Latest / The CMO Podcast with Fexingo: Marketing Leadership, Budgets, and Executive Strategy / How CMOs Are Using Contextual Targeting After the Cookie
Transcript
- Lucas: So next time you read a news article about climate resilience and see an ad for a water filter or a weather app — that's not a coincidence, and it's not because the ad platform knows who you are. It's a bet on context. Luna: Right — context over identity. That's the big shift we're seeing as third-party cookies finally go extinct. Lucas: Exactly. And it's not just a privacy workaround. A growing number of CMOs are actually reallocating budget toward contextual targeting because they're seeing better engagement, sometimes better conversion, than behavioral retargeting ever gave them. Luna: I want to get into the numbers, but first — define what we mean by contextual today, because it's not just 'put an ad for hiking boots on a hiking blog.' Lucas: No, that's the old version — keyword matching. 'Shoes' next to 'shoes.' The new contextual is semantic. Platforms like GumGum, Seedtag, and even Amazon's own contextual engine are using computer vision and natural language processing to understand the emotional tone, the sentiment, even the visual composition of a page. Luna: So it's not just what the page is about — it's how it makes you feel. Lucas: Exactly. And that's where the CMO opportunity lives. Let me give you a concrete example from a few months ago. A major CPG company — think laundry detergent, household cleaning — wanted to test contextual against their standard demographic targeting. They worked with a contextual platform to identify articles with themes around 'resilience,' 'renewal,' 'fresh start.' Not cleaning, not laundry. Emotional concepts. Luna: Interesting. So they're not targeting people who just searched for stain remover. They're targeting a mindset. Lucas: Precisely. The campaign ran across news sites — articles about people rebuilding after a natural disaster, about community cleanups, about spring traditions. The results: a 40 percent lift in brand recall versus their standard digital campaign, and a 22 percent lift in purchase intent. And zero personal data collected. Luna: That's the kind of number that makes a CMO sit up. But I wonder — does it scale? Can you get that kind of lift for a car brand or a luxury good, or is it mostly a CPG play? Lucas: It's scaling fast. The New York Times, for example, launched its own contextual targeting product called — I think they call it 'Contextual Fusion.' It lets advertisers place ads against specific article topics, but also against reader engagement signals without tracking the reader. Another example: a luxury automotive brand ran a contextual campaign against articles about 'craftsmanship' and 'heritage' — think long-form features on watchmakers, architects — and saw time-on-site for their ad units triple. Luna: So the thesis is: if you put a premium ad next to premium content that the reader actually chose to spend time with, the attention transfers. That makes intuitive sense. Lucas: It does, but it also challenges a decade of programmatic dogma. For years, the mantra was 'reach the right user at the right time' — user being the key word. Contextual flips that to 'reach the right mindset in the right environment.' Luna: And that requires CMOs to think differently about their media mix. You can't just set a retargeting pixel and forget it. Lucas: Right. The CMOs who are succeeding with contextual are investing in creative that can flex to match different emotional contexts. One ad creative might not work for both a 'resilience' article and a 'craftsmanship' article. So they're building modular creative — same brand message, different emotional tonality. Luna: That sounds like more work upfront. But if the ROI is there, it's worth it. How are budgets shifting? Lucas: I've seen some forecasts suggesting that by the end of this year — 2026 — contextual will account for about 25 percent of programmatic display spend in the US, up from maybe 12 percent in 2023. And CMOs at companies like P&G and Unilever have publicly said they're moving away from targeting segments based on inferred demographics and toward 'attention-based' buying. Luna: Which ties back to the measurement challenge. If you're not tracking users, how do you prove the ad worked? Lucas: That's the billion-dollar question. The answer today is a mix of brand lift studies, panel-based measurement, and something called 'incrementality testing' — basically, you compare a market where the contextual campaign ran to a control market where it didn't. It's not as granular as a click-through rate, but it's arguably more honest. Luna: And less prone to fraud, since you're not chasing the same fake users around the web. Lucas: Exactly. The other thing I'd flag: contextual targeting is getting a boost from generative AI. Some platforms can now generate ad copy on the fly that matches the tone of the article. So if the article is serious, the ad language is serious. If it's humorous, the ad is playful. All without a human rewriting it each time. Luna: That's wild. But also a little scary — do we want AI deciding what tone is appropriate? Lucas: Fair question. The early results suggest it works, but brand safety is a real concern. You don't want your ad for family travel to appear next to an article about a plane crash, even if the AI thinks the word 'travel' is a match. So most platforms have exclusion lists and human review loops. Luna: Okay, so let's bring it back to the CMO's desk. If I'm a CMO listening to this, what's the one thing I should do this quarter to test contextual? Lucas: I'd say pick one brand, one emotional territory — not a product feature, but a feeling — and run a two-week pilot against a control. Use a platform that offers semantic analysis, not just keyword matching. Measure brand lift through a survey-based study. And compare the cost per lifted consumer against your standard display campaign. My bet is you'll see a surprising efficiency. Luna: And the feeling you choose should be based on real consumer insight, not a guess. Lucas: Absolutely. Don't just pick 'happiness' because it sounds nice. Look at your brand's equity data, your social listening — what emotional associations do your best customers already have? Amplify that. Luna: This is genuinely one of those rare moments where regulation and technology are pushing marketing in a direction that might actually be better for everyone — better for consumers, better for brands, better for the quality of ads. Lucas: I think so too. And honestly, if this conversation sparked a tactic you'll actually try, that's the link — buy me a coffee dot com slash fexingo. People who find value in these conversations are the reason we keep them ad-free and independent. Luna: Yeah, it's a small way to say 'this was worth my time,' and it genuinely helps us keep going. Lucas: So — back to that CPG campaign. One of the things the brand learned was that the contextual ads performed best on mobile, specifically in the app environments of news publishers. Desktop was fine, but mobile in-app had the highest attention time. Luna: Because people are more focused on their phone, maybe? Less multitasking? Lucas: That's the hypothesis. And it points to another insight: contextual targeting works best when the environment is lean-back, not lean-forward. So news apps, long-form video, even audio — podcast ads are a form of contextual targeting, when you think about it. Luna: Right — you're not targeting me, you're targeting the 'marketing podcast listener' context. It's the same principle. Lucas: Exactly. And I think we'll see more CMOs treat podcast advertising as part of their contextual strategy, not separate from it. The platforms are starting to converge. Luna: Alright, so what's the biggest risk you see in this shift? Lucas: The biggest risk is that CMOs swing too far and abandon all user-level targeting. There are still use cases where behavioral data matters — retargeting a high-intent shopper who abandoned a cart, for example. Contextual won't solve that. The smart play is a hybrid model: contextual for awareness and upper-funnel, first-party data for lower-funnel. Luna: And make sure your first-party data strategy is solid before you go all-in on context. Lucas: Exactly. Contextual is not a replacement for knowing your customer. It's a replacement for surveilling them. Two very different things. Luna: I think that's a good note to end on. Next episode we'll look at how CMOs are using retail media networks as a contextual play — because that's another fast-growing area. Lucas: Yes — and we'll get into the walled garden dynamics there. For now, thanks for listening. If today's conversation was useful, you know where to find us. Luna: See you next time.