An Amazon PPC audit is a structured review of an advertising account that ends in a short list of
decisions rather than a long list of observations. Done properly it takes about two hours and finds the
same three things in most accounts.
Those three things, in order of how much they usually cost: multiple products sharing one ad group so
nothing is attributable, missing negative keywords so campaigns bid against each other, and bids set from
Amazon’s suggestions rather than from the seller’s own break-even.
Work in the order below. The sequence matters, because several later checks are meaningless until
earlier ones pass. Auditing bids on an account with mixed ad groups tells you nothing at all.
Before you start
Pull three things and have them open.
The search term report, last 60 days. Long enough for real click volume.
The bulk operations file, which shows structure, bids, and negatives in one place and is far faster to
read than clicking through campaigns.
Your break-even cost per click per product. If this does not exist, calculate it first. Everything
downstream is guesswork without it, and our guide to
Amazon bid types and strategies covers the
arithmetic.
I’d refuse to start an audit without that third item. In our experience it is missing more often than not,
and it is the reason so many audits conclude “reduce ACoS” rather than something actionable.
The ten checks, in order
1. Products per ad group
Count them. Anything above one ASIN or one tight variation family is a finding, and I’d rank it first every time because it makes the rest of your data unreadable.
2. Negative keywords present
Check whether campaigns have any. An account spending for months with empty negative lists is paying
repeatedly for the same lessons and running internal bidding wars. Our guide to
negative keywords on Amazon covers the fix.
3. Search term duplication across campaigns
Filter the search term report for terms appearing in more than one campaign. Each duplicate is you bidding against yourself, and in our experience this check surprises people most.
4. Campaigns hitting budget caps
Identify which exhaust budget before the day ends. A capped campaign is throttled rather than optimized,
and its apparent efficiency is an artifact of the cap rather than a result.
5. Bids versus break-even
Compare actual cost per click against the break-even figure per product. Any keyword consistently above
break-even and not deliberately funding a launch is losing money on purpose without anyone deciding to.
6. Placement adjustments
Check whether top-of-search adjustments exist on proven keywords. In my experience they are frequently all at zero, which leaves the best-converting placement to whoever bothered.
7. Bidding strategy per campaign
Confirm that discovery campaigns are on down-only and that up-and-down bidding, where used, sits only on
keywords with genuine conversion history.
8. Keywords paused on thin data
Look for paused keywords with fewer than 10 clicks. This is the quiet killer. Accounts pause their way
into irrelevance one under-sampled term at a time, and the paused list is where you find the evidence.
9. Products advertised while out of stock
Cross-check active campaigns against inventory. Advertising unavailable products happens more often than
anyone admits, and it is pure loss.
10. Naming and structure legibility
Read the campaign names. If you cannot tell what a campaign is for, nobody has been managing it
deliberately, and I’d treat that as the real finding rather than a cosmetic one.
What the three common findings actually cost
Putting numbers on them, because “wasted spend” is easy to nod at and hard to act on.
Mixed ad groups. Take an ad group with 5 ASINs sharing a $50 daily budget and 40 keywords. Amazon
serves whichever product it judges most likely to convert, so 1 or 2 products absorb most of the spend
and the reporting attributes the keyword performance to the ad group rather than to a product. You cannot
tell which of the 5 earned the sale. The cost is not the spend, it is every subsequent decision made on
unreadable data. In our experience that compounds for months.
Missing negatives. With no negatives, a single search term can trigger 3 of your campaigns. You win
one auction, you pay for it, and the other 2 impressions were competition you funded against yourself.
Multiply across a few hundred terms and the waste is substantial without any single line item looking
wrong.
Bids from suggestions rather than break-even. Amazon’s suggested bid reflects auction competition,
not your economics. A product with a $6 contribution margin converting at 8% supports roughly $0.48 per
click. If the suggested bid is $1.20 and you took it, you are losing money on every sale and the ACoS
figure alone will not tell you, because ACoS knows nothing about your costs.
I’d quantify all three before presenting an audit. When I have shown a seller the arithmetic rather than
the observation, the fix happened that week.
Auditing an account you did not build
Slightly different job, and worth its own note because agencies and new hires inherit accounts constantly.
Do not change anything for the first week. Read only. The temptation to fix the obvious immediately
destroys the baseline you will need to prove the work later.
Screenshot or export the current state, including bids, budgets, and structure. Bulk files are ideal.
Without a before, you cannot demonstrate an after.
Ask what changed in the last 90 days before assuming the account was always like this. A restructure
three weeks ago explains a lot of apparently irrational data.
Find out who else has access. More than one person adjusting bids without coordination produces
patterns that look algorithmic and are not.
Reading the findings
The part that turns a checklist into an audit.
Group findings by cause, not by campaign. Twelve campaigns with missing negatives is one finding, not
twelve. Sellers who list every instance produce a document nobody acts on.
Rank by cost, not by count. One mixed ad group on your best-selling product matters more than eight on
products that barely sell.
Separate structural from tactical. Structure changes take effect over weeks and reset performance
history. Bid changes take effect immediately. Do the structural work first and deliberately, then tune,
because doing it the other way round means redoing the tuning.
Write down no more than five decisions. An audit finding thirty things produces no change. When I
have handed over a five-item list with owners and dates, the work happened. When I have handed over a
thorough report, it did not.
What an audit will not tell you
Worth saying, because audits get asked to explain everything.
Whether the product is any good. Advertising amplifies a listing’s existing conversion. If relevant
traffic arrives and does not buy, the problem is upstream and our guide to
Amazon Brand Analytics covers
diagnosing which funnel stage loses the sale.
Whether your price is right. Ads cannot out-spend a competitor who is cheaper with faster delivery.
Whether the category is worth being in. Sometimes the honest audit finding is that the margin does not
support the cost per click in that category at all, and I’d say so rather than optimizing toward a
ceiling.
How often
Quarterly for a stable account, monthly after any significant change: a new product launch, a
restructure, a budget shift, or a change of whoever manages it.
A full audit more often than monthly is not useful, because structural changes need weeks to produce
readable data. The weekly work is different: search term review and bid adjustment, which our guide to
Amazon PPC campaign structure covers.
If you would rather have this run properly than scheduled and skipped, that sits inside our
Amazon PPC service. We are an Amazon Ads
partner and an Amazon SPN Verified Partner.
FAQ
How do I conduct an Amazon PPC audit?
Pull the last 60 days of search term data, the bulk operations file, and your break-even cost per click,
then work through ten checks in order: products per ad group, negatives present, search term duplication,
budget caps, bids against break-even, placement adjustments, bidding strategy, keywords paused on thin
data, out-of-stock products still advertised, and structure legibility.
What does a PPC audit usually find?
Three things in most accounts: multiple products sharing one ad group so performance is not attributable,
missing negative keywords so campaigns bid against each other, and bids set from Amazon’s suggestions
rather than from the seller’s own break-even figure.
How long does an Amazon PPC audit take?
About two hours for a mid-sized account, provided the break-even figures already exist. Calculating those
first is what usually extends it, and skipping them makes the rest of the audit guesswork.
How often should I audit Amazon PPC?
Quarterly for a stable account, and monthly after a significant change such as a launch, a restructure, or
a change of manager. More often than monthly is not useful, because structural changes need weeks to
produce readable data.
Why should structural fixes come before bid changes?
Because bid data from a campaign with several products in one ad group is not attributable to any of them.
Fixing structure first also resets performance history, so tuning done beforehand has to be redone
afterward.
What should a PPC audit produce?
No more than five decisions, each with an owner and a date, grouped by cause rather than by campaign and
ranked by cost rather than by count. An audit that produces thirty observations produces no change.
Last updated: August 31, 2026. Amazon’s advertising console, report formats, bulk operations files, and
campaign controls change over time. The Amazon Ads console and Seller Central carry the current feature
set for your account.
