Do We Want to Turn the Open Web into a Walled Garden?

A response to the first of the OWM's seven demands for strengthening the open web, and to the article "Why an ID-based advertising infrastructure would be so important" by Juliane Paperlein

The OWM has published seven demands for strengthening the open web (https://www.owm.de/presse/pressemitteilungen/pm-einzelansicht/owm-stellt-sieben-forderungen-zur-staerkung-des-open-web) – a valuable initiative that we at Welect support. Since our founding, we have contributed substantially to points 2 (Cookieless Future), 3 (Strengthening independent publishers), and 7 (Consumers at the center).

The first demand, in my view, warrants a more differentiated look – particularly because it was highlighted in Juliane Paperlein's article (https://www.horizont.net/agenturen/kommentare/id-basierte-infrastruktur-gemeinsam-einsam-230176). The demand is that the open web must reinvent itself through an ID-based infrastructure in order to keep pace with platforms from the US and China.

Do we want to transform the diversity of the open web into one giant walled garden by layering an ID infrastructure on top of it? Should we as an industry – advertisers and media alike – and as a society, after two decades of power concentration by tech giants, introduce a system that connects people across all media offerings via ID – thereby creating the risk of a new centralized power structure?

Does the idea of a shared technical foundation still sound appealing? Then consider this: doesn't it already exist? One word: Google.

Hardly a publisher in the open web operates without Google Ad Manager, Google's SSP or AdX exchange, or without relying on DV360 as a demand source. The antitrust proceedings against Google in Virginia in 2024 revealed the measures by which Google systematically expanded its market penetration and consolidated its dominant position over many years: Project Bernanke, Dynamic Allocation favoring AdX, Unified Pricing against differentiated floors, complete visibility into supply and demand data – to the detriment not only of publishers, but also of advertisers whose campaigns were met with intransparent auctions and artificially influenced prices. In Germany, the revelations from the trial received remarkably little attention. Would a more thorough discussion in Germany have brought more nuance to the current proposals about a "missing shared infrastructure"?

What has prevented the willingness to build a shared, Google-independent infrastructure? Despite – and perhaps because of – its own market power, even Google shut down its Privacy Sandbox project in mid-October due to low market acceptance, conflicting ecosystem expectations, and increasing regulatory pressure.

For publishers in the open web, an alternative explanation may apply: programmatic operates in a buyer's market with excess supply. Dynamic auctions have predictably led to a race to the bottom for many years: falling CPMs, comparability and transparency in favor of the demand side have eroded media brands through the habituation effect of "cheap prices" – and have not contributed to (3.) strengthening independent publishers.

The fact that the major platforms shielded their own inventory from transparency while publishers were fully exposed has widened the imbalance and undermined (4.) fair competition with the walled gardens.

It is also notable that the three major social media platforms – which in some markets already capture more than half of advertising budgets – offer no shared ID infrastructure. User identification has consistently worked in one direction: from the open web into the walled gardens, rarely the other way around.

This is not about questioning the programmatic ecosystem. On the contrary: I follow with great interest how the AdTech community abroad is rethinking supply paths and currently exploring a new AdCP protocol. For publishers, an ID-based infrastructure could, in my view, worsen the existing pricing market dynamics – particularly vis-à-vis the platforms – because unresolved structural problems persist and new challenges are already being felt. Which ones?

For 20 years, the vision of personalized media consumption and online targeting has been sold. But has personalized 1:1 targeting actually made the internet better? The unintended effects that come to mind include: filter bubbles (personalized algorithms preferentially show people content and advertising that matches their past behavior, causing users to repeatedly encounter similar topics, opinions, and brands), societal polarization (when people only see content that confirms their views, echo chambers form), ad fraud running into the billions (ID-based targeting is data-hungry and vulnerable to fake IDs, bots, and fake impressions), and regularly sobering targeting quality (ID-based targeting promises precision, but typically operates on outdated, modeled, and hypothetical data – while simultaneously excluding roughly 50% of users for whom no ID or other signals are available).

Until we solve these structural problems as an industry, building a new ID infrastructure is pure symptom treatment. I fear the same applies to the deployment of new AI black boxes.

Advertisers and publishers in the open web face two major challenges today, in addition to those already mentioned:

• A decline in reach due to the zero-click search effect (search queries answered without a click through to external content)
• A rapidly growing share of signal-loss traffic, which is splitting the open web into two spheres

The proposed unified infrastructure does not address the first of these two challenges and optimizes exclusively the portion of the open web that has identifiable online traffic. Online traffic without IDs and with limited data signals is by no means worthless – it is simply systematically ignored by many DSPs, particularly DV360. An ID infrastructure would not resolve consent rejections, Apple IDFA opt-outs, private browsing, or the use of ad blockers.

This is not an either/or: in the realm of authenticated traffic and the addressable open web, ID-based advertising solutions make sense – while in the signal-loss segment, other, alternative approaches are needed to effectively reach people.

Differentiation, not imitation

If we want to counter the platforms and their dominance with a global ID structure, we are playing on their turf – and that is a game that is very hard to win. The real strength of publishers and advertisers lies in differentiation: through content, formats, and user experiences.

Especially in the age of AI, where a growing generalization and self-referentiality of content is emerging under the term "slop" – because AI systems are increasingly feeding each other interchangeable content – the opportunity lies in carefully researched, original, and high-quality contributions and formats. When publishers succeed in recapturing the attention of their audiences, budgets will follow.

Advertising functioned for a century without ID infrastructures – because it was based on relevance and trust. That is something we could reconnect with today.

The major social media platforms became successful because they were founded on the principle of the intent economy: people choose for themselves what they read, whom they follow, what they like or share. Their monetization today, however, follows the logic of the attention economy – with feeds that are increasingly algorithmically controlled to maximize dwell time, something many users have long come to find intrusive. Is this an opportunity for the open web?

Today, AI produces content at high speed – but also with a tendency toward homogeneity. This opens an opportunity: media can score precisely where AI falls short – with quality, context, relevance, and emotion.

Conclusion

The open web does not need a walled garden.

The real task lies in regaining relevance and trust. The challenge for advertisers will be to reach people with advertising messages at scale – on channels that are not controlled by intransparent, externally steered systems with their own interests.

Anja Martensen and Daniel Neuhaus from The Trade Desk rightly address the issue of insufficient reach from ID-based solutions in their contribution from October 29, 2025, and propose an intelligent combination of deterministic and probabilistic approaches.

The signal-loss challenge can be addressed with a combination of a genuine intent-economy principle and the most deterministic signal of all – namely when people take over their own targeting in real time (7. Consumers at the center).

Advertising can be delivered on the basis of transparent, real-time signals from real people – acting closer to their actual interests and behavior, rather than relying on historical or modeled data. This not only increases relevance, but also the effectiveness of campaigns.

When people consume content voluntarily and with genuine interest, reach is created that enables brands to achieve real market penetration.