00:07Hello, I'm Paolo Moulet, I'm SVP Global Clients and Strategic Partnerships at Dailymotion.
00:14Before that, I was Chief Revenue Officer at Mobcoy, and before Mobcoy I was at Teads for 14 years.
00:20I joined Teads before it was actually called Teads, so it's been a great story, incredibly successful.
00:26I actually had a different career because I studied philosophy for 11 years, practicing, studying philosophy.
00:31My main mission is to allow the emotion to get closer to brands and advertisers,
00:37increase the depth of the relationship that we have directly with advertisers and their agencies, of course,
00:43and develop partnerships with them.
00:45The biggest gap I leave every day in my media life is the gap between media KPIs and business outcomes
00:54as a measure of success of a campaign.
00:57I think there's been an industry failure, like an industry-wide failure, over the past 10-15 years,
01:04where we have been almost forced to focus on proxy KPIs,
01:10which at some point were the best thing that we had to assess or predict the results of a campaign,
01:15but they turned out to sometimes be too short-term or not the best proxy possible to assess the success
01:22from a business outcomes point of view of a campaign.
01:25And that has created a fracture, not just between vendors and agencies or agencies and brands,
01:31but within brands, between media directors and CMOs.
01:35Because what looked good on an Excel as a proxy KPI, whether it was VTR, CTR for a campaign,
01:42wasn't necessarily aligning with the business outcome.
01:45So that's been a kind of an industry challenge that a lot of people have been trying to address and
01:49fix over the years.
01:50At Dailymotion, we try to solve that gap by utilizing AI.
01:55Dailymotion is a video platform network.
01:57So we have dailymotion.com, which is a huge website where 400 million users every month consume and upload video.
02:04Video. Dailymotion is also the tech stack that powers the video streaming capabilities
02:09and pre-roll capabilities of some of the biggest publishers in the world.
02:13We have 5,000 premium publishers that use our video tech to power their video,
02:18where we can obviously run advertising across.
02:21But the point is, every month we capture 5 billion video streams and all the data that comes with it,
02:28whether it's behavioral, interaction data, video encoding, metadata, video transcripts, frame-by-frame analysis,
02:38the content and the context in which the video is displayed.
02:42We've been using machine learning for 10 years to try and make sense of this huge amount of data
02:48and optimize the campaigns based on that.
02:51Machine learning is a bit dumb.
02:52It's really good at pursuing one KPI and take no prisoner.
02:56So if you train an algorithm to say, bring me the best possible CTR,
03:02it will pursue CTR at no matter what cost.
03:05And it gets a bit greedy.
03:06And maybe it doesn't deliver the right objectives for a campaign from a different point of view.
03:11AI allows us to make sense of this huge amount of data that we collect every month
03:15and turn it into signals that we can then use and leverage through an agentic platform
03:21to allow brands to better plan, better define their audiences,
03:26better reach their audiences with the right creative,
03:29analyze through AI for attention, for cognitive demand and optimize accordingly,
03:37activate through the platform and measure through the platform.
03:41So bringing in, for example, brand outcomes, brand lift results,
03:46analyze the performance in the campaign,
03:48and more importantly, bring them back into the feedback loop
03:52to then retrain and continue improving the models that AI is using.
Commentaires