Free AMC Signals End December 31: What Your Q4 Data Should Leave Behind

Amazon's free window on first-party AMC signals closes on December 31, 2026. From mid-November, that leaves roughly seven weeks, and the first three of them are Black Friday, Cyber Monday, and the December push. The richest shopper data your account will produce all year is about to land inside a window that expires before most teams come up for air.

That timing is the real story of this year's AMC changes. In June, Amazon removed the query cost on its 1P paid-feature signals through year-end (as reported by PPC Land). In July, it published a launch announcement for ten Brand+ and Performance+ updates, led by Audience Signals, which turns your uploaded audiences into a model input instead of a targeting cap. Together, they made it cheaper to learn who your best customers are and more valuable to feed that knowledge back into Amazon's automated campaigns. The question for the next seven weeks is what you keep once the price comes back.

Quick Answer

• What changed: Since June 2, 2026, Amazon's own 1P paid-feature signals in Amazon Marketing Cloud (AMC), including Amazon Audience Segment Insights and Amazon Retail Purchases, have been free to query for audience creation and measurement. The offer ends December 31, 2026. Third-party signals such as Experian still cost money.
• What else shipped: On July 22, 2026, Amazon added ten capabilities to Brand+ and Performance+, its automated campaign products inside Amazon DSP. These campaigns serve across Amazon's own sites and apps, Prime Video, audio, and third-party exchange supply. The headline feature, Audience Signals, lets your first-party audiences steer the model without capping reach.
• The opportunity: Peak-season data is the best input AMC will see all year, and you can query it at no cost until December 31. Audiences and incrementality benchmarks built from it keep their value into 2027.
• The risk: The window closes while most teams are heads-down on peak season, Amazon has not said what 1P signals will cost from January 1, and several default-on model changes will blur any year-over-year Q4 comparison.
• Our read: Treat the remaining weeks as a build-and-bank sprint. Pull what you need from Q4 data, save it outside the query layer, and set up the audiences that will feed Brand+ and Performance+ next year.

Why the Deadline Matters More Than the Launch

In June, "free through year-end" sounded like six months of runway. For most teams, summer events, back-to-school, and Q4 planning absorbed that runway, so the working window is the seven weeks that remain. Three things make those weeks unusual.

Peak season is the dataset worth querying

Holiday traffic brings new-to-brand buyers, gift purchasers, and deal-driven category switchers that barely show up the rest of the year. AMC's 1P signals are how you tell which of them were reached by your ads, which ones converted because of them, and which are worth pursuing again. Querying that data for free is the most valuable use of the offer.

Retention questions outlast the window

Whether a Black Friday buyer comes back shows up in January and February, after the free period ends. You can define and save those cohorts now. The follow-up read will run under whatever pricing Amazon sets for 2027, so the cohort definitions and baselines you capture this quarter are what make that later spend worth it.

Nobody knows the 2027 price

Amazon has not announced successor pricing for 1P signals. It could extend the offer, set a new price, or repackage the features. Planning around any one of those outcomes is a guess. Banking the outputs you need before December 31 holds up under all of them.

This matters most for Amazon sellers and brands already running Amazon DSP, including Brand+ or Performance+, with an AMC instance in place. If your spend is Sponsored Products only, read this as input for your 2027 plan rather than a Q4 task.

What Shipped This Year, in Brief

Free 1P AMC signals (effective June 2, 2026). Amazon's premium AMC paid features became free to query for audience creation and measurement through year-end. You still subscribe through the Paid Features tab in your AMC instance; the payment step is removed for Amazon's own 1P signals. Third-party paid signals still require a subscription.

Ten Brand+/Performance+ features (announced July 22, 2026). The ones that bear on the free window:

• Audience Signals (announced as Open Beta): for Brand+ Prospecting and Performance+ Customer Acquisition and Unified Consideration. First-party audiences connected through Ads Data Manager become optimization inputs at no additional fee.
• Audio for Brand+ (General Availability) and Audio for Performance+ (Open Beta): automated audio buying across 16 locales, now reaching mid- and lower-funnel tactics.
• Automatic Deal Selection for Brand+ STV (Open Beta): auto-activates and optimizes a set of premium streaming-TV deals. Enabled by default on new prospecting lines.
• Brand Outcome Optimization (Open Beta): default-on for Brand+ Reach goal lines; optimizes toward branded search, detail-page views, and add-to-cart.
• Prime Video Signal Integration (Open Beta): default-on for Prime Video Ads supply; adds streaming behavior as a ranking signal.
• Asynchronous Remarketing, Expanded Conversion Support for App Ads, and the Managed Service Campaign Proposal Builder (all GA): richer bid-time signals, broader app conversion types, and access for managed-service accounts.

The default-on features are the ones to note heading into peak season. They have been changing how your campaigns bid since late July, which means this year's Q4 results reflect Amazon's model changes as well as your own decisions. Any year-over-year holiday comparison needs that caveat.

Signal vs. Fence: Why Better Audiences Pay Off Next Year

Before Audience Signals, a first-party audience in Brand+ or Performance+ acted as a hard targeting constraint, a fence. The AI could only spend inside the segment you defined. Audience Signals turns the same audience into an optimization input, the way conversion-event data already works. The model uses your list to find high-value patterns faster while staying free to reach users outside it.

• A targeting constraint (fence) caps deliverable reach. If your list is 40,000 people, delivery stops at that pool.
• An optimization input (signal) biases the bidder toward a pattern without capping reach. The model treats your list as a directional hint about who converts.

This is where the free window and the July launch connect. Under the old fence model, a better audience mostly changed who you could reach. Under Audience Signals, a better audience changes how fast the model learns and where it spends. An audience built from real holiday purchase behavior this quarter is a stronger input for every Brand+ and Performance+ line you run in 2027.

The catch runs the other way too. The difference between a signal and a fence is not directly visible in the advertiser console, and a noisy list does not fence off bad spend; it nudges the bidder in a fuzzier direction. The quality of what you build before December 31 carries straight into next year's results.

Using AMC Data For and From BFCM

From mid-November, Black Friday through Cyber Monday is two weeks out. That splits the free window into two jobs: use AMC to set up the event, then use the event to set up 2027.

Before BFCM: set up the event with last year's data

• Replay last year's Cyber Week. If your instance has history back to November 2025, path-to-conversion queries show when holiday buyers started researching, which ad types they saw before purchasing, and where extra impressions stopped adding conversions. That tells you how early to start spending and where to set frequency caps this year.
• Build remarketing pools now. Shoppers who viewed a detail page or added to cart in the past 30 to 60 days without buying are the highest-intent audience you will have on Black Friday. AMC audiences take time to populate, so build them this week rather than the day before deals start.
• Suppress recent buyers from prospecting. Peak-week inventory is the most expensive of the year. Excluding customers who purchased recently keeps prospecting budget on new shoppers.
• Load signals before the rush, not during it. Adding Audience Signals to a Performance+ line now gives the model a head start before traffic spikes. Restructuring campaigns mid-event makes both performance and attribution harder to read.

After BFCM: turn event data into 2027 inputs

• Split holiday buyers by what they are worth. Separate new-to-brand buyers, existing customers who bought again, and shoppers who engaged but never purchased. Each becomes a different audience next year: a signal for acquisition lines, a retention segment, and a January remarketing pool.
• Read which touchpoints did the work. Path analysis on event-week conversions shows which ad types and channels assisted purchases, not only the last click. That is the input for your 2027 budget mix and your next Prime Day plan.
• Measure incrementality on the highest-spend days. Exposed-versus-unexposed conversions for BFCM week answer whether peak spend added sales or captured shoppers who would have bought anyway.
• Check deal-driven cohorts before feeding them to the model. A list full of one-time discount buyers can teach Audience Signals to chase bargain hunters. Look at repeat behavior in January before treating a holiday cohort as a high-value signal.

What the Next Seven Weeks Should Produce

A useful end state on January 1 is a short list of saved assets: an incrementality benchmark, a few durable cohorts, and audiences ready to run as signals. Here is the order that gets you there.

1. Confirm your 1P paid-feature subscription is active today. Subscription is at the instance level, so everyone on your AMC instance inherits access. Re-subscribing after a cancellation carries a 24-hour wait, which you do not want to discover on Black Friday.
2. Mark your baseline before peak traffic. Write down the dates default-on features reached your lines and any structural changes you make in November, so holiday lift can be attributed later.
3. Run an incrementality read on peak-season spend. Compare ad-exposed and unexposed conversions using exposure-based signals. This is the "did the ads cause the sales" number finance will ask for in January, and it is free to produce until December 31.
4. Define two or three cohorts from holiday buyers. New-to-brand holiday purchasers, repeat buyers, and category switchers are the archetypes that give the model a meaningful head start next year.
5. Save the outputs outside AMC before the window closes. Keep the query text, cohort definitions, result exports, and benchmark numbers somewhere your team controls, so their value does not depend on 2027 pricing.
6. Connect the cohorts through Ads Data Manager as Audience Signals. Attach them to a Brand+ Prospecting or Performance+ Customer Acquisition line ready for January, as inputs rather than fences.
7. Keep those audiences refreshed and in sync with live campaigns after January 1. Cohorts decay as shoppers move on, and a signal is only as good as its last refresh. This is the step most teams never get to, because it has to repeat every month after the deadline pressure is gone.

Example: Turning Holiday Data Into a 2027 Budget Decision

Input: A supplements brand runs Sponsored Products plus a modest Performance+ Customer Acquisition line. It has an AMC instance with a 25-month lookback and a hashed list of repeat buyers.

Workflow: In mid-November, the brand confirms its 1P subscription and logs a baseline. Through Cyber Week, it queries exposure-based conversions to estimate the true incrementality of its DSP spend. In December, it builds a "holiday new-to-brand" cohort and a "repeat buyer" cohort, exports the definitions and results, and connects both through Ads Data Manager as Audience Signals on its January Performance+ line.

Expected output: An incrementality read on the highest-spend weeks of the year, produced at no query cost, plus a Performance+ line that starts 2027 with real customer patterns instead of a cold start.

Business decision: If the incremental read holds up, the brand shifts budget from a plateaued Sponsored Products campaign into Performance+ for Q1 with evidence behind the move, and it knows which holiday cohorts are worth paying to re-measure once 2027 pricing is set.

What These Changes Cannot Do

• Free signals do not pick your audience. Removing the query cost lowers the price of measurement. It does not tell you which cohorts are worth building or how to sequence them across the funnel.
• Audience Signals will not rescue a weak segment. As an input rather than a fence, a low-quality list produces low-quality bias, not a hard stop.
• Default-on models reduce your visibility, not your responsibility. You still own goal selection, creative, product choice, retail destination, and measurement windows.
• The free window is temporary. Anything that only works while queries are free will need a budget line or a rebuild in January.
• Geography and reporting vary. Audio covers 16 locales, and the launch page lists Automatic Deal Selection for Brand+ STV as U.S.-only. Reporting that compares an auto-selected deal set with a hand-picked deal has not been fully specified.

Tools and Data You Need

• An AMC instance with 1P paid features subscribed (free through December 31, 2026).
• Ads Data Manager access to connect hashed first-party lists.
• A Brand+ or Performance+ campaign in a supported market.
• Clean conversion and margin data. The model optimizes toward what you give it, so errors in the input become errors in the bidding.
• For UK/EEA data, confirmed consent (Amazon required it for data flowing through Ads Data Manager in those regions from June 30, 2026).

Common Mistakes Before the Deadline

• Waiting until after Cyber Week to subscribe. The 24-hour re-subscribe wait and holiday workloads both argue for doing it now.
• Querying without saving. A result you only viewed in AMC may cost money to reproduce in January. Export it.
• Skipping the baseline. Launching new audiences in the same weeks default-on models and holiday traffic spike makes lift impossible to attribute.
• Confusing free 1P with free everything. Third-party signals such as Experian still cost money.
• Assuming console reporting proves the signal worked. The input-vs-fence difference is not directly visible; validate through AMC analysis.

Where Trellis Fits: Keeping the Audiences Alive After January 1

Step 7 is where the value of the free window is won or lost. The incrementality read and the holiday cohorts only pay off if they keep feeding live campaigns in February, April, and next Prime Day, long after the deadline has passed. Doing that by hand means rerunning queries, rebuilding cohorts, and keeping activation in step with campaigns every month, which is the work that stops once Q1 gets busy.

Trellis is built for that ongoing layer, so the audiences you build this quarter keep compounding into revenue and margin instead of decaying. Qore from Trellis codifies the routine from steps 3 through 6, the cohort refresh and the incrementality report, and runs it on a schedule with approval gates, so the logic stays consistent month to month. For teams that want DSP execution handled for them, Qinetix applies those audiences to live Amazon DSP campaigns. AMC standard access is free on Amazon's side; Trellis adds the operating layer, not a toll on the data. You still decide which cohorts and goals matter.

For the deeper mechanics, our guides on using Amazon Marketing Cloud data, DSP audiences for retail, and Amazon DSP best practices for 2026 go further than we can here.

Conclusion

Amazon's two moves this year pull in the same direction. Free 1P signals made it cheaper to understand your customers, and Audience Signals made that understanding more useful inside Brand+ and Performance+. With the free window closing on December 31, the most valuable data of the year is arriving just as the price clock runs out. Subscribe now, mark your baseline, read incrementality on peak-season spend, and save the cohorts and benchmarks you build. The price of querying may come back in January. What you banked before then does not have to.