Amazon Search Terms Report: How to Download and Use It

The search terms report is the most useful file in an Amazon Ads account, and in 2026 it moved. Amazon made unified reporting generally available in June and is retiring the report pages most sellers have used for years.

This guide shows where the report lives now, how to build it, what every column means, and the weekly routine we run on the accounts we manage to turn it into harvested keywords, negatives and lower ACOS (advertising cost of sales).

What Is the Amazon Search Terms Report?

The Amazon search terms report is an advertising report that shows the searches shoppers typed before clicking your Sponsored Products or Sponsored Brands ads, with the keyword or target that matched each one and its clicks, spend and sales.

Amazon's report documentation, updated May 18, 2026, sets three facts that shape how you read it:

• It includes only search terms with at least one ad click, so its impressions won't match Campaign Manager.

• Some rows are inferred: terms Amazon matched to your product in non-search placements, such as off-Amazon placements on social sites.

• Alphanumeric rows like "b00ipgvvz4" are ASINs, the product pages your ad appeared on, from auto campaigns and product targeting.

Watch for one naming trap. Brand Analytics in Seller Central has a separate dashboard Amazon also calls the Amazon Search Terms report. That one ranks the most popular searches across the whole store.

This guide is about the advertising report, and we compare the two later.

If the difference between a keyword and a search term is still fuzzy, our guide to Amazon keywords vs search terms explains it with examples.

Where the Search Terms Report Lives in 2026

Two changes this year moved the report. Both matter if you have scheduled reports or a spreadsheet that depends on the old column names.

Unified reporting is the new home. Amazon announced on June 8, 2026 that unified reporting is generally available under Measurement and Reporting > Reporting, ending a beta that began at unBoxed in November 2025. One report can now cover several accounts, ad products and countries.

The old pages close on December 31, 2026. Amazon's migration help page says the Sponsored ads reports page shuts down that day, and every saved report, schedule and historical file still on it is deleted. From December 17, you can't create or edit reports there.

The console has a new name. At unBoxed on September 29, 2026, Amazon renamed the unified ads platform Amazon Ads Agent. The report is still under Reporting in the left navigation; only the platform name changed.

Our advice: rebuild your search term report in unified reporting now, and download anything from the old page you want to keep. The migration tool moves schedules, not history.

How to Download the Search Terms Report

There are four ways to get search term data in 2026. Unified reporting is the main one, and it's the only console path left after December.

1. Open Reporting and Start a Report

In Amazon Ads Agent, click Reporting in the left navigation, then Create report. Amazon's report builder guide walks through each card: filters, columns, report period and delivery.

2. Set the Filters

Add your advertiser account under Accounts. Under Ad product, keep Sponsored Products and Sponsored Brands and untick the rest.

Agencies can add several accounts, or a whole manager account, to one report. That alone saves an afternoon a month.

3. Pick the Dimensions

Unified reporting builds the report from dimensions you tick. For a search term report, choose:

• Time: Date range (one summary row per term) or Date (one row per day).

• Level of detail: Campaign and Ad group.

• Targeting: Targeting, Targeting match type and Search term.

4. Pick the Metrics

Add Impressions, Clicks, Total cost, Purchases, Sales, Units sold and Purchase rate over clicks. Add Sales (promoted) and Sales (halo) if you sell variations or bundles.

We calculate ACOS ourselves from those two columns, ACOS = total cost ÷ sales, so it reads the same in every file.

5. Set the Period and Schedule It

Use the last 30 days for the weekly review and the longest range available for a monthly archive. Then set Report frequency to Weekly so a fresh file lands every Monday.

6. Know the Exceptions

Amazon's unified reporting overview lists what isn't supported yet. For this report, one gap matters: legacy Sponsored Brands campaigns (those without ad groups) aren't in targeting and search term reporting until later in 2026.

Sponsored Products and Sponsored Brands prompt reports aren't in unified reporting either. Pull those from the old page before it closes.

Three Other Ways to See Search Terms

• The old Sponsored ads reports page (until December 17 for new reports): Create report, Sponsored Products or Sponsored Brands, report type "Search term", summary or daily.

• The Search terms tab in an ad group. Good for a quick look, and Amazon lets you copy terms from it into other ad groups as keywords or negatives.

• The Ads API. The v3 search term report pulls up to 31 days per request and keeps 65 days for Sponsored Products and 60 days for Sponsored Brands.

Our Amazon advertising reports guide covers the other report types, such as targeting, placement and purchased product.

Every Column in the Search Terms Report and What to Do With It

Amazon has renamed most metrics. "7 Day Total Sales" is now Sales, "7 Day Total Orders" is Purchases, and "Other SKU" became halo. The old and new names mean the same thing.

Sponsored Products sales attribute over 7 days for sellers and 14 days for vendors. Sponsored Brands uses 14 days. Unified reporting also reports conversions on the traffic date, the day of the click, not the day of the purchase.

The 65-Day Lookback: Why You Need Your Own Archive

The Sponsored Products search term report has a 65-day lookback, per Amazon's report page. The Ads API keeps 60 days for Sponsored Brands. Anything older is gone from that report.

That's why we schedule a weekly download and keep every file. Two months of data can't tell you whether a term converts better in Q4 than in spring, and a year of archived files can.

Unified reporting stores up to 24 months of daily data for most dimensions, so the window may be longer there. Amazon's data availability page doesn't list a separate limit for search terms, so test it in your own account before you stop archiving.

Also ignore the last 7 days when judging a term. Sales keep attributing for a week after the click, so a term with 12 clicks and no orders yesterday may have 2 orders by next Monday.

Inferred Search Terms for Non-Search Placements

Not every "search term" in the current report was typed by a shopper. Amazon's report page now says the column includes terms "inferred as best match to your product in non search context (such as off Amazon placements on social sites)."

The column definition reads: "the search term used by the customer or inferred by Amazon in non-search contexts." Amazon doesn't flag which rows are inferred.

In the accounts we manage, three habits keep inferred rows from causing damage:

• Don't negate an odd-looking term on sight. If it has few clicks and the product is relevant, it may be an inferred match from a placement that still converts.

• Read it with the Targeting column. If an odd term sits under a target that performs well overall, give it the normal click threshold before you negate it.

• Judge on the same thresholds. An inferred term with 40 clicks and no orders is still waste. Negate it exactly as you would a typed one.

A Weekly Search Terms Report Analysis Workflow

This is the routine we run every Monday. It takes about an hour per account by hand, and most of that hour is filtering.

1. Pull the Last 30 Days, Minus the Last 7

Use the weekly scheduled report. Exclude the most recent 7 days from any "no orders" decision, since their sales are still arriving.

2. Sort by Spend

Sort descending and read the top 50 rows first. In most accounts we audit, that's where 70% or more of search term spend sits.

3. Filter Into Four Buckets

Set your rules before you look, or every term starts to look promising. Ours, with a 30% target ACOS as the example:

The threshold comes from your conversion rate. At a 10% conversion rate you expect an order every 10 clicks, so 30 clicks without one is a fair negate signal.

4. Harvest Winners Into Exact

Add each harvest term as an exact keyword in a manual campaign, and add it as negative exact in the auto or broad campaign that found it. Otherwise two campaigns bid on the same search.

The worked example and starting bid formula are in keywords vs search terms. The keyword harvesting funnel guide shows how the campaigns fit together.

5. Negate the Waste

Use negative exact for single terms. Keep negative phrase for words that are wrong in any combination, such as "replacement" if you don't sell parts.

Our negative keywords guide has the threshold math and when to negate at campaign or ad group level. Our match types guide explains why the same keyword reaches different terms by match type.

6. Run N-Grams on the Whole File

Row-by-row review misses words that waste money across dozens of small terms. An n-gram analysis splits every search term into single words (and two- and three-word phrases), then adds up spend and sales per word.

In the example above (illustrative numbers), "kids" appears in 23 search terms. None had more than 12 clicks, so none crossed the threshold alone.

Together they spent $142 for one $24 order: a 592% ACOS. If you don't sell a kids' bottle, one negative phrase on "kids" fixes all 23 rows. Our n-gram analysis guide shows how to build the pivot in a spreadsheet.

7. Log the Changes and Check Them in Two Weeks

Write down what you harvested and negated, and when. A harvested exact keyword can behave differently from the auto row that found it, usually because the bid or placement changed.

Search Terms Report vs Search Query Performance vs Brand Analytics

Amazon has four reports with "search" in the name, and they answer different questions. The search terms report shows only your paid clicks; the Brand Analytics dashboards show the whole market.

The Search Query Performance dashboard has a brand view and an ASIN view. It includes organic and Sponsored Products results from the search results page, and Amazon says its numbers aren't meant to match your ad reports.

The Search Term Impression Share report shows your share of all ad impressions on a term over the last 90 days, and your rank among advertisers.

Here is how we use them together. A term converts at 25% ACOS in the search terms report, but SQP shows a large query volume where your brand earns only a small fraction of the clicks. That term deserves more budget, a higher top-of-search bid and share of voice tracking.

Our guides to Search Query Performance and Brand Analytics cover both dashboards in depth.

Mistakes We See in Real Accounts

• Relying on the old report page. On December 31, 2026, its history is deleted. Download what you need and rebuild schedules in unified reporting.

• Judging the last week. Recent clicks still owe you orders. Terms negated on 3-day-old data often were about to convert.

• Negating at 8 clicks. Small samples throw away good terms. Use a threshold tied to your conversion rate.

• Harvesting without negating at the source. The term then serves from two campaigns, and your exact bid no longer sets the price.

• Only reading rows. Words that waste money across many small terms only show up in an n-gram view.

• Ignoring ASIN rows. Converting product pages belong in product targeting; poor ones go in negative product targeting.

• Comparing SQP clicks with ad clicks. They measure different things, and Amazon advises against comparing them.

A full Amazon PPC audit checks search terms alongside bids, budgets and structure.

How We Run the Search Terms Report in Adbrew

Everything above works by hand: one scheduled report, one spreadsheet, one hour a week per account. It stops working at twenty accounts, when the review is the first job that slips.

Adbrew is an agentic retail media management platform for Amazon and Walmart, and the weekly routine runs inside it on rules you write and can read:

• Harvesting and negation in one rule. Set the lookback, purchase threshold and target ACOS. Adbrew adds the converting term as an exact keyword and the negative exact at the source in the same step.

• Threshold-based negation. Negate relevant terms only after a click threshold you set, or one based on each campaign's own conversion rate, with brand terms excluded.

• Search terms across every campaign. Filter terms account-wide in Sponsored Ads management and act on them with bulk operations.

• Agents for the reading. Ask Adbrew Intelligence for "search terms with over $40 spend and no purchases in the last 30 days, excluding brand terms". The same works from Claude or ChatGPT through the Adbrew MCP server, and nothing applies until you approve it.

Every change lands in Change History with the rule that made it. Brands on Adbrew cut ACOS by 46% on average, and Seller Presto saves more than 250 hours of manual optimization a month.

Make the Report a Weekly Habit

The search terms report hasn't lost any of its value in 2026. It has moved, it renamed its columns, and it now includes terms nobody typed.

Set up the unified report this week, archive it, and run the routine every Monday. Want harvesting and negation running on rules you control? Book a demo and we'll walk through your search terms report with you.