This week, Shirley Marschall looks at why a message that can feel well-placed in one environment may feel intrusive in another. Like hair, of course…

Picture beautiful hair. Long, healthy, glossy hair. It's striking, isn’t it?

Now picture a single hair on… a hotel pillow, in the sink, in your soup. 

Identical hair. Completely different reaction. The disgust is immediate and reflexive, and no amount of rational thinking makes the ick go away. The hair didn't change: the context did.

This, in a single (disturbing) image, is the story of digital advertising over the past two decades and why the industry re-discovered contextual advertising over and over again. This same message that can feel premium, relevant, and well-placed in one environment may feel intrusive, jarring, or vaguely suspect in another.

But what actually happened to contextual, and where does it sit today? Still thriving with specialist vendors? Absorbed into DSP and SSP feature sets? Quietly inside curation? Still strategically important, just no longer marketable as its own category? Or has the market simply moved on? Its latest comeback came when the cookie crumbled, positioned as a smarter, more sustainable corrective to identity-heavy targeting, though that already feels like a Google U-turn and a(i)ges ago…

Going back in time, advertising was built on contextual logic long before programmatic made audience precision the industry’s dominant obsession. Entire media strategies were once shaped around a fairly intuitive principle: place the right message in the right environment and the message lands differently.

Haircare ads sat beside beauty editorial. Financial services aligned with business journalism. Automotive campaigns appeared in enthusiast publications. The environment was part of the message. Then, somewhere along the way, digital advertising became increasingly focused on who was seeing an ad rather than where they were seeing it. At peak programmatic confidence, buying against environments could suddenly feel almost quaint. Ever pitched "right message, right time, right person"… and felt, slightly more sophisticated than the non-programmatic people in the room, asking: why target a publication when you could target the exact person reading it?

Anyway, precision is the metric of sophistication and identity the strategic centre of gravity. Context never completely disappeared, but it often got demoted to supporting status at best, while the industry chased ever finer-grained audience signals. 

So again, what happened? Was it ad tech’s habit of reducing enduring capabilities to temporary talking points, revisiting them only when market pressure demands a new answer to an old problem? Partly, for sure, because market attention is elsewhere right now. AI dominates the discourse, retail media continues its expansion and curation has become the supply-side buzzword of choice. Or is it the simplest explanation, that contextual stopped being novel? Something changed, that’s for sure, even if it’s just the wording, to fulfil today’s need to sound relevant. After all, nowadays everything has to be a signal, nuanced, AI-infused. 

Some of it still lives with pure-play specialists. Companies that built their business around contextual intelligence and continue refining, rebranding, reframing, and AI-enhancing their offer. They are still there, still innovating, but no longer carrying the same industry-wide narrative weight they briefly held during the post-cookie panic.

Some of it has been absorbed by DSPs and SSPs, where contextual features become part of the underlying machinery, inside broader optimisation and targeting frameworks, rather than a standalone strategy. 

Some of it has been folded into curation, just not necessarily marketed as such. Much of today’s curation conversation is about quality, efficiency, and packaging. Contextual is there but plays a supporting role. One among many others.

And some of it may simply have dissolved into newer environments altogether. Take GenAI chats and ads. When advertising becomes a natural (or unavoidable) part of AI-driven interactions, relevance will depend mostly on the immediate context of the exchange itself, which in many ways makes it contextual advertising in one of its purest forms. Different surface, same contextual principle, even if the context itself becomes something entirely new.

Andrea Tortella, co-founder and CEO of Thrad, captures the shift neatly: "The conversation about contextual got stuck in the 2010s, debating keyword targeting and page-level signals. Meanwhile the surface itself moved with billions of people now spending real time inside AI chats, and every message they send is pure context." His company serves ads inside AI conversations, where, as he puts it, targeting is almost trivially simple because the user has already told you what's on their mind.

A signal of where contextual logic is ultimately headed may go far beyond ad placement altogether, moving upstream into the system that produces and serves advertising.

On one side, platforms are beginning to reshape the environment itself. Recent Google patent filings describe systems that dynamically adjust landing pages around immediate user intent and situational context. Not just matching an ad to a page, but reshaping the page itself around the person in the moment they’re reading it. 

On another, the creative layer is becoming fluid. Meta’s AI-driven advertising tools increasingly generate and modify ad variations in real time, adjusting messaging, format, and presentation based on inferred context. The ad is no longer a fixed object placed into a context, but something that can be reconstructed by it.

That’s, structurally, contextual logic. Just on a far more interventionist level. 

So maybe that’s what keeps happening to contextual: not disappearance, but absorption into whatever environment comes next. Even contextual depends on the context…

Read all of Shirley's columns here, and find her on LinkedIn