Key takeaways

  • Set every target from product margin, then manage the account to TACoS, not to a single ROAS number.
  • Since January 1, 2026, default view metrics for vCPM Sponsored Brands and Display use a shorter window, so compare like with like.
  • High reported ROAS on branded and remarketing ads can hide a loss; test incrementality before you scale them.
  1. 1Set targets from margin

    Break-even and target ACoS per ASIN

  2. 2Split by intent

    Branded, non-branded, competitor, remarketing

  3. 3Fix your reports

    Rebuild in unified reporting before Dec 31

  4. 4Test incrementality

    Two-week branded holdout, AMC new-to-brand

  5. 5Move budget weekly

    Steps of 20% max, judged on TACoS

What Amazon advertising optimization really means

Most guides treat optimization as a list of tweaks: raise bids on winners, add negatives, shift placement modifiers. Those tweaks matter, but they all optimize toward the number the console reports. If that number overstates what your ads cause, every tweak pushes money in the wrong direction faster.

A better definition has three parts. First, a target for every product that comes from its margin, not from a benchmark. Second, a view of which campaigns create sales that would not have happened anyway. Third, a clear job for each ad type, so Sponsored Products, Sponsored Brands, Sponsored Display, Sponsored TV and DSP are judged on what they are built to do.

The tactical work then becomes simple: move budget from spend that fails its target to spend that beats it, one controlled change at a time.

What changed in Amazon Ads measurement in 2026

Three changes make 2025 playbooks unreliable this year. Most ranking articles on this topic were written before them.

On January 1, 2026, Amazon introduced a shopping-signal enhanced last-touch model for view attribution. It applies to Sponsored Brands and Sponsored Display campaigns billed on viewable impressions (vCPM), including branded keyword campaigns, and to Amazon DSP ads served in the Amazon store. Default purchases, sales and ROAS now use a shorter view window that Amazon has not published. The old 14-day view numbers survive as separate "all views" metrics such as Purchases (all views). Click attribution did not change.

The practical effect: if you compare a vCPM campaign's ROAS this year with last year, you are comparing two different rulers. Pick one view of the data, label it, and use it for every comparison.

Second, Amazon's unified reporting left beta on June 8, 2026, and the legacy Sponsored Ads reports page and Amazon DSP reports page are set to retire on December 31, 2026, according to PPC Land. Any scheduled report, spreadsheet or tool that still depends on those pages needs to be rebuilt before year end, or your optimization data simply stops arriving.

Third, Amazon Marketing Cloud has been open to all sponsored ads advertisers in the ads console since September 18, 2025, with no-code templates for path to conversion and audience insights. Incrementality and new-to-brand analysis that used to need a DSP contract is now within reach of most brands.

Start from margin: the only targets worth optimizing to

Optimization needs a finish line. For each product, take the price, subtract landed cost, Amazon referral and FBA fees, and any returns allowance. What is left, as a share of price, is your break-even ACoS. Spend above it and every attributed sale loses money.

Your target ACoS should sit below break-even by the profit you want to keep. A product with a 40 percent break-even and a goal of 15 percent profit after ads has a 25 percent target. Launch products can run closer to break-even for a planned period, because the goal is rank, but that should be a written decision with an end date.

Then manage the account to TACoS, total ad spend divided by total sales. If ACoS holds steady while TACoS falls, ads are lifting organic sales too. If TACoS climbs while ACoS looks fine, ads are mostly buying sales you would have made anyway.

Worked example: the best ROAS campaign that loses money

Illustrative numbers, not a real account. A brand sells a $30 product with a 40 percent break-even ACoS, so each dollar of sales leaves 40 cents before advertising. It spends $10,000 a month across two Sponsored Products campaigns.

Branded search: $2,000 spend, $20,000 attributed sales, a reported ROAS of 10. Non-branded search: $6,000 spend, $20,000 attributed sales, a reported ROAS of about 3.3. On the console, branded looks three times better.

Now the brand runs a two-week pause test on branded keywords where no competitor bids on its name, and sees that about 80 percent of branded sales still arrive organically. Only $4,000 of the $20,000 was incremental. That $4,000 earns $1,600 of contribution against $2,000 of spend: a $400 loss each month. Assume non-branded sales are 80 percent incremental: $16,000 of new sales earns $6,400 against $6,000 of spend, a $400 gain.

The campaign with the best reported ROAS is the one losing money at the margin. The fix is not to turn branded off everywhere. It is to cap branded bids at the level that still defends the top slot, and move the freed budget into non-branded terms that beat target.

  • Reported ROAS ranks branded first: 10 vs 3.3
  • Incremental contribution ranks it last: minus $400 vs plus $400
  • Change made: cap branded bids, move budget to non-branded winners

Give each ad type one job, and judge it on that job

Mixing ad types into one ROAS number is the most common reason optimization stalls. Each format reaches shoppers at a different moment, so each needs its own measure of success.

  • Non-branded Sponsored Products: buys new demand. Judge on ACoS against the profit target and on organic rank for the terms you advertise.
  • Branded Sponsored Products: defends your name. Judge on whether competitors take your top slot when you bid less, not on ROAS.
  • Sponsored Brands: wins the category search and the Brand Store visit. Judge on new-to-brand share of orders and on clicks, since many view conversions now sit in a shorter default window.
  • Sponsored Display and remarketing: closes shoppers who already viewed you. Judge on incremental lift, because last-touch credit flatters it.
  • Sponsored TV and DSP: build awareness above the search. Judge on branded search volume and new-to-brand customers over months, not same-week sales.

Amazon advertising optimization playbook: 6 steps for this quarter

Run these in order. Each step names who owns it, where it happens and the metric that proves it worked.

  • Finance or the brand owner: build a margin sheet per ASIN with price, landed cost, fees and returns, and set break-even and target ACoS. Metric: every active campaign has a written target.
  • Ads manager: in the Amazon Ads console, split campaigns into branded, non-branded, competitor and remarketing so no campaign mixes intents. Metric: each campaign can be read against one job.
  • Ads manager: rebuild scheduled reports in the unified reporting builder before December 31, 2026, and add Purchases (all views) next to default purchases for vCPM campaigns. Metric: no gaps in your weekly data in January.
  • Analyst or agency: run one holdout test on branded Sponsored Products, pausing for two weeks on terms where no rival bids on your name, and compare total sales for those ASINs with the weeks before. Metric: share of branded sales that stayed organic.
  • Analyst: open Amazon Marketing Cloud in the console and use the path to conversion and new-to-brand templates to find which ASINs bring in new customers. Metric: new-to-brand share by ASIN, used to rank where non-branded budget goes next.
  • Ads manager: move budget weekly from campaigns above target to non-branded campaigns below target, in steps of no more than 20 percent, and log every change. Metric: TACoS falling or flat while total sales hold or grow.

When this advice does not apply

Holdout tests need volume. If a product sells only a handful of units a week, two weeks of noise will drown the signal, and you are better off watching TACoS over a quarter.

Branded pause tests are also risky when competitors already bid on your brand name or when a product is in its launch window, because losing the top slot can cost rank that takes months to rebuild. Seasonal products should not be tested in their peak weeks, when last year's pattern is not a fair baseline.

Finally, if listings convert poorly, no ad optimization will rescue the economics. Fix images, price and reviews first, then optimize the spend.

How PPC Spider runs Amazon advertising optimization

PPC Spider sets every campaign's ACoS cap from the product's real margin and manages the account to a TACoS that leaves money after fees, with both shown in the weekly report. SpiderOS, our own platform, tracks profit per product and monitors bids, budgets and search terms 24x7.

Amazon PPC management is a flat monthly retainer, never a percentage of ad spend. Launch is $995 a month for up to $10k of monthly ad spend with weekly optimization. Growth is $1,250 a month for up to $40k, adding daily bid and budget management, a campaign architecture rebuild and AMC basics. Scale is $2,200 a month for up to $120k, adding AMC custom audiences, Sponsored TV and a dedicated strategist. Larger accounts and DSP get an Enterprise quote. Every plan starts with a 90-day term, then runs month to month.

If you want to see where your current spend leaks before you change anything, start with a free Amazon audit.

Sources

  • Amazon Ads: View attribution updates for Amazon store ads, effective January 1, 2026 (advertising.amazon.com/resources/whats-new/view-attribution-updates-for-amazon-store-ads)
  • Amazon Ads: Sponsored ads advertisers can now access Amazon Marketing Cloud, September 18, 2025 (advertising.amazon.com/library/news/amc-for-sponsored-ads)
  • PPC Land: Amazon Ads unified reporting exits beta, June 2026 (ppc.land/amazon-ads-unified-reporting-exits-beta-and-takes-two-old-tools-with-it)
  • Feedvisor: Amazon incrementality guide, March 2026 (feedvisor.com/university/amazon-incrementality)
  • Intentwise: 3 ways AMC can level up your Sponsored Ads (intentwise.com/blog/3-ways-amc-can-level-up-your-sponsored-ads)
Get the next change before your competitors do

We send the marketplace changes that matter each week, with what to do next.

By subscribing you agree to our privacy policy. Unsubscribe any time.
Want this done for your account?

We run these playbooks on client accounts every day. Start with a free audit.

Get a free audit