way we made media mix decisions. Our process was unscientific and flawed, and the solutions on the market were imperfect so we said, “let’s create our own media mix model using our own historical data, built against our ratings outcomes.” This project ended up unlocking significant functionality that we never originally envisioned such as return curves and machine learning simulators.

We then built an optimizer with AdPredictive to run spending algorithms, and found ways to sift through millions of combinations of decisions to hone in on choices that will maximize exposure against key target audiences.

Discovery’s commitment to its marketing intelligence infrastructure shows precisely how big data can and should deliver benefits to more than just the team in charge of running banner ads. The company has developed a data-fueled approach to increase the efficiency and value of its media while delivering tangible, bottom-line results across the entire organization. AdPredictive was chosen by Discovery a key partner to overcome data and media hurdles by building fully functional planning tools and workflow products.

Share this post
1 Comment

Leave a Reply

Your email address will not be published.