Jules Minvielle (Olyzon): "We now steer CTV media planning based on how programmes resonate with a brand's DNA"

Open Garden. You moved to New York in early September to accelerate a US expansion that started the same year with the hire of Zach Sorscher. What is your first read on the American market?

Jules Minvielle (Olyzon). Extremely positive. Being on the ground changes everything: physical proximity with the teams, lunches, meet-ups. It gives you access to events you would never fly in from France for, but which are essential to widen your prospect network.

As in France, getting into the big holding companies is still a long process, but we lean on our case studies with the French entities to open doors. In 2025 we already ran around fifteen campaigns in the US, with WPP Media and advertisers such as Doordash, EY and Dell.

What strikes you is the change of scale. Particularly on the independent agency side, which combines the fast decision-making typical of that profile with striking firepower: briefs for campaigns worth several hundred thousand dollars come quickly. Amounts you see far less often among independents in France.

As a result, our first six months in the US have been better than our start in France, which was already excellent. US revenue already accounts for 20% of our 2025 turnover and we expect to reach parity with the French market during 2026.

Did you use the move to shift Olyzon's positioning?

Our mission has not changed: help advertisers navigate the transition to CTV. We work particularly with those coming from linear who are trying to recover the contextual transparency they lost in that environment.

According to an eMarketer report published with Olyzon, knowing exactly where your advertising runs is a marketer's number one priority in 2026. As is being able to optimise a CTV plan accordingly, favouring the most suitable delivery contexts.

We have invested heavily on the technology side to give marketers the answers they are looking for. Our platform can now analyse the signals around each CTV programme before it even airs. Whether that is information shared upfront by the broadcaster or the conversations the programme generates on social media, on Reddit and elsewhere.

That lets us score each programme against the DNA of the brand in question. We call it the resonance score, a concept inspired by Nielsen, which established the formula: reach x resonance = reaction.

In the US people talk a lot about outcomes. An outcome is not just a click, it is the overall impact generated by a perfect alignment between the creative and the viewer's state of mind. An impact that moves upper-funnel KPIs (awareness, consideration) or lower-funnel ones (sales, downloads, site visits).

That lower-funnel space is dominated by players like Vibe, Mntn or TV Scientific, which target US SMEs. Do you want to go there too?

I am convinced that measuring CTV campaign outcomes in an ultra-short attribution window, looking at whether someone bought in the minutes or hours that followed, gives a biased and partial read on performance.

First because it mechanically pushes you to "top up" with other devices to capture an immediate action signal. Second because TV's impact is by nature more diffuse and longer.

The goal is not necessarily that someone buys straight away, but that TV takes its place in the messy middle and moves the decision forward. Obviously you do not sell a saloon car the way you sell an app to install on a phone.

So we do not settle for measuring a click or an install. We look at other consideration and intent signals to judge a campaign's effectiveness. Which is why we focus on advertisers coming from linear TV, who are structurally better suited to that logic: longer decision cycles, more complex trade-offs, but far better business quality behind it. Notably much lower churn than on the SME side.

Could you go after the linear TV market?

That is the plan. We will extend our expertise to linear TV this year. The textual and contextual signals we qualify in CTV transfer perfectly to linear. It is the logical next step for Olyzon.

In the meantime you are also moving on agentic. You recently launched an agent-to-agent activation via AdCP with Swivel. What did you learn?

Everyone talks about agentic AI and the term is increasingly overused. Or at least no longer differentiating, since it is part of operational reality.

We worked with Swivel, a platform founded by one of the founders of SpringServe, the ad server Magnite acquired in 2021. Swivel helps publishers build seller AI agents. Those seller agents are what we wanted our buyer agents to talk to, in order to remove operational friction.

Running formats like Home Screen or Pause Ad in CTV is a logistical nightmare today, whether you go through IOs or programmatic, since depending on the inventory type you often have to go through several DSPs. The agentic approach smooths those exchanges, whether through AdCP as here, or through APIs.

Each mode has its advantages. AdCP gives the agent more freedom to chase optimisation but is barely used to date. An API connection constrains the agent somewhat but lets you deploy right now...

Exactly. And we have no dogma on this. We obviously know agentic will let us grow revenue and cut operational costs. But we have no certainty on whether that happens through AdCP or not, precisely because of the adoption issue you mention. Either way, it is a chance to show what our tech can do.

We talked earlier about creative, which is how you entered the CTV market. Today everyone offers Pause Ads, Home Screen or L-Shaped formats. Is there still room for innovation there?

We are not here to push formats at any cost, and the L-Shaped example, which has a real scalability problem today, shows it.

There are, on the other hand, formats the market has adopted at impressive speed, like Pause Ads or Home Screen.

Here again Olyzon's approach stays the same: improve the advertising experience by making each spot more informative and interactive, while avoiding the "tunnel of ads" running one after another with no distinction. Always building on sentiments and moods extracted from an advertising context.

We have launched two major concepts. First the enriched ad break, an ad format that builds on classic video spots (pre-roll, mid-roll) to make them more engaging and effective. An L-band is added alongside the video ad, displaying the brand logo and the promise from the first second to maximise recall. Interactive elements can be added, such as a QR code redirecting to a landing page or a quiz that boosts engagement.

Then there are transition cards, an ultra-short video format (two to five seconds) designed to turn the entry into the ad break into a fluid, contextualised and engaging experience. Running just before the advertiser's video spot, they adapt to the programme being watched to create a natural transition between editorial content and advertising. The result: stronger attention, better ad acceptance and higher-quality exposure for the brand message. So to answer your question, there is still room for innovation on formats.