Stop treating every channel like its own business

The most dangerous marketing dashboard is the one that looks clean.

Meta has a number. Google has a number. Amazon has a number. TikTok has a number. Each one appears to tell you whether that channel is working. Each one gives the team something tidy to optimise. And each one quietly encourages the same mistake: treating your business like a bundle of separate little P&Ls instead of one demand system.

That mistake is becoming more expensive.

The received wisdom in paid media is still platform-first. If Meta ROAS drops, cut Meta. If branded search looks profitable, protect it. If Amazon is reporting conversions, feed it. The channel with the cleanest attribution gets the money, even when it may simply be collecting demand that another channel created.

This is how brands end up overpaying for their own name.

The useful question is not “which dashboard claims the sale?” The useful question is “which activity created a customer who would not otherwise have arrived?”

So the budget flows towards the collectors

Branded search often looks heroic because it sits close to the point of purchase. Amazon can look unbeatable because people go there when they already know what they want. Google can look clean because intent is visible.

Meanwhile, the channels doing the harder job, creating awareness, changing preference and starting the journey, can look weak inside a 30-day reporting window.

The useful pattern shows up repeatedly in channel case studies: a large share of platform-attributed conversions can come through branded terms. When that happens, the platform may be collecting demand the rest of the system already created.

That does not prove every brand should slash Amazon or branded search. It proves something better: platform attribution is not a strategy.

Improve measurement without worshipping it

You can improve measurement. You can run holdouts. You can separate branded from non-branded demand. You can look at new visitor rates, assisted journeys, incrementality, contribution margin and payback.

You can absolutely get smarter. But you are still managing a system with lag, overlap and confounding variables. The goal is not perfect attribution. The goal is better budget decisions.

Split demand creation from demand capture

For mature brands, every serious media review should split demand creation from demand capture. Branded search, retargeting and Amazon brand terms need to be treated as collectors unless proven otherwise.

Meta, TikTok, native, YouTube, CTV and organic content need evaluation windows that match their job. If cold traffic typically takes 60 to 90 days to turn into customers, killing the test after one monthly review is not discipline. It is impatience dressed up as financial control.

For younger brands, the answer is different. Do not use this as an excuse to spray money at awareness before you have conversion proof. Build demand capture first. Get the offer, landing page, retargeting, search hygiene and sales path working. Then test demand creation in controlled slices.

The practical move this week

Pick one product line or campaign and build a one-page blended scorecard:

• Total spend
• Total revenue
• Units sold
• New customers
• Blended CAC
• Contribution margin
• Branded vs non-branded search split
• New visitor rate by channel

Then ask which channels are creating demand, which are collecting it and which are being overpaid because they stand near the till.

That one exercise will tell you more than another week of staring at platform ROAS.

The brands that win from here will not be the ones with the prettiest attribution dashboard. They will be the ones brave enough to manage the whole system.