Every time Melinda Han Williams, chief data scientist at Dstillery, hears someone in the ad industry talk about a “data clean room,” she has to chuckle.
Before joining Dstillery in 2013, Williams was a physicist studying “electronic transport in nanostructured graphene devices.” (Tune in to find out what that is.) This involved donning a one-piece head-to-toe cleanroom suit and entering a sterile environment to work with one-atom-thick honeycomb sheets of carbon atoms.
“I actually spent a lot of my time in a literal, physical clean room,” Williams says on this week’s episode of AdExchanger Talks. “I’d put on the Tyvek bunny suit, like in the old Intel commercials, and go into this special lab where the air is super clean and filtered and no dust can get in.”
These days, Williams spends her time focused on using AI to develop cookieless solutions, including a behavioral targeting solution called ID-free that doesn’t rely on identifiers in order to reach people.
“Instead of focusing all of our AI predictive power on understanding the person we’re trying to reach, we shift the focus of our AI to understand the moment when we’re trying to reach that person with an ad,” she says. “We use a neural network to build a 128-dimension map of digital behavior by looking at opt-in digital journey data.”
But despite the product’s seemingly self-explanatory name – ID-free – a surprising number of people have trouble wrapping their heads around the fact that there really isn’t any identifier involved.
“We tell them about ID-free,” Williams says, “and they’re like, ‘OK, so it’s a new identifier – how does your new identifier work?”
Also in this episode: Why Williams made the move from physics to cookieless, using AI to reduce bias in advertising and getting more women into executive roles in ad tech.
