Cannes always sets the agenda for the year ahead. This year, we are leaving with two announcements that take Vidium well beyond where we started. The first opens up managed access to the most sought-after video inventory on the planet. The second moves our AOS optimisation from the report into the live bidstream. Two different problems, one organising idea: quality should decide where video budgets go.

Part One: Vidium Direct, premium video access made simple

Premium video is where audiences are, and it is also where buying is hardest. Streaming, premium publisher OLV and the big platforms each sit in their own silo, with their own targeting logic and their own way of reporting success. Reach is fragmented, targeting is inconsistent, and measurement rarely travels across screens.

‍

Vidium Direct is our answer and our first managed-service offering. We layer our technology on top of the leading DSPs to give you a single route into Netflix, Broadcaster Apps, Disney, Paramount, Spotify, YouTube, Reddit and premium publisher OLV. One brief, one team, one set of outcomes, across CTV and online video.

‍

Underneath that simplicity sits an agentic orchestration layer, built across the leading DSPs and platform ad managers. It runs the entire campaign as a single system. Activation happens from one point rather than platform by platform. Optimisation runs continuously across the entire portfolio, moving budget to what works in concert rather than in isolation. And insights come back unified, so you read one clear story rather than stitching together a dozen dashboards. The fragmentation stays under the hood; the experience stays seamless.

‍

What makes it different is how we target. Most premium video still leans on blunt demographics or tired segments. We have built a unified targeting layer from Trends data: what people are actually watching, drawn from the world's leading video platforms, combined with what people are actually saying, drawn from the web's most active conversation communities. Pairing viewing behaviour with live conversation lets us follow genuine cultural attention and spend on it consistently, no matter which platform the impression runs on.

‍

And every impression is measured against our Attention Opportunity Score (AOS), the attention dataset we built specifically for OLV and CTV. So you are not just reaching premium environments; you are reaching the ones that actually earn attention, with proof that travels across the whole buy.

Part Two: Quality scoring in the bidstream

On the programmatic side, attention has always had a timing problem. It usually arrives after the auction, in a report, once the money is already spent. Useful for the next campaign, too late for this one.

‍

We have changed that. Working inside the cloud environment of one of the world's leading exchanges, we will run AOS as a containerised service co-located with the auction itself. Opportunities are scored in real time, inside the bidstream, within the milliseconds an auction allows.

‍

The result is a step change in control. Instead of optimising at the deal or domain level, you can now bid and optimise on quality and attention at the individual impression, as it happens, with insight flowing back in real time. Higher-attention impressions get the budget they deserve; the rest stop draining it. Same supply, sharper decisions, measured at the point they are made.

One idea, two fronts

These launches point in the same direction. Vidium Direct brings attention-led buying to the premium managed market, across the platforms everyone wants and few can unify. The exchange cloud integration brings the same discipline into programmatic, at the impression level and in real time.

‍

Premium reach, culturally relevant targeting, and attention you can act on rather than just admire after the fact. That is the standard we are building toward, and both launches will be live during Q3/H2. 

‍

Want to see either in action? Get in touch at vidium.co.

‍

It’s widely reported that anywhere from 20% to 30% of programmatic video spend is lost due to supply-side misrepresentation, invalid traffic and low-quality inventory. That’s billions in wasted ad spend – not to mention billions in ad revenue siphoned away from premium publishers – further eroding high-quality video inventory.

If programmatic video is going to mature beyond the “scale at any cost” phase, we need to rethink how we activate and optimize toward higher-quality inventory. Self-declared inventory (accurate or not) doesn’t fully capture multidimensional video inventory, and static metadata snapshots don’t verify both the ad and content aspects of the video experience.

We need to start looking at the deterministic signals that tell us how video ads and the surrounding content behaves on the page to gain a much more relevant, fine-grained understanding of the actual viewing experience.

1 - How the video experience gap grew so wide

It’s easy to blame opportunistic publishers masking low-quality inventory for our current status quo, but we arrived here in a much more organic way.

Over the past five years, programmatic fundamentally democratized access to video. But to scale quickly, the industry developed labels as a proxy for ad experience. “Instream,” for example, implied a pre-, mid- or post-roll ad within real editorial video content. “CTV” implied a premium, lean-back viewing experience on a television screen. And premium OLV environments were expected to mirror that same immersive, video-first experience within web and mobile environments.

But as programmatic video supply has expanded and supply-side competition has intensified, that connection between the label and the lived ad experience is now far more ambiguous.

Inventory categorized as instream isn’t always anchored to meaningful editorial video content. In fact, instream-labeled inventory often sits on sites that don’t even have a valid video player to serve the format.

And in CTV, “premium” does not always deliver the immersive value advertisers assume they’re buying. Case in point, a large volume of premium advertisements are delivered on web or mobile browsers when most buyers expect their CTV ads to run on full-screen TVs.

That 20% to 30% inefficiency lives in the gap between label and experience.

2 - What closing the gap actually requires

So why does the inefficiency gap persist? Most ad technologies were built for static digital ads and later adapted to video. But video is an experience that unfolds over time. Quick snapshots of a page attain surface-level metadata. They miss key facets of how video actually behaves on the page.

And so we need to ask these important questions:

· First and foremost, is there a video player on the page?
· If yes, what type of video player is it – and is it one via which we can buy/transact?
· Does the player autoplay or require a click?
· Does it initiate muted or with sound?
· Does it stay fixed on the page or collapse into a small floating unit that follows the user?
· Is it embedded in meaningful editorial content or dropped into an outstream container between paragraphs?
· Does it expand, reposition or interrupt the reading experience?

These nuances make all the difference for whether an impression feels premium and immersive – or intrusive and ignorable. But parsing those data points requires looking beyond what the code says to observe how the player behaves in real conditions.

That means analyzing video placements at the URL level, staying on the page long enough to see how the player functions and capturing execution-level signals that the quick crawls of traditional ad verifiers typically miss. It also means evaluating not just whether a player exists, but the size, placement integrity, movement behavior and the ratio of ads to content surrounding it.

Looking at video at the placement level rather than broadly categorizing an entire domain gives brands a more precise picture. Two placements on the same site can deliver dramatically different experiences.

Once you see the full picture, you can act on it. Spend can be directed toward truly premium, high-impact placements. Intrusive sticky units can be filtered out. In-article video that commands attention can be distinguished from low-attention outstream ads. Private marketplace deals can be structured around player behaviors instead of broad site lists.

3 - Smarter spending strengthens the entire ecosystem

The advertiser benefit of these deeper insights is obvious. If even a fraction of that 20% to 30% inefficiency is reclaimed, performance and ROI lifts will be sizeable. But the ripple effects benefit the entire ad ecosystem.

When loosely categorized or lower-quality supply competes in the same marketplace as truly high-quality inventory at a lower price point, it penalizes publishers that are investing in those more premium video experiences. By contrast, better insights drive more ad dollars toward publishers that have invested in premium experiences, such as real video infrastructure, editorial content and thoughtful ad deployment. This strengthens the economic foundation for high-quality content creation, supporting the exact environments that brands say they want to fund.

Even consumers win in this shift, because they experience fewer cluttered deployments, fewer intrusive ad placements and more relevant ad content delivered in environments that feel intentional rather than opportunistic – not to mention providing economic support for the high-quality content that we all want.

4 - Why this shift will happen faster than people expect

The best part of this shift is that it doesn’t require a heavy infrastructure rebuild. The high-quality video environments brands want already exist – and there’s more than enough scale within placements that deliver strong player integrity, thoughtful deployment and immersive viewing contexts. It’s just hard to find them in today’s cluttered supply landscape.

Capturing and applying these execution-level signals will drive immediate impact, redirecting spend and producing near-term gains. As those gains become visible, momentum will build quickly across the industry. More brands will embrace this refined approach, more ad dollars will flow to high-quality environments and premium publishers will have the economic support to expand their high-quality video inventory.

5 - The future of programmatic video is about precision

This is a classic story of market maturity.

Programmatic video was built to scale as fast as possible and it succeeded, exploding across the open web and CTV in less than a decade.

But scale breeds complexity, and inclusion lists of domains and apps now stretch into the thousands, making it incredibly complex to verify how video is actually deployed. Moreover, when we focus too much on CPM efficiency to drive decision-making, qualitative differences blur. And when labels are largely self-declared, inconsistencies persist.

That’s why every fast-growing market eventually transitions from expansion to refinement. Programmatic video is at that inflection point.

The inefficiencies we see represent the tremendous opportunity in front of us all right now. The next decade will be about assembling supply more intelligently, moving from one-dimensional labels and surface-level assumptions to looking deeper and having a behavioral understanding of execution.