There is a version of Amazon PPC account management where complexity is the deliverable. New campaigns get added regularly. Old campaigns never get deleted. Over time the account accumulates hundreds, then thousands of campaigns, each created for a reason that made sense at the time and none of which have been audited since.
The account looks comprehensive. It isn't. It's a graveyard with a management fee attached.
How accounts get here
Amazon PPC proliferates by default. You run an auto campaign. You mine it for search terms and create exact match campaigns. You add broad match to capture variations. You test competitors. You run DSP retargeting. You have a campaign for every ASIN, a campaign for every keyword cluster, a campaign for every seasonal push.
Over three years, that's 3,000 campaigns. Not a hypothetical number. One account we reviewed had exactly that. Three thousand active campaigns across a catalog that didn't justify it. The agency running the account treated the count as evidence of thoroughness.
The problem isn't the number by itself. The problem is what 3,000 campaigns means for management quality. No human being can meaningfully optimize 3,000 campaigns. What actually happens is that maybe 50 to 100 campaigns get regular attention, and the rest run on autopilot spending money at mediocre ACoS against targets no one remembers setting.
The 70% buy box filter
Before you can run an effective PPC campaign on an ASIN, you need to hold the buy box. When you don't hold the buy box, your ad runs, someone clicks it, and the buy box goes to someone else. You paid for the click. Someone else got the sale.
The 70 percent buy box share threshold is the filter we apply: any campaign pointing to an ASIN where you're winning the buy box less than 70 percent of the time is not a real campaign. It's a donation to Amazon's ad revenue.
When we applied that filter to the 3,000-campaign account, here is what happened: 2,947 campaigns were ineligible under the rule. The brand was actively winning the buy box at 70 percent or above on ASINs supported by 53 campaigns.
Fifty-three.
That's the actual number of campaigns worth managing. The other 2,947 were real campaigns by Amazon's definition, live and spending, but they were running against ASINs where the brand wasn't reliably winning sales even when they generated clicks.
What "optimization" usually is
The agency managing that 3,000-campaign account would send weekly reports showing bid adjustments, new negative keywords added, campaign-level ACoS percentages. It looked like active management.
It was portfolio management of a broken system. Adjusting bids on campaigns pointing at ASINs you don't control is rearranging furniture in a room with a flood. The bids don't matter. The buy box is the problem.
Real optimization starts with the precondition list. Before spending on an ASIN:
The ASIN needs to hold the buy box above 70 percent. If it doesn't, find out why. Is there a competitive seller with lower price? Is there a newer offer? Is there an account health issue suppressing your listing? Fix the buy box problem before you put a dollar into advertising that ASIN.
The ASIN needs enough sales velocity to generate statistically meaningful data. The general rule: 80 to 100 units per month minimum before PPC spend makes analytical sense. Below that, the data is too thin to optimize against.
The ASIN needs a margin structure that can absorb ad spend. Run the contribution margin before you run the campaign. An ASIN at 15 percent gross margin cannot support a 25 percent ACoS without losing money per unit sold.
When you apply those three filters, most large Amazon accounts collapse to a manageable number of campaigns that actually deserve active management.
What to do with the rest
Pause, not delete, campaigns where the buy box issue is temporary. Seasonal products, newly repriced ASINs, products where a suppression issue is being fixed. Pause them and set a reminder to revisit when the underlying condition is resolved.
Delete campaigns that have run for 90-plus days with zero impressions or zero clicks. These are structural dead ends. No amount of bid adjustment will revive a campaign that Amazon's system has decided is irrelevant.
Archive data before any bulk deletion. Download the campaign performance reports so you have historical ACoS and spend data. You'll want it for the accounts that actually matter.
The right account looks different
A well-managed Amazon PPC account has somewhere between 20 and 200 active campaigns depending on catalog size. Every campaign has a defined purpose. Every ASIN being advertised holds the buy box above the threshold. The person managing the account can tell you, from memory, what the top five campaigns are, why they exist, and what they're returning.
If your PPC manager cannot do that, the problem isn't the campaigns. It's the account design. And adding more campaigns to an account that already doesn't work won't fix it.
