The lines between streaming and traditional TV ad tech and viewing habits are only getting blurrier.
DatafuelX, a predictive analytics startup founded in 2021, is a data and insights company that aims to optimize the value of TV and digital media. It is also one of many startups born of the need to capitalize on the trend of channel convergence. But the company has linear roots: Its foundation centers on the business of data-driven linear (DDL), and most of its early clients are traditional broadcasters, such as Hallmark and Fox.
Earlier this year, the startup rehired Dan Aversano, one of the company’s co-founders, to take the helm as CEO. Aversano will lead the company through its next growth phase, which revolves around – surprise, surprise! – connected TV, according to the company.
After co-founding datafuelX, Aversano worked at TelevisaUnivision, where he was responsible for setting up an advanced advertising business. Before that, Aversano spent nearly 10 years leading analytics strategy at Turner and WarnerMedia (both before and after AT&T’s acquisition).
Last year, datafuelX’s board was considering selling the company. Aversano and a few other shareholders dissented, insisting that the company is still in a promising growth phase and has the potential to multiply its revenue far beyond what it is today thanks to the quality of its technology and analytics.
“A couple of other shareholders and I were lobbying the board to say there’s more here,” Aversano told me.
“And that was the bet that they ultimately put on me.”
AdExchanger: What was the rationale for establishing datafuelX five years ago?
DAN AVERSANO: I noticed a few industry trends and opportunities for improvement during my time at Turner-then-WarnerMedia-then-AT&T.
One of the more obvious trends was the intensifying push for data. Every agency demanded more data, holding companies were buying identity solutions and media companies were investing in their own first-party data.
But what I noticed was the disparity between data science and media strategy. In other words, not all engineers working on data science teams for media and advertising know the intricacies of how our industry functions. If you have a world-class data scientist who doesn’t understand this ecosystem, that’s a minimum two-year learning curve. That’s an expensive and prolonged learning curve that the industry can’t really afford at a time when it’s changing and evolving so quickly.
This was the sort of insight that inspired the creation of datafuelX with the aim of establishing data forecasting and analytics to help optimize convergent media buys.
To date, the company has been growing between 20% and 30% in annual revenue terms. I think that number can reach 50% with a couple of tweaks.
And what are those tweaks?
Expanding our existing technology and applying it in new directions – meaning, cross-platform use cases.
One of our core products is M3, our yield optimization platform for publishers that powers our DDL business. This year, we’re focused on pushing aggressively to increase adoption among digital publishers across connected TV, free ad-supported TV (FAST) and online video. Anyone who owns inventory needs to better forecast demand for their inventory. And we believe that inventory should be aggregated because three-quarters of the work involved in actual media decisioning still lives in Excel. Better consolidation allows publishers to focus more on maximizing the value of their supply.
Another core product is PrecisionX, an analytics tool that forecasts ad exposures based on digital IDs. Which means we can predict who is going to see an ad in a particular content title. Before patenting the tech, we ran a proof of concept with NBCUniversal with our model, which connects a set of digital IDs to a historical viewing data set and predicts where an advertiser will run over a specified time frame. Meaning, it predicts the probability that someone represented by an individual ID is going to be exposed to an ad in a particular show at a given time.
When we ran the proof of concept, we saw north of 80% accuracy on predicting spot-level exposure. Which is a promising number, considering all the industry concern about the data inaccuracy behind many TV and video media plans.
How exactly does datafuelX plan to expand further into the digital realm?
Programmatic expansion is one piece of the puzzle. We’re hoping to announce two major DSP partnerships this year. These conversations are underway, and they started at CES.
A DSP user would be able to drag and drop their linear plan into the DSP and immediately use that plan as part of their targeting criteria for digital-based buys to build incremental reach. A comprehensive view of targeting across different channels would help buyers ensure, when they pursue reach as a campaign goal, that they’re actually reaching net-new viewers instead of double-hitting the same people.
Closer integrations with DSPs should also give buyers more options on how to plan and target their campaigns. For example, a buyer would be able to activate a campaign based on their choice of audience demos or impressions across platforms, then buy that campaign via programmatic guaranteed or private marketplace deals. This degree of control in a biddable environment also means higher yield for publishers.
Will this expansion into programmatic change how datafuelX positions itself in the ad ecosystem?
To date, we’ve 100% just been just focused on publishers. That will remain a big part of what we do. But there are also ways that our technology can benefit the buy side, too – particularly by making media decisioning more sophisticated via integrations with DSPs.
You mentioned the company’s intention to work more closely with DSPs – but what about SSPs?
We’re absolutely open to working with SSPs. But supply-path optimization is a priority. The problem is that most publishers use multiple SSPs, and many SSPs specialize in a particular area of the ecosystem.
Have thoughts or tips? Hit me up at alyssa@adexchanger.com.
For more articles featuring Dan Aversano, click here.
