Amazon Account Audit: The 30 Things We Check

Key takeaways: An account audit is 30 checks across six levers: advertising, listings, creative, catalog health, inventory and unit economics. Each lever gets a 1 to 5 score, and the plan starts on the lowest score with the biggest sales behind it. The audit takes about two weeks with access, and the output is a scorecard and a written 30/60/90-day plan, not a sales pitch.

Why the same 30 checks every time

An audit that changes shape with each account cannot be compared with anything, and an audit run by whoever is free that week misses what they do not usually look at. Ours is a fixed list. The checks below are the ones that, across the accounts we have run, most often explain why the numbers are what they are. The scoring and the plan that follow are set out in A6 · Audit, diagnose, plan — your account health scorecard; this article is the checklist itself.

Advertising: 10 checks

1. Wasted spend share. Search term spend with zero orders over 60 days, as a percentage of total. Above a fifth is a structural problem.
2. Campaign purpose. Does every campaign have one job: discovery, control, defense or brand? Overlapping campaigns bid against each other.
3. Match type discipline. Are proven terms in exact match with their own budgets, or still living in broad and auto where they cannot be controlled?
4. Negative coverage. Are negatives applied where discovery terms are graduated, so the same term is not bought twice?
5. Placement performance. Top-of-search versus rest-of-search versus product page ACoS, and whether modifiers match the result.
6. Budget utilization. Campaigns running out early in the day, and campaigns spending under 70% of budget. Both waste.
7. Bid logic. Whether bids reflect the job of the keyword, ranking or profit or discovery, or one blanket figure.
8. Ad type mix. Sponsored Brands and Display spend as a share, and whether each is measured on its own outcome.
9. Branded versus non-branded. How much of the reported ad sales are shoppers who searched your brand name, which inflates ACoS performance.
10. Change history. Whether edits were logged and whether the account has been touched at all in the last 30 days.

The advertising portion follows the seven audits in B12 · How we audit your advertising account, and the bid-level detail is in the checklist below.

Amazon PPCAmazon PPC Audit Checklist: 20 Checks Before We Change a BidThe 20-point audit we run on every Amazon advertising account before touching anything: setup, structure, targeting, bids, budgets and the listings ads land on.

Listings: 6 checks

11. Indexing on core terms. Is the listing indexed for the ten terms that matter, verified by search rather than assumed?
12. Title structure. Within the length limit, leading with the terms that rank, readable on a phone.
13. Bullets and description. Do they answer the questions shoppers actually ask, in language that indexes, without policy risk?
14. Backend search terms. The 250-byte field used fully, without repeats of words already in the title or bullets.
15. Attributes and variations. Missing attributes that block filters, and variation families that should be merged or split.
16. AI-search readiness. Whether the listing answers the comparison and use-case questions Rufus draws on.

Creative: 5 checks

17. Main image. Click-through rate against the results-page competition, and whether the image shows the product's size and use.
18. Gallery completeness. All seven slots used, each with a job, and whether infographics are legible on a phone.
19. A+ Content. Present, current, comparison module included, and measured on conversion before and after.
20. Brand Store. Live, structured by category, receiving traffic from Sponsored Brands, and not a single page of tiles.
21. Video. Present on the listing and in the Store, and whether it shows the product in use in the first five seconds.

Catalog health: 4 checks

22. Suppressed, stranded and inactive. Anything not live that should be, and why.
23. Account Health. Open policy notifications, listing policy violations and the odds they escalate.
24. Buy Box share. Featured Offer percentage on each main ASIN, and who else is on the listing.
25. Brand Registry status. Enrolled, and using the protections and tools it unlocks.

Inventory: 2 checks

26. Days of cover. Per child ASIN, against true lead time, and whether any ranking ASIN is within six weeks of zero.
27. Excess and aged stock. Units over 120 days of cover, aged-inventory surcharges due, and the IPI score trend.

Unit economics: 3 checks

28. Profit per unit. Built per ASIN from fees, landed cost, returns and promotions, giving break-even ACoS.
29. Return rate and reasons. Against the category, with the reason codes read for expectation gaps.
30. Promotion hygiene. Coupons, deals and Subscribe & Save discounts still running that no longer have a purpose.

From 30 checks to a plan

Each lever is scored 1 to 5 from its checks. The scores are then weighted by the sales behind the ASINs affected, because a weak listing on a product that does 60% of revenue matters more than a weak one on a product that does 2%. The plan starts on the lowest weighted score: the first 30 days go there, the next 30 to the second, and the last 30 to scaling what the first two unlocked. One primary KPI is agreed for the whole plan, and it is the first line of every report thereafter.

What the audit is not

It is not a sales document. A good audit will sometimes conclude that the account is in decent shape and the fix is a routine, not a retainer. It is not a rebuild plan either; most accounts have more that works than the owner thinks, and the plan should say what to leave alone. And it is not free of your input: the unit economics checks need your cost data, and the audit is weaker without them.

A6Audit, diagnose, plan — your account health scorecarda 1–5 score on six levers and a 30/60/90-day plan that starts where it will make the biggest difference.

B12How we audit your advertising accountseven audits that show where money is wasted, which placements convert and how much we expect to save — before we change anything.