The omnichannel argument: why brands shouldn't have four agencies

Most consumer brands I work with sell in five places. Their own Shopify store. Amazon. Walmart. Meta. Google.

They have, on average, three agencies. Sometimes four. One handles paid social. One handles Google. One handles Amazon (often grudgingly, often badly). One handles Walmart, if Walmart matters enough to merit its own. The brand internally tries to coordinate.

Each agency reports separately. Each agency uses its own dashboard, its own attribution model, its own metric definitions. ROAS in the Meta report and ROAS in the Amazon report are not the same number. ACOS doesn’t exist outside Amazon. Total ad-attributed revenue means one thing to the Google agency and a different thing to the Amazon agency.

The brand gets four reports per month. They internally try to combine them into a single view. They almost never succeed, because the metrics aren’t comparable, the time windows don’t align, and the agencies aren’t incentivized to make their data play with anyone else’s.

What the brand ends up with: a vague, channel-by-channel view of performance and no way to make portfolio decisions. They can’t tell whether a dollar is better spent on Meta or on Amazon DSP, because the two channels don’t report in the same language. They can’t tell whether their Walmart growth is cannibalizing Amazon, because the two reports are produced by different teams with different attribution windows. They can’t see incrementality across the mix because nobody is set up to measure it.

This is the structural problem that drove me to start AdOdyss. The fix isn’t better reports from each agency. The fix is one agency that runs the whole mix.

The argument against: specialists are better than generalists. A pure Amazon agency knows Amazon better than an omnichannel agency. A pure paid-social agency knows Meta better than a generalist.

The argument is partially right. A specialist agency in 2020 was almost certainly better at its specialty than a generalist. In 2025, with the platforms converging on similar mechanics (creative-led performance, audience-driven targeting, attribution models that increasingly look the same), the gap is closing. The cost of running multiple specialists has stayed the same. The benefit of running one generalist who can see across channels has grown.

The brands that will outperform over the next five years are the ones who can see their portfolio in one view, make decisions across the portfolio, and run a shared operating system across channels. Not the ones who hire the best Amazon specialist plus the best Meta specialist plus the best Google specialist and try to coordinate them.

The “you have a product, we sell it everywhere” thesis is just this argument applied operationally. One company. Same structure across platforms. One combined number. The brand makes portfolio decisions instead of channel decisions. The math becomes legible.

I’m building toward this with AdOdyss. We’re not there yet — Amazon and Walmart are the strongest pieces today, with the other channels growing — but the trajectory is intentional. The next five years of agency work, in my view, belong to the firms that solve this problem properly.