The problem with scale

Why ad accounts fall apart as they grow

Most large ad accounts are not badly managed. They are badly structured. They grew one campaign at a time, each added to solve a problem that week, until the account became a pile of overlapping campaigns bidding against each other on the same terms. Nobody set out to build that. It accreted.

The cost is invisible until you look for it. Two campaigns compete for the same keyword and drive up your own cost. Branded and non branded spend blur together so your reported efficiency flatters you. Budget sits on tired terms while new demand goes unfunded. A clean architecture fixes all of this at once, because it gives every dollar a clear job and a clear place to live.

The building blocks

What each campaign type is actually for

Amazon gives you a handful of advertising products. Used well, each does a distinct job. Used carelessly, they overlap and compete.

Sponsored Products

The workhorse. It places your product in search results and on product pages, and it carries the majority of most accounts’ spend and sales. This is where your keyword and targeting structure has to be cleanest, because it is where the money moves.

Sponsored Brands

Available to brand registered sellers, it puts your logo and a custom headline at the top of search, sending shoppers to your storefront or a curated set of products. Its job is brand presence and defending the top of the page, not squeezing the last point of efficiency.

Sponsored Display

It reaches shoppers on and off Amazon, including retargeting people who viewed your product but did not buy. Its job is to stay in front of demand you have already touched and to defend your own product pages from competitors.

Amazon DSP

The programmatic layer for audiences and remarketing at scale, with its own minimums and its own learning curve. DSP earns its place once the lower tiers are running clean and you have a real audience strategy, not before. It is an amplifier, not a fix for a messy account.

The structure

Separate by intent so nothing cannibalizes

The principle that holds a large account together is separation by intent. Branded search, competitor and category exact terms, broad generic terms, and discovery each get their own campaigns with their own budgets and their own targets. When they are mixed, you cannot tell what is working, and your campaigns end up bidding against your own other campaigns.

Branded campaigns defend your name cheaply and should run at the most efficient target in the account. Generic and discovery campaigns buy new customers and should be allowed a more tolerant target, because the value of a new buyer is not in the first sale. We cover the full intent tier logic and how to set a target for each in a dedicated piece, linked below. The structure here is what makes that targeting possible: you cannot set a target by intent if your intents are all in one campaign.

Use negative keywords as the seams between tiers. A term that belongs to your exact campaign should be negated in your broad campaign, so the broad campaign keeps discovering new terms instead of re buying the ones you already own. Without that discipline the tiers leak into each other and the separation collapses.

Branded defense

The branded question every brand gets wrong

Branded search is the most argued over line in Amazon advertising. One camp says never pay for your own name, since those shoppers would find you anyway. The other says always defend it, or a competitor will buy the space above you. Both are too absolute.

The honest answer is that branded spend mostly harvests sales you were likely to win, so it should be cheap and measured, not a budget you pour into because the ACOS looks beautiful. Defend the top of your branded results enough that a competitor cannot sit above you uncontested, but do not fool yourself that branded efficiency is growth. It is protection. The growth lives in the tiers where shoppers do not yet know your name.

The judge

Total ACOS is the only honest scoreboard

Inside a clean structure it is tempting to optimize each campaign to its own target and call the job done. That is how accounts drift. Branded looks efficient and tempts more budget. Discovery looks expensive and tempts cuts. Optimize each campaign in isolation and you can shrink advertised sales while convincing yourself you improved.

The number that judges the whole machine is total advertising cost of sale: your entire ad spend divided by your total sales including organic, not just the sales the ads get credit for. When total ACOS trends down over time while you keep funding discovery, the architecture is working, ads are building organic rank and the brand is compounding. When it stays flat no matter how you tune campaigns, the spend is circular and the structure needs another look.

The cadence

Run it on a rhythm, not on panic

A seven figure account is not optimized in heroic weekend sessions. It is maintained on a rhythm. A weekly pass to harvest converting search terms into your exact campaigns and negate the waste. A regular review of budgets against each tier’s job. A slower cadence for the structural questions, like whether a new product line needs its own campaign set or whether DSP is finally worth turning on.

The brands that win on Amazon advertising are rarely the ones with the cleverest single tactic. They are the ones whose structure lets them see clearly and act consistently, week after week, while competitors thrash. Build the architecture once, then run it with discipline.