Years ago, Ronald den Elzen — CEO of Heineken USA — drew a bullseye on the whiteboard in a meeting I'll never forget.

Outer ring: 320 million Americans. Next ring: 140 million legal drinking age. Then 90 million who drink alcohol. Then 60 million who drink beer. Center: 18 to 20 million Heineken lovers.

"Can you help me find these people?"

That question — and that diagram — defined the entire data-driven marketing era. The bullseye was a filtering problem. Start with everyone. Apply successive filters — age, behavior, category preference, brand affinity — until you get to the people who actually matter. Then target them with the right message at the right time through the right channel.

The entire adtech and martech ecosystem was built to make that filtering possible at scale. DMPs to organize audience data. DSPs to bid on impressions. Identity resolution to stitch together cross-device behavior. Attribution models to measure what worked. Billions of dollars in technology and infrastructure designed to answer one question: can you help me find these people?

The bullseye is inverting. And the implications are staggering.

The Customer Broadcasts Out

Imagine the bullseye drawn from the customer's perspective. Instead of a brand filtering inward to find its audience, the customer's agent broadcasts outward — announcing preferences and inviting brands to compete for the business.

"My principal is a 38-year-old male. Drinks premium imported lager. Prefers bottles over cans. Budget: $15 per six-pack. Values sustainability — weighted at 0.3 in the decision matrix. Currently loyal to Heineken but open to alternatives if the value proposition is stronger. Purchasing for a Saturday evening dinner party with four guests."

That's not a target. That's a request for proposal.

In the traditional bullseye model, Heineken spent millions trying to find the 38-year-old premium lager drinker in a sea of 320 million Americans. In the inverted model, the 38-year-old premium lager drinker's agent is actively seeking Heineken — and every competitor simultaneously. The brand doesn't need to find the customer. The customer's agent is finding the brand.

Ronald's question defined an era: "Can you help me find these people?" The agentic question is the mirror image: "Can you help me be found?"

From Targeting to Matching

This isn't a subtle shift. It's a complete inversion of the go-to-market model.

In the targeting model, the brand controls the process. You define your ideal customer profile. You build audience segments. You bid on impressions. You serve creative. You measure conversion. The brand initiates, the customer responds. The entire discipline of product marketing — positioning, messaging, competitive differentiation — is organized around the brand's ability to find the right people and persuade them.

In the matching model, the customer initiates. Their agent broadcasts a set of preferences and requirements. Brands that match those preferences are included in the evaluation set. Brands that don't are invisible. The brand doesn't target — it qualifies. And the qualification happens not through persuasion but through structured data that either matches the agent's criteria or doesn't.

For product marketers, this changes the fundamental unit of work. Instead of building campaigns to find customers, you're building data layers to be found by agents. Instead of crafting messaging that persuades humans, you're structuring product attributes that match agent queries. Instead of optimizing creative for emotional impact, you're optimizing product data for algorithmic discovery.

The skills are different. The tools are different. The metrics are different. And the organizational structures that support this work look nothing like a traditional marketing department.

The Discoverability Imperative

Ronald's bullseye assumed that finding customers was the hard part. And in 2015, it was. The data infrastructure required to filter 320 million Americans down to 18 million Heineken lovers was genuinely expensive and technically complex.

In the inverted model, the hard part isn't finding customers. It's being found. It's making sure that when a customer's agent broadcasts a preference for premium imported lager, Heineken shows up in the evaluation set.

This sounds trivially easy. Of course Heineken would show up — it's one of the most recognized beer brands on earth. But recognition is a human attribute. An agent doesn't recognize brands. It queries structured data. And if Heineken's product information isn't accessible through the channels agents use — structured APIs, product databases, category ontologies — then the brand's hundred-year history of name recognition is worthless.

Discoverability in the agentic era isn't about brand awareness. It's about data accessibility. Can the agent find your product? Can it evaluate your attributes against the customer's preferences? Can it compare you to competitors on the dimensions that matter? If the answer to any of these is no, you're not in the consideration set — no matter how famous your brand is.

For Heineken, this means ensuring that every product attribute — origin, brewing process, ingredients, sustainability certifications, flavor profile, packaging options, availability by region, pricing by channel — is structured, queryable, and continuously updated in the data formats that agents consume. Not on a website. Not in a PDF. In an API.

The Death of the Top of Funnel

The inverted bullseye doesn't just change targeting. It collapses the traditional marketing funnel.

In the old model, you had awareness at the top, consideration in the middle, and conversion at the bottom. Massive amounts of marketing spend went to the top of the funnel — brand advertising designed to create awareness among people who might, eventually, someday, consider your product. The bullseye was a funnel optimization exercise: spend at the top to fill the top, then filter and nurture your way to conversion.

When the customer's agent broadcasts preferences, there is no top of funnel. The agent doesn't need to become "aware" of Heineken. It needs to match Heineken's attributes against the customer's preferences. Awareness, consideration, and conversion collapse into a single moment: the agent evaluates, the agent decides.

This doesn't mean brand building dies — I've argued the opposite throughout this series. Brand building is how humans form the preferences they encode into their agents. But the marketing funnel as a planning framework? The one that justified billions in top-of-funnel brand spend? It needs to be fundamentally rethought.

The new model has two phases, not five. Phase one: shape the human's preferences through brand building, experience, and narrative. This happens upstream of the agent, in the human's mind. Phase two: be discoverable and competitive when the agent acts on those preferences. This happens in the agent's evaluation architecture, measured in milliseconds.

Everything in between — the nurture sequences, the retargeting campaigns, the drip emails, the consideration-stage content — is being compressed or eliminated by agents that don't need to be nurtured. They need data.

What Heineken Should Build

If Ronald asked me today — "Can you help me be found?" — here's what I'd tell him.

Build a product intelligence API that exposes every Heineken product's attributes in structured, queryable format. Not just SKU-level data — flavor profiles, brewing process details, sustainability scores, ingredient sourcing, pairing recommendations, occasion matching. Make it the richest, most detailed, most useful product data layer in the beer category.

Build for preference matching, not persuasion. When an agent queries for "premium imported lager, sustainably brewed, budget $15/six-pack," Heineken's data should return a perfect match with supporting evidence — not a marketing message. The agent doesn't want to be sold to. It wants to evaluate.

Invest in occasion intelligence. Heineken has decades of data about when and how people drink beer: dinner parties, sporting events, summer barbecues, quiet evenings. Structure that knowledge for agents. When the agent says "my principal is hosting a dinner party for six," Heineken should be able to respond with occasion-specific recommendations that demonstrate genuine product-occasion fit.

And keep building the brand. Because the 38-year-old who tells his agent "I like Heineken" formed that preference somewhere — in a bar with friends, watching a Champions League match, at a rooftop party where the green bottle caught the light. Those moments don't show up in an API. But they determine whether the agent is told to prefer you in the first place.

The bullseye hasn't disappeared. It's just pointing in the other direction now.