317 campaigns, 2.2% of spend: what campaign bloat actually costs

Campaign bloat is usually sold as a money problem: too many campaigns, too much wasted spend. In the account behind our shower curtain liner audit, the numbers say something else. The extra campaigns barely spent anything. What they took away was the ability to read the account.

The account

By most measures this is a healthy account. Click conversion was 37.8%, and long-tail search terms with zero sales cost $1,393 — 0.3% of spend. The keyword book was clean.

What happened on July 29

The account’s change log shows a bulk tuning wave on July 22–23 (171+ bid changes, 37+ placement-premium changes), then, on July 29, a “pipeline” system of more than 300 campaigns built in a single day. Campaign names followed a strict template, for example:

SKU-72X72|SP|KW|MKW|01|Shower Curtain Liner|E|Mid|TOS|3

— suffix-numbered up to 39.

Three weeks later the split looked like this:

If you only look at dollars, the conclusion is easy: 317 campaigns spending 2.2% are a rounding error. Leave them.

What the 317 cost instead

The audit grouped the dead weight into four buckets. They overlap — some campaigns sit in more than one — so they don’t add up to 317:

None of these is expensive. Each one makes a different part of the account harder to read.

1. No campaign gets enough data to judge a bid

Seventy-four campaigns shared one keyword root — the same root the account’s two test campaigns were already running as an exact keyword, with $98,884 of spend between them. The 74 pipeline campaigns spent $895 together: about $12 each over three weeks (our arithmetic).

Twelve dollars is about seven clicks at the account’s average CPC of $1.61. You can’t tell a good bid from a bad one on that, and you can’t tell whether one variant of the campaign is better than another. Seventy-four rows of noise sit next to two rows that carry the signal.

A common worry here is that duplicate campaigns bid against each other and push up your own CPC. Amazon doesn’t let one advertiser’s campaigns compete in the same auction, so that usually isn’t the cost. The cost is that your data for one keyword is scattered across dozens of campaigns, and no single one of them is readable.

2. Budgets stop meaning anything

The 114 campaigns with $50 daily budgets spent nothing in 30 days. That makes the account’s total daily budget a number nobody can plan with: it includes capacity that isn’t used today and could switch on tomorrow if search volume or a bid elsewhere changes. When you can’t tell which budget is real, pacing and forecasting are guesses.

3. Changes become hard to review

The account logged 1,738 changes in the 30 days. Most of them were reasonable — the July 22–23 tuning wave was followed by orders up 31% from week 2 to week 4 with ACOS holding at 36.6%. But when hundreds of campaigns move at once, finding the one change that mattered (or the one that hurt) means reading through hundreds that didn’t.

4. The working core gets harder to see

The audit’s campaign map shows a tight gray cluster at 33–37% ACOS: the legacy campaigns that do almost all of the work. Its advice for that cluster was short — leave it alone. Only six campaigns needed a hand. With 317 extra rows in every report, that cluster is still there, but it takes longer to find, and every export, filter and pivot gets slower.

How to collapse them without orphaning a keyword

The audit’s cleanup action took the account from 422 campaigns to 130 or fewer, in an estimated two to three hours, at no cost:

1. Pause the 114 zero-spend campaigns.
2. Collapse the 74 duplicates on one keyword root into 2.
3. Keep 1 of the 104 Sponsored Brands minis.
4. Audit, then pause, the 57 that spent without selling.

The one rule that matters: before you pause a campaign, check whether it is the only place a keyword lives. Bulk-built systems often contain a handful of keywords that exist nowhere else. Pause their only host and you’ve silently stopped advertising on them.

A practical way to check: export your targeting report, group it by keyword and match type, and list every keyword whose campaigns are all on your pause list. Move those keywords into a surviving campaign first — or, as the audit puts it, reopen any campaign that turns out to have been a keyword’s only host.

What cleanup won’t fix

Collapsing 317 campaigns gave this account clean reads. It didn’t change its ACOS much, because the money was never there. The audit found it in two places:

• Placement premiums on the wrong words. The curtain words — about $77K of spend at 47–64% ACOS, converting at a half to two-thirds the rate of the liner words — sat in campaigns with top-of-search premiums of up to 200%. (More on that finding.)
• The price itself. At $1.61 a click and 42% click-to-unit conversion, every unit carries $3.83 of ad cost on a $10.49 average price. That is a 36.6% ACOS by arithmetic, and even the best keyword in the account only gets the floor down to about 33%. (Try it with your own numbers in the ACOS calculator.)

That is why the report’s bottom line wasn’t “cut campaigns”. It was: “Don’t cut keywords. Reprice the structure.”

Is your account bloated?

A quick check you can run today:

• Count vs spend. Sort campaigns by spend. As a rule of thumb, if the top quarter of campaigns carry more than 95% of spend, the rest are mostly noise.
• Budgets with no spend. Filter for campaigns with a daily budget and zero spend in the last 30 days.
• Many campaigns per keyword. Group targets by keyword and match type; anything hosted in more than three campaigns deserves a look.
• Bulk creation dates. Dozens of campaigns created on the same day usually came from a template — check whether that template is still earning its place.

If two or more of these are true, a cleanup will make every other decision easier — just don’t expect it to be the decision that moves your ACOS.