The Activation Reset: Why Platform Performance Isn't Enough

The value of the walled gardens remain essential to part of a plan. They offer scale, data, optimisation and buying simplicity.

The problem starts when convenience becomes dependency.

Walled garden platforms can tell you what happened inside their own system. They can report attributed conversions, reach, frequency, engagement, ROAS and modelled outcomes. All of them are useful signals, but they are not always proof.

The harder question to ask is, “what changed because the media ran?”

A campaign can be attributed without being incremental, so making that distinction is increasingly important. Conversions can follow an ad without being caused by it, and a channel can look efficient because it’s close to existing demand, not because it created new demand.

This is where performance marketing needs to get more grown up.

Too simplistically and too often, open web is framed as a brand environment, while platforms are where performance happens.

The open web has its own challenges; supply quality, transparency, measurement consistency and fragmentation all need active management, but treating it as leftover reach is a mistake.

As platform ecosystems become more closed, the open web can offer things that matter. In terms of independent environments, publisher quality, verification, supply-path choice and more flexibility in how a campaign is built.

The point is to reach a balance, the point is not “open web good, walled gardens bad”.

Mature activation plans shouldn’t be built around which platform has the neatest dashboard. They need to be built around what behaviour needs to be influenced, which environments are best placed to influence it, and how we will know whether that influence was real.

The industry is moving in this direction for a reason.

IAB Europe’s 2026 Attitudes to Digital Advertising report found that performance and efficiency remain major drivers of digital investment, while media quality issues such as fraud, brand safety, viewability and transparency remain key barriers to growth.

Marketers want performance, but they also want confidence that the performance is real. Can you feel the tension?

Incrementality is still a large part of the overlooked answer because it asks the better question, “What outcomes did the campaign cause that would not have happened anyway?”

That doesn’t mean every campaign needs a perfect scientific test, but it does mean we should stop treating correlation as causation.

Clicks, platform ROAS, view-through metrics, etc all matter, but they should sit inside a broader evidence plan, not become the plan itself.

This is where retail media is a useful example. Its growth is partly driven by the promise of clearer links between media exposure and outcomes, but even there, the same discipline has needed to apply. More measurable does not automatically mean more meaningful.

The next phase of activation should be less platform-led and more proof-led. Meaning asking better questions before the plan is built:

• What behaviour are we trying to change?
• Are we creating demand, capturing demand, reassuring people, or reinforcing a choice?
• Which environments are best suited to that job?
• What would count as evidence that the behaviour has moved?
• Which metrics are signals, and which are proof?

This is where behavioural planning matters, by defining the human job-to-be-done, “What needs to change in someone’s mind or behaviour, and what should happen next?” Gives the media plan a clearer organising principle before channels enter the conversation.

Once that behavioural job is clear, you’re able to fit an activation solution to the response. The value is not adding more complexity, it’s giving teams more ways to execute, optimise and learn across environments, without defaulting to whichever platform is easiest to buy or easiest to report.

That is the real media activation reset, not fewer platforms, but better reasons for using them.

Walled gardens are useful, but they’re not enough on their own and there’s a tendency to over-rely on them. Brands need activation plans that balance walled gardens, the open web and emerging commerce environments around the behaviours they are trying to influence. That means defining the job first, choosing solutions second, and measuring with a clear distinction between correlation and causation. The result is a media plan that is easier to defend, easier to learn from, and less likely to mistake a good-looking dashboard for real growth.