How to find wasted Amazon ad spend in 15 minutes

You can find useful leads in an Amazon ad account in fifteen minutes. You cannot prove that every click without a sale was wasted. The aim is to leave with a short list of changes worth checking, each tied to a product and a dollar amount.

This walkthrough focuses on Sponsored Products. Keep other ad formats separate because their reports and attribution can differ. The examples below are invented arithmetic examples, not customer results. Use your own fees, costs and account data before making changes.

Minutes 0–3: set the date and the margin

Open the correct advertising account and marketplace. Record the account, currency, campaign type and date range in a review sheet. Use a period long enough to include ordinary buying behavior. Avoid judging a campaign only on today's incomplete data.

Check the attribution window and reporting notes for the report you download. A click near the end of a period can produce an attributed order later. Recent rows may still change. Label them provisional rather than treating a blank sales column as final evidence of failure.

Next, choose one advertised SKU. Work out the amount left from a sale before advertising:

Contribution before ads = net product revenue − product cost − selling fees − fulfillment − other variable costs.

Include a reasonable allowance for returns and storage where they matter. If an input is missing, write “unknown” and flag the product for cost review. Do not substitute zero.

For an example product with $30 of net revenue and $21 of non-ad variable costs, contribution before ads is $9. The break-even advertising cost of sales, or ACOS, is $9 ÷ $30 = 30%. That leaves nothing from an advertised sale toward fixed expenses at break-even.

Minutes 3–6: download the right report

Get the Sponsored Products search term report from the reporting area of the advertising console. Menu labels can change. Look for the report type rather than relying on an old screenshot.

Keep campaign, ad group, targeting, match type, customer search term, clicks, spend, attributed orders and attributed sales. Include advertised product identifiers when available. If the report combines several products, use the matching product report to understand which margins apply.

Amazon explains the difference between the phrase you bid on and the shopper query that triggered an ad in its Sponsored Products targeting guide. That distinction matters: a broad keyword can produce both useful and irrelevant queries.

Save an untouched copy of the export. Add calculations in a working copy. Sort by spend, highest first, so the largest decisions get your attention first. Do not start by deleting every row with zero sales; those rows are one of the things you need to inspect.

Minutes 6–10: find three kinds of concern

Start with queries that clearly describe a different product. A seller of replacement bottle caps might appear for someone seeking complete bottles. Read the search term and check the advertised listing before calling it irrelevant.

Next, find spend without attributed orders in mature rows. Compare spend with the product's contribution before ads. Spending $18 without an order on the example product has used the contribution of two ordinary sales. That is a reason to inspect the query, not a universal rule to block it.

Finally, look at terms that do produce orders but leave too little contribution. Suppose the example SKU produced $150 of attributed sales from five units and cost $60 in ads. ACOS is 40%. Estimated contribution after ads is 5 × $9 − $60 = −$15, before fixed costs.

That calculation assumes all five units have the same revenue and cost. If a report includes other purchased products, bundles or multiple units, use the actual mix. A tidy ACOS percentage does not fix a mismatched margin.

Keep brand searches and discovery traffic separate in your review. They can serve different purposes. A cheap brand campaign does not establish that new-customer acquisition is profitable. An expensive discovery term also does not establish that all of its sales were incremental.

Minutes 10–13: choose a small action

Write one proposed action for each serious concern. Options include reviewing a bid, adding a carefully chosen negative target, fixing a listing mismatch, or waiting for attribution to mature. A campaign with sparse evidence may need a bounded observation period instead of an immediate cut.

Use the narrowest change that addresses the problem. Negative phrase targeting can block more searches than negative exact targeting. Product targeting needs product negatives, where supported, rather than assuming a text keyword will cover it. Amazon's guide explains the available targeting approaches; confirm the options in your account.

Check existing negatives and account rules before adding anything. Protect terms you have decided must remain available. A useful review log includes the query, product, evidence, proposed change, old value, new value and reason.

Do not raise a campaign's budget simply because it reaches its limit. First establish that the marginal sales can afford more spend and that stock is available. Conversely, a low-spend campaign might be limited by poor relevance rather than by a budget that needs increasing.

Minutes 13–15: save a review point

Record what you approved and when. Choose a review date that gives the campaign enough time to collect relevant data and complete its attribution window. Compare the same product, marketplace and time basis after the change.

Watch both spend and sales. Cutting spend by $50 is not a $50 profit improvement if contribution from lost orders fell by $80. Check total product sales alongside attributed advertising sales, while recognizing that a simple before-and-after comparison does not prove causation.

Track other changes made during the period. A price cut, listing edit, stockout or seasonal event can explain a result that otherwise looks like a bidding win. If the account is large enough, a comparable unchanged group can give you a better reference.

What you should have at the end

Finish with a handful of documented concerns, a cost-based limit for the products reviewed, and a clear owner for each action. Keep uncertain rows marked as uncertain. Repeat the review on a regular schedule so it becomes a habit rather than an occasional account cleanup.

If you do not know your product margins yet, begin with our profit per sale guide. Better costs often improve an ad decision more than another report filter.

Want a second set of eyes on your own numbers? Get a free BrimScout profit audit. We will start with the evidence and the changes worth discussing.