Hitmetrix - User behavior analytics & recording

A Tale of Two Rocket Fuels

Once upon a time, a successful programmatic media-buying platform went to sleep…and when it woke up, it was an AI-driven predictive marketing solution!

Well, life isn’t quite like that of course, but Eric Duerr’s reaction when I told him I saw signs of Rocket Fuel rebranding itself was: “It’s great if you’re seeing that. That’s my job.”

Not that the Redwood City, CA-based vendor isn’t comfortable in its own skin. It’s repeatedly been ranked as one of North America’s fastest growing tech companies, and it serves major brands, from Denny’s to Toshiba. Rocket Fuel still does a lot of business, Duerr says, with traditional insertion orders, ensuring brands hit campaign objectives by helping them execute media buys in the fast-moving, large-scale programmatic space.

Beyond programmatic

“We’re leading companies beyond the traditional programmatic space,” Duerr emphasizes. The high profile 2014 acquisition of the x+1 DMP, and it’s integration with the Rocket Fuel platform at server level, created “whole new capabilities,” Duerr says. Some buyers wanted to keep data management and media-buying separate, Duerr says, “but we saw the way the world was going.” Namely, “activating data in real-time” — which calls for seamless collaboration between the data layer and Rocket Fuel’s higher level solutions.

On the one hand, brands doing what Duerr calls “interesting stuff” are buying into integrated suites of tools; he mentions IBM, SAP, and Adobe. On the other hand, single-point solutions — he calls them the “ankle-biters” of the space — are faster and more innovative. Open APIs, which make the larger marketing suites accessible to third-party solutions, are game changers.

“We don’t think of ourselves as a platform company,” Duerr says, acknowledging that Rocket Fuel does use the “platform” term. “We have data, and data storage, but we don’t have a developer community building Rocket Fuel apps.” What Rocket Fuel does enable, for example, is a brand using Adobe to integrate Rocket Fuel’s AI and predictive capabilities with Adobe’s core services. 

But why?

Real-time decisioning

There are two ways to use data, Duerr says: for storage or for access. A traditional CRM stores data. Rocket Fuel’s mission is to activate data to inform real-time decisioning. As a use case, he offers a Rocket Fuel integration with another, more generalized AI solution. “[That] AI, and ours, are connecting, and looking at every web experience.” Wait: Double the AI? “Theirs is general purpose AI,” he says, “ours is really good at a few things.”

In terms of web experience, the parner AI ingests the make-up of a page, including content and sentiment — “the page experience” — as well as ad placement data. Rocket Fuel brings the “ad experience.” Mining some 200 terabytes of data every data, Rocket Fuel, scores every single media-buy bid, based on contextual data (from time of day to weather) and a “holistic version of you.” The direct integration of data with programmatic buying allows Rocket Fuel to score “that moment.” Essentially, it’s predictive modeling of whether an ad will be effective.

Sometimes the score will be low — there’s a realization that the ad will be ill-timed or annoying for the recipient. Sometimes, “Hey! There’s receptivity here,” Duerr explains.

Interestingly, the gradient for success in any individual campaign is slow. The Rocket Fuel solution starts out with many low bids across the board. The feedback loop helps the AI understand what works and what doesn’t — “This is marketing that learns,” Duerr says. With time, and especially with large volumes of data, “the learning curve is awesome. The more data you have, the better it will get.”

Anticipating the consumer

There’s more to do, Duerr admits, like connecting probabilistic data (for example, from social media) with existing, rich, but primarily first-party, data sets. But it’s the future. “We get used to intelligent service,” he explains. “Anticipatory experiences are setting the bar.” If machine learning algorithms can do so much to steer us comfortably through our connected daily lives, they should be able to predict what we need to hear from brands, and where and when we need to hear it.

Being able to anticipate the moment and deliver the right message addresses, Duerr says, “a real pain that marketers are feeling.”

Total
0
Shares
Related Posts