AMZTool/Blog/Reporting & Measurement

Amazon Ads Guide · Reporting & Measurement

Amazon Ads Reporting & Measurement: A System You Can Actually Operate

Learn what the figures mean, choose the right report and reconcile totals before deciding what to investigate.

Also available in: 中文版 · Deutsch

William, founder of AMZTool
William
6 min read·Published: November 13, 2026
Share:

Start an Amazon Ads review with the question you need answered: are the ads getting seen, are clicks turning into orders, or is spend out of proportion to sales? Each question calls for different detail. One hypothetical insulated-tumbler example will connect the metrics, report choices and reconciliation below.

Follow the numbers from an impression to a sale

Suppose a group of Sponsored Products campaigns in one marketplace records 20,000 impressions, 200 clicks, $100 in spend, 10 ad-attributed orders and $400 in attributed sales over the same period. These are teaching values, not account results or recommended targets.

Impressions count appearances of an ad; clicks count interactions. Click-through rate is 200 ÷ 20,000 = 1%, and average cost per click (CPC) is $100 ÷ 200 = $0.50. Dividing the report's orders by its clicks gives an example order conversion rate of 5%. Orders and units sold are different measures, so check which column you are using.

Each number describes a different part of the journey. Fewer impressions do not establish a conversion problem, and more clicks do not guarantee more orders. Find where the change occurs before opening more detailed reports.

Which report answers your question?

Report names and availability vary by ad product, account and interface. The table describes common Sponsored Products analysis tasks; use the options actually available in your account.

Your questionReport detail to examineWhat to look for
Which campaigns account for the spend?Campaign reportSpend, clicks and attributed sales by campaign
Which keyword or product target used the money?Targeting reportTargets, match types and their performance
What did shoppers actually search?Search term reportQueries and the targets that matched them
Does performance differ by placement?Placement reportTraffic, cost and conversions at each placement
Which advertised item received traffic?Advertised product reportThe advertised SKU or ASIN, not an assumption that every attributed purchase was that item

A campaign report can identify which campaign drove a spend increase. To see whether a particular query attracted more clicks, continue into the search term report. A targeting keyword is not necessarily the phrase a shopper typed.

What do ACOS, ROAS and TACOS tell you?

With $100 spend and $400 attributed sales, ACOS = spend ÷ ad-attributed sales × 100% = 25%. ROAS = ad-attributed sales ÷ spend = 4. They describe the same relationship in opposite directions. ACOS has no ordinary ratio when sales are zero; ROAS has the same problem when spend is zero. Do not silently turn an undefined ratio into zero in a spreadsheet.

If total sales for the same product scope, period and currency are $1,000, TACOS = ad spend ÷ total sales × 100% = 10%. Total sales must come from matching retail or business data. Renaming the advertising report's attributed sales does not turn them into total sales.

A 25% ACOS does not establish profitability: product, marketplace and fulfillment costs still matter. Amazon's ACOS guide and ROAS guide explain the formulas. When combining campaigns, divide total spend by total sales rather than averaging row-level percentages.

Why do the console and your download disagree?

Align the account, marketplace, ad product, currency, timezone, start and end dates, and filters before comparing totals. A campaign summary and search-term detail should not automatically be treated as having identical coverage.

For example, the console may show the whole campaign while the file is filtered to one ad group. Missing spend is then a scope difference, not evidence of lost data. Remove the filters, compare campaign totals for the same dates, and narrow the selection again.

Generation time also matters. Orders can be attributed after a click, so sales for recent dates may change. Attribution definitions depend on the ad product, account and selected metric; one fixed number of days does not fit every report. Record when the file was generated and reload recent dates using the relevant definition. Do not use reporting delay to explain every unexplained discrepancy.

Turn the reports into a useful investigation

Suppose the next comparable period still has 20,000 impressions and 200 clicks, but only five orders. That points you toward what happened after the click: query mix, offer changes and the product page deserve investigation before a budget increase. If clicks and orders both halve while impressions stay steady, examine click-through rate and traffic composition first.

These are leads, not proven causes. Record which figures changed, which report supplied them and what information is missing. For a working analysis file, follow the Excel export and analysis tutorial. Once the numbers make sense, use the campaign review guide to investigate settings and decide what to change.

William, founder of AMZTool

I've worked in Amazon operations for five years, with more than $10M in sales. I now build practical tools that help sellers reduce repetitive advertising work and avoid manual errors.

Published in:Amazon Ads ReportingPPC MeasurementAnalyticsComplete Guide
Share:
Browse this topicReporting & Measurement · 4 articles