The Future of Ad Targeting: Why "Deep Context" Beats Cookies Every Time
What I learned from my conversation with Brendan Norman, Co-Founder of Classify
I just recorded a podcast with Brendan Norman, the founder of Classify. Before his current gig, Brendan helped build the Facebook Audience Network, so he has seen the ad tech machine from the inside out.
It’s funny to think a lot of us have spent the last five years obsessing over the death of the cookie. But while the industry was panicking about tracking, Brendan was building something better.
My conversation with him made thing clear that the future of advertising isn’t about finding new ways to track people. It’s about using AI to understand the context of where they are right now.
Here are my biggest takeaways from our conversation:
Contextual 3.0 is not your grandfather’s keyword targeting
For years, “contextual targeting” just meant spotting a keyword like “skiing” and slapping a ski ad next to it. Brendan argues we are now in Contextual 3.0. This is where AI reads a page like a human does—understanding sentiment, entities, and nuance. It knows the difference between a page selling ski gear and a news article about a ski accident.
“Contextual 2.0... is having more automated machine learning systems... And then three is really reading the entire content of the page and attributing hundreds and hundreds of different vectors... It’s talking about this specific ski brand. It’s talking positively and negatively about it.”
My take: This is the breakthrough B2B has been waiting for. If you can target “positive sentiment” around “cybersecurity implementation” rather than just the keyword “cybersecurity,” you aren’t just buying impressions—you’re buying intent. It moves us from “spray and pray” to surgical precision.
AdTech is just “Old City Plumbing” (and it’s leaking)
Brendan used a perfect analogy: AdTech is like New York City plumbing. It works, but the pipes are old, lead-lined, and leaky. You can keep patching them (cookies, IDs), or you can lay new pipe. Brendan is betting on the Ad Context Protocol (MCP), a standardized way for data providers to talk directly to ad buyers without the messy middleware.
“The plumbing in New York City is old... these pipes have been around a long time... But if you can build a brand new something from scratch right now using modern technology, it is just by default going to be much more efficient.”
My take: We often forget that “innovation” in marketing usually just means adding another tool to the stack. Real innovation is removing the friction entirely. If MCP takes off, it could democratize access to premium targeting for everyone, not just the big players. It’s time to stop patching the leaks and start replacing the pipes.
Privacy isn’t “Theater”—It’s a Performance Hack
I asked Brendan if contextual targeting was just “privacy theater”. He pushed back hard. His view? Privacy means not tracking Kevin. Contextual targeting doesn’t care who you are; it cares what you are engaging with right now. And because the ad matches the moment, not the person’s past history, it often performs better.
“The fact that a certain ad has the ability to show up against a specific type of content... is completely disconnected from any concept of privacy at all... If you see an ad about something that’s really related to the article that you’re reading... you are much more likely to engage with that because you’re already in the mindset.”
My take: We’ve gotten lazy with behavioral tracking. We assumed that because we could follow someone across the web, we should. Brendan’s approach reminds us that relevance (right message, right time) beats surveillance (right person, wrong time) every day.
The “Agentic Web” will negotiate for you
We are moving toward a future of “AI Agents.” Brendan predicts that media buying will soon shift from manual dashboard management to autonomous agents negotiating with other agents. These systems will identify the best inventory, price, and creative fit in real-time, freeing up marketers to focus on strategy.
“Agents talking to other agents... They can figure out what’s the best ad creative to deliver for this specific campaign... without necessarily having to have a human involved in making a lot of these manual decisions.”
My take: This scares some people (”will it take my job?”), but Brendan sees it as liberation. If AI handles the negotiation and placement, marketers can get back to what we’re supposed to be doing: strategy and creative. The “Agentic Web” doesn’t replace us; it promotes us.
The bottom line:
The death of the cookie isn’t a crisis; it’s a correction. By using AI to understand “deep context,” we can finally stop stalking users and start actually connecting with them. Brendan is building the infrastructure to make that happen, and honestly, it can’t get here fast enough.
Listen to the full conversation:
Catch all the Tech Marketing Rewired Podcasts here. Have an idea for a guest or podcast topic, let me know!