Sören Stamer, auteur chez State of Digital Publishing

Publishers have been told for years that personalization is the answer to declining engagement and the fight for loyal relationships, yet 84% admit they're still running generic campaigns. According to Forrester's 2025 CX Index, 21% of organizations globally saw their customer experience scores decline last year, while only 6% improved. The industry has spent a decade investing in personalization technology. The results suggest something more fundamental is broken. Behind the scenes, the workflow is the problem. The industry has been solving the wrong problem When personalization fails, the instinct is to add more technology. Better AI, richer data, and more sophisticated segmentation. But in conversations with the teams using it, I see the same pattern repeatedly: the tools aren't the problem. The organizational design around them is. The people who buy personalization platforms are rarely the people who create content. The people who set up audience segments are rarely the people who understand why a piece of content was written the way it was. And the editors who actually publish content every day rarely have visibility into any of it. It's the result of how technology stacks were built: tool by tool, team by team, each layer added to solve a specific problem without anyone stepping back to ask whether the overall system was actually delivering relevant experiences. Increasingly, it isn't. Fragmented tech stacks became a liability, and the problem is compounding What's less discussed is why the fragmentation problem gets worse rather than better, even as organizations continue to invest in personalization technology. Each new tool added to the stack creates a new handoff point The marketing team buys a personalization platform. A separate team configures the segments and triggers. Editors keep working in the CMS as if none of that exists. What starts as a coordination problem becomes an organizational one: different teams, different priorities, different definitions of what "personalized" means. Over time, personalization only gets done for high-priority campaigns, because the cross-team coordination required makes it too expensive to do for anything else. Everything else goes out flat, to everyone. The speed problem compounds this further Visitors make decisions in seconds, expecting the kind of contextual, relevant experiences they get from social media and AI tools every day. Yet most campaigns are still built on rigid schedules rather than driven by real-time engagement. A visitor who has viewed a page three times needs a different experience than someone arriving for the first time. Someone hesitating at checkout needs reassurance, not a generic banner. By the time a campaign is optimized and the right message is ready, the moment has passed. Then comes AI The assumption is that AI will fix the personalization gap by automating what humans can't do at scale. But AI doesn't fix fragmented data, it amplifies it. When content is scattered across multiple systems with no single source of truth, it goes out of date. Visitors get different messages depending on which channel they came from. Data signals are incomplete or stale. When AI orchestrates experiences on top of all this, inaccurate outputs go out faster and wider than any editor could catch. The question the industry isn't asking The right question isn't "how do we get better data?" or "how do we make AI smarter?" It's: why are the people creating content disconnected from the systems that determine how that content reaches your audience? Content governance, knowing what content exists, where it lives, who owns it, and whether it's current, is what makes personalization at scale possible. Not as a compliance function, but as the operational foundation that everything else depends on. Without it, every investment in personalization technology produces diminishing returns, because the inputs are unreliable and the people closest to the reader have no way to act on what they know. The structural fix isn't another tool. It's connecting content, data, and AI in the same workflow so that editors, the people who understand the audience, the tone, and the content, can make targeting decisions right at the point of publishing. A practical example: Personalization for a German publisher One of Germany's largest publishing groups with over 40 imprints, faces a version of this problem at publishing scale: 2,500 new titles a year and a catalog that has to stay relevant across genres, reader demographics, and constantly shifting market trends. Their vision was clear from the start: a scalable, modular content platform that could support real-time publishing, user-centric personalization, and deep integration with business systems like Salesforce CRM. Because in publishing, timing is everything, the foundation needed to be one where content, data, and audience behavior work together from the start. Running on the CoreMedia Experience Platform, a CMS with native personalization capabilities, bestseller status, pricing, and category structures update in real time. User interactions, follows, bookmarks, and newsletter sign-ups feed directly into the CRM for audience segmentation, so notifications and campaigns are relevant and timely rather than generic. Editorial teams can enrich content with author bios, events, and awards directly on the content item, without having to rebuild from scratch or switch tools. Personalization became practicable when it stopped being a separate workstream and became part of how content actually gets created and published. Personalization should be as simple as publishing The conversation is often about what personalization tools can do, what AI can predict, and what data can reveal. That matters. But in practice, the most sophisticated tools stay theoretical if they sit separate from the editors who need them most. The data lives in one system, the content in another, and the people publishing to readers every day have no way to connect the two. That distinction is still not well understood, and it's costing the industry more than it realizes. The publishers closing the personalization gap didn't get there by adding more tools. They got there by treating personalization as a workflow problem, not a technology problem.

SODP

SODP Media

Sören Stamer

La véritable raison de l'échec de la personnalisation : ce n'était jamais le rôle du rédacteur en chef

À propos

Liens utiles

Articles en vedette