00:00You heard me mention the volume of creative that we see today, roughly 50,000 creatives that we
00:05then use AI to look at what's happening. I'll give a real quick example. If a kid's not wearing a
00:09helmet in a commercial, we're not running it, right? So you can train the model to be able to
00:14do that. The other thing is we were talking earlier about, so there's a creative itself,
00:19and then who are you going to run it against? And I love Dominic's background, and he's also
00:23ran the data practice, right? So he has a data background, he has a finance background,
00:27and he has a creative background. And on the data background, to figure out who you're looking to
00:32target, I'll give an example where you take an advertiser's first-party data set, you combine it
00:37with Disney Advertising's first-party data set, you do that in a privacy-compliant way in a cleanroom,
00:42you identify the most valuable people to target, and then you go ask the model to go find more
00:48people like that. And then you A-B test that against the advertiser's data, which they can run
00:53by themselves, and what the AI and ML models came back with. They find more people, they find
00:59performance that's totally in line with the first-party audience segments, and that was a
01:04new concept probably two years ago or so. That's where many of the conversations that we're having
01:09right now is how do you take the data models, apply those with lots more creative to different
01:14audiences, and then you go test those things, and then you need to be able to move really quickly
01:19and respond. I love the example of going right to Slack and being able to respond to something in
01:23real time. The speed at which modern marketing works at takes all those skill sets, right? And so
01:29it's great to see when those things are all kind of coming together.
Comments