When Wolves Teach Sheep to Build Their Own Kitchen

The Pattern Everyone Sees, But No One Questions

There's a strange pattern emerging across advertising right now. Agencies are reselling Amazon Ads. Media companies are packaging Amazon Ads. Retail media networks are trying to position themselves as Amazon Ads equivalents. At the same time, executive teams are actively recruiting former Amazon Ads leaders to "Amazon" their organizations.

On the surface, it looks like alignment with the future. But underneath it, there's a much more uncomfortable question.

If the wolf is teaching you how to build your kitchen, who exactly is dinner?

Reselling Is Not Owning

Amazon may look like just another media channel, and much of the ecosystem is treating it that way and drinking the kool-aid. That creates real short-term advantages. Faster go to market, easier monetization, and a clean story to tell internally and externally.

But that framing misses what is actually happening. The more your business runs through someone else's platform, the less control you have over the things that matter. You do not control the data, the roadmap, or the economics.

You are participating in someone else's system, not building your own.

When Everyone Sells the Same Thing

There is a second order effect that is already starting to show up. If everyone is reselling the same underlying product, differentiation begins to disappear. And when differentiation disappears, the only remaining lever is price.

In advertising, price shows up as CPM, which ultimately means margin. What starts as a strategy conversation slowly becomes a pricing conversation. What starts as value creation quietly turns into value compression.

To compensate, most players try to layer in differentiation.

• Amazon, but with our data.
• Amazon, but with our service.
• Amazon, but with our proprietary audiences.

These additions can help in the short term, but they are still built on the same core product.

Toppings and sprinkles don't change who owns the kitchen.

The System Is Learning Faster Than You Think

At the same time, Amazon is aggressively building agentic systems that connect creative generation, audience construction, media planning, buying, optimization, and measurement directly to transaction level outcomes. Not as separate tools, but as a unified system.

That shift matters because the current reseller ecosystem exists in the gaps between those functions. Planning, execution, and measurement have historically required coordination across tools and teams, which created space for intermediaries. As those gaps close, the value chain compresses.

And when the value chain compresses, intermediaries get squeezed.

The Margin Illusion

Right now, pricing dynamics feel favorable. Lower take rates, more competitive CPMs, and faster adoption all make it seem like the ecosystem is winning.

But platforms do not compress margins out of generosity. They do it to accelerate control. Lower the barrier to entry, capture more demand, aggregate more data, and train better systems. Then, once the system is embedded, rebalance the economics.

This playbook is not new. We have seen versions of it before with a Search Giant. Incentives and favorable economics exist while adoption is being built, and tend to disappear once dependency is established.

If you are not paying the right price for a product, you are the product.

In this case, what is being collected is not just revenue. It is signal. Every campaign, every optimization, and every measured outcome feeds the system that will eventually automate more of the workflow.

What Happens Next

Project this forward a few years. It's a Black Mirror episode, and the year is 2028. Most of the industry has standardized around reselling the same infrastructure. Pitch decks look nearly identical. Media plans follow the same patterns. Differentiation is explained through small variations in data or service.

Then the platform releases a fully realized agent. An advertiser defines an objective, connects their data, and the system handles creative, audience, budget allocation, optimization, and measurement in one place. No manual workflows, no fragmented systems, and far fewer reasons to rely on layers in between.

What exactly is left to resell?

How Not to Become Dinner

If this trend continues, the question is not whether platforms get stronger. They will.

The real question is how you avoid becoming fully dependent on them.

A few principles are starting to separate the companies that will keep leverage from the ones that won't:

• Own data that does not originate from the platform — If your most valuable signals come from someone else's ecosystem, you are training their system, not yours.
• Prioritize outcomes, not access — Access can be commoditized. Outcomes tied to real business results are much harder to replace.
• Treat individual platforms as inputs, not your foundation — The moment a platform becomes your business model, you have already given up leverage.

A Different Path Forward

There is an alternative to this trajectory, but it requires a different mindset.

Instead of building on top of a single platform, the next generation of companies are building systems that sit above and across them. They treat individual platforms as inputs, not the foundation, and build their strategy on systems that allow them to operate across all of them without giving up control.

The distinction is not platform versus no platform. That is not realistic. The distinction is whether the platform increases your leverage or replaces it. The safest position is not avoiding platforms entirely, but using systems that help you maintain control of data, relationships, and outcomes across environments rather than locking you inside a single ecosystem.

They focus on owning the connection between audience, activation, and outcome, not just access to inventory.

The advantage is not in access to media. It is in co-owning the workflow between identity, activation, and measurable outcomes.

That requires infrastructure that can operate across the open internet, not just inside a single walled garden. It requires tying real world outcomes back to the household or customer level, and using that signal to continuously inform planning and activation.

And increasingly, it requires an intelligent layer that can connect all of it. Not just automate tasks, but help agencies plan, activate, optimize, and measure across identity, media, and outcomes in a continuous loop.

In that model, platforms like Amazon still play a role.

They are just no longer the center of gravity.

The Only Question That Matters

This is not an argument against platforms. Platforms will continue to shape the future of advertising. But building entirely on top of one, especially one that is actively working to collapse the layers you sit within, comes with real risk.

The companies that win in the next phase will not be the best resellers of someone else's platform. They will be the ones that own something no single platform can replace. Proprietary data outside the platform, workflows that span across environments, direct relationships, and closed loop measurement tied to real business outcomes.

If the platform keeps getting better, cheaper, and more automated, what do you actually own?

Because if the honest answer is access, packaging, or slight variation, that is not a durable strategy. It is a dependency.

And dependencies tend to end the same way.

The wolf does not need help forever. It just needs it long enough to learn how the kitchen works.