What Brands Want to See in a Weekly Amazon Ads Report

There is a specific silence that follows a weekly Amazon ads report. Fourteen tabs of data land on Monday morning and nobody says anything. On Thursday, someone asks in a separate email why sales were down last week.

The answer was in the report. It just was not findable.

Whether you build these reports yourself or receive one from an agency every Monday, the same standard applies. This is what a weekly report owes you, and what to push back on when it does not deliver.

After nine years of running Amazon accounts and inheriting them from other agencies, we have read a lot of weekly reporting. The pattern is consistent: most reports are built to prove work happened, not to help someone decide anything. They are defensive documents dressed up as informative ones.

A weekly Amazon ads report has one job. It should let a busy owner understand their business in under five minutes and know what happens next. Everything that does not serve that job is decoration.

Why Most Weekly Reports Fail

The failure is rarely missing data. It is almost always too much data, arranged without a point of view.

Three habits cause most of the damage:

• Exporting instead of reporting. A campaign-level dump is raw material, not analysis. If the reader has to sort it themselves, you have handed them homework.

• Reporting metrics instead of movements. ACOS was 24% last week, but was that up, down or flat, and does it matter at the account's current stage?

• Leaving out the human. The most valuable line in any report is what you concluded and why, and that is the part software cannot generate for you.

There is also a quieter failure. Many reports cover only advertising, while the brand owner experiences one business.

If sales dropped because a hero ASIN went out of stock, an ads-only report shows rising ACOS and falling revenue with no visible cause. The reader is left to guess.

The Five Questions a Weekly Report Must Answer

Build every report backward from these five questions. If a section does not serve one of them, cut it.

1. Did We Make Money This Week?

Start with the numbers that reflect the whole business, not just the ad account: total sales from Business Reports, ad sales, ad spend and TACoS. Put last week's figures beside them, along with the four-week average.

TACoS matters more than ACOS here because it is the view that shows whether advertising is buying growth or subsidizing it. An account with a healthy 18% ACOS and a climbing TACoS is quietly getting worse.

Give this section three lines of plain English. Instead of "TACoS increased 2.1 points," write: "TACoS rose from 9.4% to 11.5% because organic sales fell while ad spend held flat. That points at ranking, not at the ads."

2. What Changed, and Why?

This is the section brands actually read, and the one most reports skip.

Pick the two or three largest movements of the week and explain each in two sentences. Not every fluctuation deserves a paragraph: a 4% swing on a mature campaign is noise, while a 40% drop in impressions on your top ASIN is a story.

When you cannot explain a movement yet, say so. "Impressions on B0XXXXXXXX fell 31% starting Wednesday. We are checking whether this lines up with the Featured Offer (Buy Box) loss we saw on Tuesday." Honest uncertainty builds far more trust than confident nonsense.

3. Where Is the Money Leaking?

Every account wastes some spend. That is normal and not an admission of failure; hiding it is the failure.

In our own account audits, we consistently find a meaningful share of monthly ad spend going to search terms with no conversions in the trailing sixty days. The exact figure varies by category and account maturity, but it is almost never zero, even in well-managed accounts.

Show the waste plainly:

• Spend on zero-conversion terms from the search term report.

• Spend on targets running above your break-even ACOS.

• Budget lost to campaigns capping out before the day ends, which Amazon's own budget guide treats as missed sales.

Then show what you did about each one.

4. What Did We Actually Do?

List the changes made, with the reasoning attached: bids adjusted on which campaigns and why, negative keywords added, new campaigns launched, dayparting rules changed.

Keep it to what materially affects performance, because nobody needs a log of every automated bid adjustment, so if rules-based automation is running, summarize what the rules did rather than listing each action.

If you are assembling this yourself, the Amazon advertising reports available to you are worth knowing, because most of this section can be sourced directly from them. Amazon's unified reporting became generally available in June 2026, and the legacy Sponsored Ads and DSP report pages retire on December 31, 2026, so rebuild any scheduled exports there now.

5. What Happens Next Week?

State two or three priorities as intentions with expected outcomes. For example: "We are pulling budget from the display ads retargeting campaign and moving it into top-of-search placement on the two converting keywords. We expect ad sales to hold and TACoS to improve by roughly a point."

This is what turns a report into a working document. It also gives you something concrete to measure yourself against seven days later, which is uncomfortable and exactly why it works.

Metrics You Can Safely Cut

Some numbers earn their place every week, and many do not. If you are trying to get a report down to something readable, these are usually the first to go:

• Impressions in isolation. Useful for diagnosing a specific drop, but meaningless as a standing weekly headline.

• Click-through rate without context. CTR only means something against your own recent baseline and your category, which Amazon's benchmarks reporting now shows.

• Every campaign, every week. Report on the campaigns that moved, and list the rest in an appendix if the client wants it.

• Placement breakdowns by default. Bring these in when you are making a placement decision, not as furniture.

• Vanity comparisons. Year-over-year on an account that changed strategy in March is not a comparison; it is two unrelated numbers next to each other.

Year-over-year needs extra care in 2026. Amazon changed how it credits views for some ad types on January 1, 2026, so compare those lines on the "all views" metrics rather than the standard columns.

The goal is not a shorter report. It is a report where every number present is one the reader would have asked for. Our guide to Amazon PPC metrics covers what each one tells you.

Format and Cadence

Weekly beats daily for reporting, even when your data updates hourly. Daily reporting invites reactions to noise, while weekly gives enough sample size to see a trend without letting a real problem sit for a month.

Keep the structure identical every week: the same order, the same sections and the same place for the summary. The reader learns where to look, and after a month they can scan it in ninety seconds. Redesigning the template every quarter destroys that.

Put prose first, not a chart. Open with three to five sentences covering the state of the account, the one thing that changed most and the one decision that needs making. Charts go below for anyone who wants them.

The report should arrive on the same day each week, before anyone has to ask for it. A report you have to chase is not a reporting system.

Automate the Assembly, Not the Judgment

Pulling data, calculating deltas and formatting tables should not consume your Monday. That work is mechanical and should be handled mechanically, which is where AI in Amazon advertising is genuinely useful today rather than aspirational.

What does not automate well is the part your client is actually paying for: deciding which two movements matter out of forty, connecting an impression drop to a listing change nobody flagged, or knowing that a particular ACOS spike is seasonal because you watched it happen last year.

The best reporting setups we have seen use software for everything up to the conclusion, then a human writes the conclusion. Fully automated reports end up technically accurate and unread, which is the same outcome as no report at all.

What Good Looks Like

One page, five sections, and a plain-English summary at the top you could forward to a business partner without explaining it.

The test is simple. If the report leaves you with an opinion about next week's plan, it worked. If your only response is "thanks," it did not.

That gap is where most agency relationships quietly go wrong: not through bad performance, but through months of reports that were technically complete and practically useless. Clear reporting will not fix a struggling account, but it will make sure you find out early enough to do something about it.