As AOL integrates with its new parent Verizon and even newer sibling Yahoo to form one of the largest consumer data powerhouses around, classically trained data scientists are working together with specialists in less obvious fields including genetics, neuroscience and electrical engineering.
Verizon employees with traditional STEM backgrounds have been joining biweekly meetings of AOL's Science Guild, so named by AOL CEO Tim Armstrong, to advance capabilities like ad viewability optimization. The Science Guild takes a grad school-esque approach to guiding independent research toward common company goals such as developing algorithms, building products and determining where to source data.
To ensure AOL makes the most of Verizon's data and capabilities for advertising-related services, the company has also recruited from outside traditional data science practice areas.
The company would not share names of new hires, leery of poachers from rival firms, but described recent recruits including a PhD in computational neuroscience who developed statistical models and machine learning techniques to analyze brain function at MIT; a PhD in statistics who built gene selection and expression mapping models while conducting genetic and molecular biology; and a PhD in feedback control theory who studied control system design for things like hard disk drives.
One recent hire has an electrical engineering background, and some have histories studying auction theory, helping them measure how ad prices change throughout the day, said Rob Luenberger, chief scientist and senior VP of research and development at AOL Platforms. It's no longer necessary to find people who can write certain algorithms that have become standardized, he noted. Instead, AOL seeks people who can "pull out signal from noise," he said.
"How do you formulate the whole problem and realize what are the parts?" he asked rhetorically. "How is this going to react with the other parts of the system that we have?"
On ad viewability, AOL is developing tech that goes beyond measuring whether ads appear on-screen long enough for consumers to potentially notice. It uses predictive modeling to gauge the chance that an ad impression will be viewable, letting marketers alter their bidding strategy based on an ad's likelihood of being seen.
"This is the live application of data science," said Seth Demsey, chief technology officer of AOL Platforms.
AOL employees and their new colleagues have also developed an optimization feature that guesses consumer demographics using models based on "seed populations" of Verizon users who have opted in to share anonymized and aggregated data on their ages and genders.
The tool helps predict whether someone is likely to click on a certain ad or buy a product.
"This is an opportunity for us each to do what we do best," Mr. Demsey said.
Verizon and AOL have yet to say much about how the Yahoo acquisition will be integrated, but Verizon's roll-up of AOL and mobile ad firm Millennial Media, acquired by AOL in 2015, hints at the future. AOL already allows advertisers to target based on Verizon's carrier data, for example, and began making its Precision Market Insights data available to select clients for refining ad campaign optimization and measurement earlier this year.
Precision Market Insights uses location data generated by Verizon Wireless-enabled devices to determine demographic and other information about people in certain locations and connect it to other places those people visited. The result is anonymized data showing, for instance, which percentage of an audience segment that attended a basketball game then visited a sponsor's burger joint at a particular time on a given day afterwards.
~ ~ ~
CORRECTION: An earlier version of this article erroneously said Yahoo employees were joining meetings of AOL's Science Guild. That won't happen until Verizon's acquisition of Yahoo officially closes.