Amazon Ads has published a set of recommended actions tied to its peer benchmarks, spelling out what advertisers should consider when their results on eight metrics sit above or below those of comparable brands. The guidance appears on an Amazon Ads Support Center page titled "Understanding benchmark metrics and recommended actions," which carries the line "Updated on Nov 11, 2026." The page does not say what was changed on that date.

The page covers eight metrics, each with one set of actions for performance above benchmark and one for performance below it: 16 recommended actions in all. On PPC Land's count, every one of the eight favorable positions leads to a suggestion to scale campaigns, raise investment, increase budget or bid higher. None of the 16 suggests reducing a budget or pausing a campaign. In two rows, a cost that is already lower than the peer benchmark comes with a suggestion to "test increasing bids."

The same page explains how benchmark percentiles are labeled, where the 25th and 75th percentile values can be found, and that the figures rest on "ad-spend weighted calculations." That last detail sits uneasily with how the page describes the median.

In Short

Amazon now tells advertisers what to do when their ad results are better or worse than similar brands, across eight metrics and 18 countries. It matters because the advice shapes how brands read Amazon's own benchmark numbers, and every good result is paired with a suggestion to spend or bid more, while no result is paired with a budget cut. The median in these benchmarks is weighted by ad spend, so "average" leans toward the biggest spenders in a category rather than the typical brand.

Who gets benchmarks, and where

According to the page, benchmarks are available in Australia, Brazil, Canada, Egypt, France, Germany, India, Italy, Japan, Mexico, the Netherlands, Saudi Arabia, Spain, Sweden, Turkey, the United Arab Emirates, the United Kingdom and the United States. That is the same list of 18 countries Amazon gave when benchmarks reached general availability on May 18, 2026, ending a beta that had been limited to the US.

"This feature is available to brand owners and representatives for both sponsored ads and Display, Video, and Audio campaigns," the page states. The May announcement drew the eligible group a little wider: brand owners and representatives registered in Brand Registry, plus endemic advertisers in 15 of the 18 markets, with Saudi Arabia, the UAE and Egypt left out of that second group. The November page does not mention endemic advertisers. It does not say whether their access has changed.

The product list has moved too. In May, Amazon listed Amazon DSP, Sponsored Products, Sponsored Brands, Sponsored Display and Sponsored TV campaigns. The November page says benchmark metrics in unified reporting "are available for Sponsored Products, Sponsored Brands, Streaming TV, and DVA+." Sponsored Display does not appear in that sentence. Nor does the page say whether Streaming TV is the same product that was called Sponsored TV in May, or whether Sponsored Display benchmarks remain available through another route such as the Reporting API.

DVA+ is itself recent. PPC Land reported on September 29 that Amazon had put a four-step DVA+ campaign builder into an extended closed beta in the US, with an open beta expected in late October 2026 and general availability in 35 markets targeted for the first quarter of 2027. Around the same date, Amazon retired the Amazon DSP name in favor of DVA and renamed its platform Amazon Ads Agent.

How the percentiles are read

Amazon divides the metrics into two groups and displays them in opposite directions.

Rate-based metrics are shown from lower to higher values, "as higher percentages typically indicate better performance." The page names click-through rate, video completion rate, purchase rate, detail page view rate and return on ad spend (ROAS) as examples. For these, the bottom 25% "Shows baseline performance level" and the top 25% "Shows strong performance level," with the median in between.

Cost-based metrics - CPC, CPM, cost per purchase and cost per completed view for video ads - run the other way, "as lower costs typically indicate better performance." Here the labels follow the value rather than the quality of the result. "Top 25%: Shows higher-cost level," the page states, while "Bottom 25%: Shows lower-cost level." A brand in the top 25% for CPC is therefore among the most expensive in its peer group, the opposite of what the same label means for click-through rate.

Figures without a percentile label are the advertiser's own. "The values shown without percentile labels (p25, p50, or p75) represent how your brand is performing based on campaigns in your advertising account," according to the page.

Where the numbers appear depends on the tool. "In the campaign manager insight card, you'll see the median benchmark to provide a quick performance reference point," the page states. The 25th and 75th percentile values sit in unified reporting and the Reporting API. In unified reporting, each benchmark metric can be added at three levels, which the page illustrates as "CTR (p25), CTR (p50), and CTR (p75)."

Benchmark dimensions can be added alongside, "such as Benchmark brand name, Benchmark browse category path, and Benchmark campaign goal," which Amazon says show "the peer group for each comparison." The page does not specify whether the brand name dimension refers to the advertiser's own brand or to members of the peer set. Amazon's benchmarks white paper states that no single advertiser's performance is identifiable.

The core of the page is a table. For each of eight metrics, Amazon gives a "Planning" action and an "Optimization" action for performance above benchmark, and the same pair for performance below it. The eight are completion rate (video ad), cost per completed view (video ad), cost per purchase (new to brand), CPC, CPM, CTR, percent of purchases new to brand, and purchase rate (new to brand). These are the eight metrics Amazon listed at general availability in May.

Amazon frames the table with a caveat. "These suggestions serve as general guidance to help inform your strategy, but are not exhaustive," the page states. "Your specific business goals, market conditions, and overall advertising strategy should guide how you act on benchmark insights."

When results are better than peers

A favorable result means above benchmark for the four rate metrics and below it for the four cost metrics. Across those eight cells, the actions point one way.

A completion rate above benchmark leads to "Increase video ad investment." A click-through rate above benchmark leads to "Scale successful campaigns, test increasing bids." A share of new-to-brand purchases above benchmark leads to "Increase budget allocation to prospecting campaigns targeting new customers" and "Expand targeting to similar audiences, increase bids on high-performing new audience segments." A new-to-brand purchase rate above benchmark leads to "Scale budget for successful new customer acquisition campaigns."

A cost per completed view below benchmark leads to "Scale successful video campaigns." A new-to-brand cost per purchase below benchmark leads to "Consider scaling successful campaigns targeting new customers."

CPC and CPM go a step further. When CPC is below benchmark, the optimization action is "Test increasing bids to capture more premium placements." When CPM is below benchmark, it is "Test increasing bids to access premium inventory while maintaining efficiency." Put plainly, an advertiser already paying less than its peers per click or per thousand impressions is pointed toward bidding up, a move that, if it raises costs, would bring those costs closer to the peer level the brand was beating.

When results trail peers

The eight unfavorable cells are framed as reviews and reallocations rather than reductions. A completion rate below benchmark leads to "Test shorter video lengths, improve opening sequences." A click-through rate below benchmark leads to "Review ad spend distribution" and "Test new ad copies, improve keyword relevancy." A low share of new-to-brand purchases leads to "Balance budget between prospecting and retargeting campaigns." A low new-to-brand purchase rate leads to "Reallocate budget to better-performing products/campaigns" and "Review pricing strategy, enhance product content, adjust bidding strategy."

On the cost side, a CPM above benchmark leads to "Review inventory and audience strategy" and a cost per completed view above benchmark to "Review video campaign strategy and placement mix." The closest the table comes to spending less is the new-to-brand cost per purchase row, where a cost above benchmark leads to "Refine audience targeting, adjust bidding strategy to reduce costs while maintaining new customer acquisition." Even there, the goal stated is lower unit cost with acquisition held steady, not a lower budget. No cell mentions pausing a campaign or lowering a bid by name, and "reallocate" and "balance" move money between campaigns rather than taking it out.

Terms the page leaves open

The CPC row, for results above benchmark, recommends "Adjust bidding strategies, improve quality score." The page does not define quality score or say where an advertiser would see one. The term is most often associated with keyword ratings in search advertising. Amazon did add a content quality rating to A+ Content Manager in February 2026, but that rating grades product detail page content as "Needs Improvement" or "Meets Standards" and is not described as part of any ad auction. Which of these, if either, the table means is not stated.

The table also applies one set of actions across products with different mechanics. "Improve keyword relevancy" fits Sponsored Products and Sponsored Brands, which run on keywords, but the CTR row does not distinguish them from DVA+ display or Streaming TV campaigns, which the page also lists as covered by benchmarks.

What "median" means in a spend-weighted benchmark

The page's methodology section is two short paragraphs. "Benchmarks use ad-spend weighted calculations where each peer brand's influence reflects their market activity, with balancing controls to prevent any single brand from skewing results," it states. "This approach accounts for different brand types and market positioning to ensure that larger brands aren't disproportionately affected by smaller competitors in the same category."

That second sentence names whose interests the weighting protects: larger brands, so their benchmark is not pulled around by smaller competitors. The flip side is that a smaller brand's peer benchmark is shaped mostly by the biggest spenders in its peer set.

Amazon's white paper on the subject, "The New Standard: How Benchmarks Are Transforming Performance Evaluation," which the company promoted in a post dated March 27, 2026, sets out the mechanics in more detail. According to the paper, peer groups, which Amazon calls PeerSets, are built with a hybrid of K-Nearest Neighbor matching, based on brand-level revenue and average sales price within a category, and a large language model that assesses competition, market segment overlap and customer base similarity. Campaigns must meet operational status, minimum activity and attribute thresholds, for which the paper gives no numbers. Metrics are aggregated per brand by summing across valid campaigns, then combined into spend-weighted percentiles in which each brand's influence is proportional to its ad spend.

The "balancing controls" on the help page correspond, in the white paper, to a cap: no single brand can contribute more than 49% of the total weight, with the excess redistributed to the remaining peers. At least five peer brands are required, according to the paper, or no benchmark is shown. The paper gives no lookback window or refresh cadence.

A 49% cap still lets one brand account for almost half of a benchmark built from five or more peers. The median, too, is spend-weighted: the white paper works through a three-brand example in which an unweighted median click-through rate would be 10% while the spend-weighted median is 3.00%.

What, then, does the median on the help page describe? The page itself calls it "the middle performance level where half of peer brands perform above and half perform below." Under the white paper's method, that is not what the figure measures. A spend-weighted median splits peer spend, not peer brands, into two halves. In a peer set where one or two large advertisers carry most of the spend, most brands could sit on the same side of the "median." The page does not reconcile its plain-language description with its own statement that calculations are weighted by ad spend.

Metrics beyond the eight

The help page's list of rate-based examples includes ROAS, purchase rate and detail page view rate. None of these was among the eight metrics Amazon named in May, and the white paper describes support for eight metrics - the same eight that appear in the recommended actions table. The page does not say whether ROAS or detail page view rate benchmarks now exist. It directs readers to a separate "Benchmark reporting" page "for the full list of metrics with benchmarks and their definitions."

The three new-to-brand metrics among the eight are older than the benchmarks themselves. Amazon launched new-to-brand metrics in January 2019, counting a purchase as new to brand when the shopper had not bought from that brand on Amazon over the prior year.

A benchmark report that cannot yet migrate

The timing of the update matters because Amazon is retiring its old reporting pages. When unified reporting reached general availability on June 8, 2026, Amazon set the Sponsored Ads reports page and the Amazon DSP reports page to be sunset on December 31, 2026.

A separate Amazon Support Center page on migrating scheduled reports, updated on November 12, 2026, moved the shutdown of both legacy pages to January 31, 2027, with the old pages becoming read-only on December 17, 2026 and remaining saved reports, schedules and historical data deleted at shutdown. That page lists the benchmark report among ten report types that are not yet available in unified reporting and cannot be migrated.

The two pages describe different things. The November 11 page says benchmark metrics can be added as columns to a unified reporting report at p25, p50 and p75. The November 12 page says the legacy benchmark report type has no equivalent in unified reporting yet. A scheduled benchmark report left on the old pages will be deleted on January 31, 2027. The closest substitute described on Amazon's pages is a unified reporting report built with benchmark columns, and neither page explains how closely that matches the old report.

A CPM benchmark across a counting change

One more change will run straight through two of the eight metrics. From January 1, 2027, Amazon will count DVA display impressions on its own store only once an ad appears on a shopper's screen, and expects impression counts on those placements to be about 63% lower on average. On PPC Land's earlier calculation, a campaign whose spend and clicks stayed the same would report a CPM 2.7 times higher, with click-through rate rising by the same factor.

If peer benchmarks are computed over the same periods as an advertiser's own figures, as the daily, weekly and monthly views suggest, peers face the same change. But the effect depends on how much of each brand's activity sits in Store display placements, which the change covers, versus video, audio, third-party inventory or sponsored ads, which it does not. A brand's position against the CPM or CTR benchmark could shift in January without any change in what shoppers saw. Under Amazon's table, a brand that crosses the benchmark line would also face a different recommended action.

Why the bidding advice draws attention

Amazon's suggestion to "test increasing bids" when costs are below peers lands while the pricing of its auctions is in court. On August 31, 2026, the Federal Trade Commission and 22 states sued Amazon, and the complaint alleges that the share of charged clicks priced at the advertiser's own full bid reached 79.1% in 2024. Amazon rejects the case and says no advertiser paid more than its submitted bid. If the complaint's figure is accurate and that rate still holds, a higher bid would, on most charged clicks, translate directly into a higher cost per click.

Amazon's own bidding education has also stopped short of evidence. An Amazon Ad Lab video in October compared fixed and dynamic bidding for Sponsored Products and named four metrics to watch without publishing any test results. The benchmark page likewise gives direction without outcomes, citing no tests that show brands following the recommendations improved their position against peers.

None of this removes the practical use of the benchmarks themselves. Since Amazon first launched self-service benchmarks in November 2025 through two report types in its Reporting API, they have offered a category-level reference point produced by the platform itself, rather than drawn from a third-party tool's sample of accounts. The November page makes that reference point easier to read. It also attaches a set of instructions to it, written by the company that sells the inventory being benchmarked.

Timeline

Summary

Who: Amazon Ads, and the brand owners and representatives running sponsored ads and Display, Video, and Audio campaigns in 18 countries who can see peer benchmarks in Amazon's tools.

What: Amazon updated its Support Center page on benchmark metrics with an explanation of how p25, p50 and p75 levels are labeled for rate and cost metrics, where each level appears, a table of 16 recommended actions across eight metrics, and a short methodology note saying benchmarks are ad-spend weighted. Every favorable result in the table is paired with a suggestion to scale, invest more or bid higher, two rows suggest higher bids when costs are already below peers, and none suggests cutting a budget. The page describes the median as the point where half of peer brands sit above and half below, while Amazon's white paper describes a spend-weighted median.

When: The page carries an update date of November 11, 2026. Benchmarks reached general availability on May 18, 2026, and the legacy reporting pages, whose benchmark report cannot yet be migrated, close on January 31, 2027.

Where: In the Amazon Ads console, through the campaign manager insight card, unified reporting and the Reporting API, across Australia, Brazil, Canada, Egypt, France, Germany, India, Italy, Japan, Mexico, the Netherlands, Saudi Arabia, Spain, Sweden, Turkey, the United Arab Emirates, the United Kingdom and the United States.

Why: Amazon presents benchmarks as a reference point to help brands judge their performance against peers. The recommendations it now attaches to them shape how that reference point is acted on, and the spend weighting determines whose results define it. Both matter for budget decisions as Amazon moves reporting to unified reporting and changes how display impressions are counted from January 2027.