Amazon Advertising Reports: Which Ones Actually Matter

Dana had 47 spreadsheets in a folder called "PPC Data."

Every Monday for eleven months she had logged into the ad console, requested the same four reports, downloaded them, renamed them by date, and dropped them in the folder. Every Monday she opened the newest one, scrolled for about four minutes, felt vaguely uneasy, and closed it.

In eleven months she had changed exactly two bids because of something she read in a report.

That is the real state of Amazon advertising reports for most sellers. Not "I don't have the data." Everybody has the data. The problem is that Amazon hands you nine different reports with overlapping columns, no guidance on which one answers which question, and no hint about which numbers are worth acting on. So the reports get downloaded and never used.

This guide fixes that. Here is what every Amazon advertising report actually contains, which five carry almost all of the decision-making value, the workflow that turns a download into a change, and the traps that make sellers act on numbers that were never saying what they thought.

Where Amazon Advertising Reports Live (And What They Are)

Amazon advertising reports are downloadable exports of your ad performance, generated on demand from the Measurement and Reporting → Sponsored ads reports section of the advertising console. They are separate from the numbers you see on screen in Campaign Manager.

That separation matters more than it sounds. Campaign Manager shows you performance at the level you are currently looking at: campaign, ad group, or target. The reports show you dimensions Campaign Manager cannot display at all — the actual customer search terms behind your broad keywords, which specific ASINs people bought after clicking, which placement your money went to. Those dimensions are where the money is.

When you request a report you choose four things:

• Report category: Sponsored Products, Sponsored Brands, or Sponsored Display. Each ad type has its own report set, and they are not interchangeable.
• Report type: the nine options covered below.
• Time unit: Summary (one row per item for the whole period) or Daily (one row per item per day). Summary for decisions, Daily for trend analysis.
• Report period: a date range, generally reaching back around 60 days in the console.

A practical consequence of that last point: your reporting history is not permanent. Amazon keeps generated reports available for a limited window, and the requestable range only goes back so far. If you want a year-over-year view of your own account, you have to build the archive yourself. Set a calendar reminder for the first of every month, pull a summary report for the previous month, and store it. Sellers who did this two years ago can answer questions today that nobody else can.

Every Amazon Advertising Report, and the Question It Answers

Nine report types, each with one job. Here is the honest version of what each is for.

Search Term Report

The most valuable report Amazon gives you. It shows the actual phrases customers typed before clicking your ad, mapped to the keyword or target that matched them.

This is the only place you can see the gap between what you bid on and what you actually bought. You bid on "stainless steel water bottle." You paid for "water bottle for toddlers," "bottle for hot soup," and "steel water bottle 2 litre." Two of those will never convert for your product, and you are paying for both.

Everything downstream flows from here: negative keywords, new exact-match keywords harvested from converting terms, match type decisions. We have a full workflow for this one in the Amazon search term report guide. If you read only one report, read this.

Targeting Report

Performance of the things you chose to bid on: keywords, product targets, category targets, audiences. Clicks, spend, sales, ACoS, orders per target.

Where the search term report tells you what you caught, the targeting report tells you how the net is performing. It is the right report for bid decisions, because it is scoped to entities you can actually change. Sort by spend descending, then look at the ones with zero orders. That list is your bid-cut queue.

Advertised Product Report

Performance by ASIN or SKU (the product being advertised, not the keyword). Includes ad spend, ad sales, and units.

This is the report that answers "which products are my ads actually working for?" Most accounts are wildly uneven here. One or two ASINs earn their spend, a middle group breaks even, and a tail of products quietly burns budget because they were added to a campaign eighteen months ago and never reviewed.

It is also where you get the raw numbers for TACoS by product, once you pair it with total sales from Seller Central. If you have not done that comparison, start with our TACoS guide. Ad sales alone will mislead you on which products are genuinely profitable.

Purchased Product Report

The one almost nobody opens, and one of the most interesting.

It shows the ASINs customers actually bought after clicking your ad, including products that were not the advertised one. Someone clicks your ad for the 12-ounce bottle, browses, and buys the 24-ounce. That sale shows up here.

Two uses. First, it reveals halo sales: campaigns that look mediocre on their own ASIN may be driving real revenue to siblings in the same variation family. Second, it exposes cannibalization patterns: if your ads for Product A mostly sell Product B, you are paying to advertise the wrong product.

Placement Report

Spend and performance split by where the ad appeared: Top of Search (first page), Rest of Search, and Product Pages.

Top of Search almost always costs more per click and almost always converts better. The question is by how much, in your account, on your products. That ratio is the entire basis for setting placement bid multipliers. Guessing at those multipliers is one of the most common ways to overspend. The placements guide walks through how to calculate them from this report instead of guessing.

Campaign Report

Top-level performance per campaign. Useful as a scoreboard, close to useless as a diagnostic: it tells you a campaign is underperforming without telling you why.

Treat it as the entry point: campaign report to find the problem, targeting and search term reports to find the cause.

Performance Over Time

Clicks, spend, CPC, and impressions plotted by day. Not for optimization. For context.

When your ACoS jumps in a week, this report tells you whether CPCs rose across the account (an auction-level change: new competitors, seasonal pressure) or whether one campaign changed behavior. That is a genuinely different diagnosis and it leads to a genuinely different response.

Budget Report

Shows how much of your daily budget each campaign consumed, and how often campaigns ran out of budget.

Every hour a profitable campaign spends "out of budget" is inventory you paid to earn and then handed to a competitor. Filter for campaigns hitting 100% of budget with an ACoS below your target. Those are the ones to raise, immediately. If you use scheduled increases, pair this with the budget rules guide.

Gross and Invalid Traffic Report

Total traffic vs. the portion Amazon classified as invalid (bots, click fraud, internal traffic) and removed before billing you.

You do not act on this report; you check it when something looks wrong. If a campaign shows a spike in clicks with no matching sales, this tells you whether the traffic was real. Amazon filters invalid traffic automatically and you are not charged for it, so the report is a sanity check, not a to-do list.

Sponsored Brands sellers also get a Search Term Impression Share report, which shows the share of impressions you won for a given search term against the total available. It is the closest thing Amazon offers to a "how much of this keyword do I own?" number, and it is the right way to size the opportunity on head terms before you raise bids on them.

Which Reports Actually Matter: The Short Table

Five reports (search term, targeting, advertised product, placement, and budget) carry roughly 90% of the value. If a report is not on your calendar with a decision attached to it, you are collecting spreadsheets, not managing ads.

The Weekly Workflow That Takes 30 Minutes

Tom sells kitchen equipment and spends about $9,000 a month on ads. His entire routine is three reports and half an hour, every Monday morning.

Minutes 1–10: Search Term Report (last 14 days, summary). Sort by spend descending. Scan for terms with meaningful clicks and zero orders. His threshold is 10+ clicks and no sale. Those become negative exact keywords. Then he flips the sort: terms with 2+ orders that are not already exact-match keywords get harvested into his exact campaign at a bid derived from their current CPC. Two lists, one pass. The negative keywords guide covers where to apply each negative so you don't accidentally block a whole ad group.

Minutes 11–22: Targeting Report (last 30 days, summary). Anything above his ACoS target with 15+ clicks gets a bid cut proportional to how far off it is, not a blanket 20%. Anything below target with impressions available gets a raise. Anything with high spend and zero conversions after 30 days gets paused, not cut.

Minutes 23–27: Budget Report (last 7 days). Any campaign that hit its cap while running under target ACoS gets more money. That's it.

Minutes 28–30: Bulk file. He applies every change in one upload instead of clicking through the console. If you are still making changes one at a time, bulk operations will give you back more hours than any other single skill in Amazon PPC.

Monthly, he adds the advertised product and placement reports and does a deeper pass. Quarterly, the purchased product report. That is the whole system.

Notice what the routine has that Dana's eleven months of downloads did not: every report is attached to a specific decision made with a specific threshold. The report is not the work. The decision is the work.

Why Your Numbers Don't Match (And Which One Is Right)

The single most common report question sellers ask: why does the report say something different from Campaign Manager, and why did last week's number change since I looked at it?

Attribution lag. Amazon attributes a sale back to the day of the click, not the day of the purchase. Sponsored Products uses a 7-day attribution window; Sponsored Brands and Sponsored Display use 14 days. So a click on Monday and a purchase the following Sunday both land on Monday's row, so Monday's row keeps growing for a week or two after Monday.

The practical rule: never judge the last 48 hours. They will look worse than they are, always, and sellers who cut bids based on yesterday's ACoS are systematically cutting winners. Wait a week before you trust a number.

Different totals across reports. The purchased product report will not tie out to the advertised product report, because it counts different things (products bought vs. products advertised). Sponsored Display reports include view-attributed conversions that Sponsored Products reports do not. These are not errors; they are different questions.

Same-day data is provisional. Numbers restate as Amazon processes traffic and removes invalid clicks. Same-day reporting is directional at best.

If two reports disagree, the answer is usually that you are comparing two different attribution models, not that one is broken.

Five Ways Sellers Misread These Reports

1. Judging targets on ACoS alone. A keyword at 45% ACoS on a product with 55% margin is profitable. A keyword at 22% on a 20%-margin product is losing money. Your ACoS target is a per-product number, not an account number. Our ACoS guide has the math.
2. Killing search terms too early. Four clicks and no order is not evidence of anything. Statistically, you need to see roughly 2–3× your typical clicks-per-conversion before "no orders" means "will never convert."
3. Reading auto campaigns as a performance problem. Auto campaigns are supposed to have a messy search term report. That is their function. They are a discovery engine. Judge them on the quality of terms discovered, not on ACoS. Our auto vs. manual guide explains the split.
4. Ignoring impression data. Sellers obsess over ACoS and never look at impressions. A target at 12% ACoS with 300 impressions is not a success story. It is a bid that is too low to matter. Low ACoS plus low impressions means raise the bid, which is the opposite of what most people do.
5. Downloading Daily when you want Summary. A 30-day daily export of 400 targets is 12,000 rows of noise. Summary aggregates it for you. Use Daily only when the shape over time is the question.

Frequently Asked Questions

How far back do Amazon advertising reports go?

The console generally lets you request roughly the last 60 days, and generated reports stay available for a limited window after that. For longer history, export monthly and keep your own archive. Amazon will not do it for you.

Which report shows customer search terms?

The Search Term Report. Campaign Manager and the targeting report show only the keywords and targets you created, never the raw customer queries behind them.

Why is my ACoS different in the report than in Campaign Manager?

Almost always attribution timing. Reports and console views can be pulled at different moments, and sales keep attributing back to the click date for 7 days (Sponsored Products) or 14 days (Sponsored Brands and Display).

Can I automate report downloads?

Yes, through the Amazon Ads API, which is how third-party tools pull data. In the console itself you can schedule recurring reports, but you still have to open and read them.

How often should I pull Amazon advertising reports?

Weekly for search term, targeting, and budget. Monthly for advertised product and placement. Quarterly for purchased product. More often than that and you are reacting to attribution noise.

The Honest Limit of Reports

Reports are a photograph. You look at Monday's picture, make decisions, and then the account runs unattended for six and a half days while conditions change: competitors adjust bids, a rival stocks out, a term that converted all month goes cold on Thursday.

That gap is not a discipline problem. Nobody is pulling the search term report at 11 p.m. on a Wednesday. Even sellers with a genuinely good routine are making a week's worth of decisions from a single snapshot, and reviewing maybe the top 50 rows of a report that has 4,000.

A quick way to find out how much that costs you: run a PPC audit and count the changes you would have made on days you weren't looking.

Dana, from the top of this article, deleted 45 of her 47 spreadsheets. She kept the workflow and gave up the archive. Her spend did not change much. Her ACoS dropped 6 points in two months, entirely from negatives she had never applied and bids she had never raised. The data had been in the folder the whole time.

Reports tell you what happened. Something still has to decide what to do about it, every day, on every target. Not once a week on the rows you had time for.