22 March 2026
It replaces the 13-month limit with two years of behavioral data, allowing you to track year-over-year trends and bypass 90-day retargeting limits to capture true customer lifetime value.
Focus on three key segments: Event Loyalists (seasonal deal-hunters), Lifecycle Upgraders (buyers due for long-term replacements), and Long-Cycle Researchers (hesitant shoppers researching high-ticket items for months).
Yes. AMC requires complex SQL queries to extract audiences from raw data. To bypass this technical barrier, most SMBs partner with agencies that use pre-built, proprietary SQL scripts.
No, these audiences only route to Sponsored Display and DSP. Avoid Amazon’s massive monthly minimums by partnering with an agency to use their shared DSP seats at lower costs.
For years, Amazon sellers have been operating with a severe case of short-term memory loss.
If you wanted to understand your customer’s journey, calculate their true lifetime value, or build a highly specific retargeting audience, you were constrained by a frustratingly narrow window of time. The standard Amazon Advertising console gave you 60 to 90 days of lookback data. When Amazon Marketing Cloud (AMC) first became accessible to smaller brands, it offered a 13-month lookback. For most SMB sellers, that already felt like a major improvement. But even a 13-month window still left many long purchasing cycles invisible, especially for products that customers replace or upgrade only once a year or less.
That limitation is officially gone.
The expansion of the Amazon Marketing Cloud lookback window to a full 25 months is arguably one of the most significant data unlocks Amazon has ever handed to third-party sellers. This change goes beyond a simple feature update. For many sellers, it fundamentally changes how they can analyze long-term customer behavior and plan full-funnel advertising strategies.
You now have over two years of granular, event-level data at your fingertips. You can see the shopper who clicked a Sponsored Brand ad in the spring of 2024, finally bought your product during Prime Day in July 2024, and is mathematically primed to buy a replacement or an upgraded model for Black Friday 2025.
For many smaller sellers, this is where things become difficult. The data exists but turning it into usable audiences requires a technical layer that most teams never had before AMC. Many brands now have access to AMC, but the advantage usually comes from how the data is used. Raw data alone does not create value. It only becomes useful when it is translated into specific audiences and campaigns that advertisers can actually act on. This is where many smaller Amazon sellers run into a wall. The data may exist, but without the ability to structure queries and extract meaningful segments, most of it remains untouched.
In this comprehensive guide, we are going deep into the mechanics of the 25-month AMC lookback. We will break down exactly why standard Amazon retargeting fails for long-term growth, and we will map out the precise architecture for the three most profitable custom audiences that you should be building, syncing, and bidding on today to aggressively scale your market share.
Before we dive into the advanced capabilities of Amazon Marketing Cloud, we must understand the fundamental flaws of the out-of-the-box retargeting tools available in the standard Amazon Ads console.
If you are currently running Sponsored Display (SD) purchases remarketing or standard Amazon Demand Side Platform (DSP) retargeting, you are likely relying on preset lookback windows. You tell the system, “Show this ad to anyone who viewed my product but didn’t purchase in the last 30 days,” or “Retarget past purchasers from the last 90 days.”
These standard campaigns are incredibly effective for fast-moving consumer goods (FMCG) like supplements, paper towels, or coffee. If a customer buys a 30-day supply of vitamins, hitting them with a Sponsored Display ad on day 25 is a brilliant, highly profitable strategy.
But what if you sell a premium leather duffel bag? What if you sell a high-end espresso machine, a memory foam mattress, or a baby stroller?
The purchase cycle for these items does not fit neatly into a 30, 60, or even 90-day window. If someone buys an $800 espresso machine from your brand, hitting them with a retargeting ad for another espresso machine 45 days later is not just a waste of ad spend; it is actively annoying your customer. By the time they might actually be in the market for a complementary product—like a high-end burr grinder or a descaling kit a year down the road—they have completely aged out of your standard retargeting audiences.
Your standard advertising console has forgotten they exist.
This short-term memory inherently forces sellers to constantly acquire net-new customers at a premium cost-per-click (CPC), rather than monetizing the customers they have already paid to acquire. In an ecosystem where inbound placement fees, fulfillment costs, and CPCs are rising year over year, relying purely on net-new customer acquisition is a mathematically unsustainable strategy. You must increase the Customer Lifetime Value (LTV), and you can only do that if you can actually remember who your customers are.
Amazon Marketing Cloud is a secure, privacy-safe “clean room.” It allows advertisers to access event-level data (every click, every view, every add-to-cart, every purchase) completely stripped of Personally Identifiable Information (PII). You cannot see that “John Doe from Chicago” bought your product, but you can see an anonymous user ID that represents a unique shopper, and you can track every interaction that unique ID had with your brand’s advertising and product listings.
For many advertisers, this concept feels unfamiliar at first. The idea of analyzing anonymized event data rather than customer profiles requires a different mindset.
When the window was limited to 13 months, you could barely scrape together a Year-over-Year (YoY) analysis. If you wanted to compare Prime Day 2024 to Prime Day 2025, you only had a one-month overlap.
By extending the window to 25 months, Amazon has given you two full, uninterrupted years of behavioral data, plus a one-month buffer. For many analysts, this is the first time they can clearly see the full customer journey over multiple seasons.
You can now use SQL (Structured Query Language) within AMC to filter this massive dataset and create highly bespoke audiences. Once you write the query and build the audience in AMC, you can push that audience directly into your Amazon DSP console to run highly targeted, highly efficient display and video ads across Amazon.com, Twitch, Freevee, and third-party publisher websites.
Instead of relying on Amazon’s algorithmic guesswork, you are dictating exactly who sees your ads based on two years of proven behavioral history. Let us look at the three most powerful audiences you can build with this data.
The biggest shopping days of the year—Prime Day, the Fall Prime Big Deal Days, Black Friday, and Cyber Monday—bring a massive influx of traffic to the platform. However, the shoppers participating in these events often possess a very specific psychological profile: they are deal-hunters.
Many of these customers will not buy your product at full price in March, but they will absolutely buy it when it features a red “Prime Day Deal” badge in July.
With a standard 90-day lookback window, it is impossible to retarget the people who bought from you during Black Friday of last year when Black Friday rolls around this year. They have aged out of your standard audiences.
The 25-month AMC lookback solves this beautifully. It allows you to build an audience of “Event Loyalists.”
The goal here is to isolate the user IDs of every shopper who purchased your product, or engaged heavily with your ads, during a specific major deal event in the prior year, and then aggressively target them in the weeks leading up to the current year’s event.
For example, an AMC SQL query can be written to: “Find all unique user IDs who purchased a product from my brand catalog between November 20th and November 30th of 2024. Exclude anyone who has made a purchase in the last 60 days.”
Many sellers eventually notice that a portion of their customers behaves this way year after year. They rarely purchase outside major promotions, but when deal events arrive, they return quickly.
Once you push this audience to your DSP console, you do not want to hit them with a generic brand awareness ad. You know exactly what motivates this specific cohort: major seasonal discounts.
Two weeks before Black Friday 2025, you launch a DSP campaign specifically targeting this AMC audience. The creative should heavily tease your upcoming Black Friday promotions. When the event goes live, you switch the creative to highlight the massive discounts on your new product lines or complementary accessories.
Because you are targeting proven buyers who have a demonstrated history of converting during high-velocity deal events, your conversion rates on this specific DSP campaign will often dwarf your standard non-branded category targeting. You are turning one-time deal hunters into repeat, annualized buyers, effectively doubling their LTV while completely bypassing the bloodbath of generic keyword bidding during Q4.
One of the most tragic wastes of ad spend on Amazon is the failure to map advertising to the natural lifecycle of a product.
If you sell a consumable product like dog food or vitamin C, Amazon’s standard “Subscribe & Save” feature handles the replenishment cycle fairly well. But what if you sell products that have a 12-to-18-month lifecycle?
Consider a brand that sells baby gear. A customer buys a newborn bassinet. Twelve months later, that child has completely outgrown the bassinet and desperately needs a toddler car seat or a heavy-duty stroller. The customer is actively back in the market.
If you use standard Sponsored Display, you cannot reach that customer. Your 90-day window expired nine months ago. If you do not have AMC, you are forced to bid on highly competitive, expensive keywords like “toddler car seat,” fighting against massive legacy brands, hoping your previous customer happens to see your Sponsored Products ad and remembers they liked your bassinet.
With the 25-month AMC lookback, you can engineer the “Lifecycle Upgrader” audience. In reality, these lifecycle patterns look different across product categories. Some products are replaced every six months, others every eighteen. The longer lookback window finally makes those patterns visible.
This strategy requires you to understand the exact timeline of your product catalog’s ecosystem. You must map out the natural progression of your customer.
Let’s say you sell premium water filtration pitchers, and the core unit comes with a filter that lasts exactly 12 months.
A typical AMC query for this audience might look like this: “Find all user IDs who purchased the core water pitcher ASIN between 11 and 13 months ago. Ensure that these specific user IDs have not purchased the replacement filter ASIN within the last 6 months.”
You have just built a hyper-targeted audience of people whose water filters are currently expiring or have just expired. They literally need your product right now.
You push this custom audience to your DSP console. You set up a highly aggressive campaign with a strong call to action: “Time for a fresh filter? Get 15% off a 3-pack today.” Because the relevance of the ad is incredibly high, and the audience is perfectly timed to their exact point of need, the click-through rates (CTR) and Return on Ad Spend (ROAS) on these campaigns are typically astronomical. You are no longer guessing when a customer might need you; you are using 25 months of historical data to show up exactly at the moment of friction. This type of strategy often works particularly well for categories such as tech upgrades, automotive parts, beauty tools, and other multi-stage product ecosystems.
Many high-ticket purchases rarely happen on impulse. If you are selling a $1,500 luxury mattress, a $600 home security system, or a high-end robotic vacuum, the customer journey is rarely linear.
A shopper might click your Sponsored Brands video ad in February, read the reviews, check out your brand store, and then abandon the purchase because they aren’t ready to spend that kind of money. They might research competitors for months. They might wait until they get their tax refund in April, or until a major holiday sale in November.
In many cases, the same shopper may return several times over months before making a final decision. Standard advertising attribution rarely captures this type of slow research behavior.
In the standard Amazon Ads console, a shopper who clicks an ad in February and buys in November is considered a completely new, un-attributed organic sale or a net-new ad acquisition if they clicked a different ad that day. The standard 14-day attribution window completely loses the narrative. You have no idea that your top-of-funnel video ad from nine months ago actually drove the consideration that led to the sale.
More importantly, you lose the ability to nurture that researcher over their long consideration phase. The 25-month AMC lookback allows you to build the “Long-Cycle Researcher” audience, ensuring you stay top-of-mind for high-ticket buyers.
This audience focuses on high-intent engagement without a purchase over an extended period.
A common AMC query for this audience might look like this: “Find all user IDs who have clicked on our Sponsored Brands or Sponsored Display ads, or who have visited our brand store, at least three times in the last 12 months. Cross-reference this list with our purchase data and exclude any user ID that has actually purchased a product from our catalog.”
You have just isolated a group of highly interested, highly hesitant shoppers. They know your brand. They have researched your products multiple times. But something is holding them back, usually price or a lack of final trust.
You push this audience to DSP. Since these are high-ticket items, you can afford a slightly higher CPA (Cost Per Acquisition) to finally convert them. You serve them specific ad creatives designed to overcome objections. You show them DSP video ads highlighting your product’s multi-year warranty. You serve display ads featuring quotes from five-star reviews.
Finally, when you run a major promotional event, you bid incredibly aggressively on this specific AMC audience. You know they have been researching your $600 robotic vacuum for six months. When you drop the price to $450 for Prime Day, you want your DSP ad to be the very first thing they see when they log onto Amazon or read an article on a partner site.
In practical terms, this audience allows brands to reconnect with shoppers who showed strong interest but never completed the purchase. Of course, not every researcher will convert. Some shoppers simply continue comparing options across multiple brands for months before making a final decision.
Reading about these advanced, 25-month custom audiences is incredibly exciting. It feels like unlocking a cheat code for Amazon growth. However, we must address the massive, glaring elephant in the room: the technical barrier to entry.
Amazon Marketing Cloud is not a user-friendly, drag-and-drop dashboard. It does not have pretty pie charts or simple toggle switches. It is a raw data environment. To extract these custom audiences, you must know how to write complex, perfectly structured SQL queries. You must know how to join massive datasets, handle overlapping user IDs, and structure the output so that it is accepted by the Amazon DSP API.
Furthermore, running these audiences typically requires access to Amazon DSP, which historically carried massive minimum monthly ad spends (often $35,000 or more if managed directly through Amazon’s managed services).
For a small to medium-sized business doing $2 million or $5 million a year on Amazon, hiring a full-time SQL data scientist and committing to a $35k/month DSP budget is completely out of the question. This is exactly why AMC was, for a long time, the exclusive playground of Fortune 500 enterprise brands.
This is precisely where the value of a highly technical, data-driven Amazon Ads agency comes into play. Agencies like SellerMetrics bridge the gap between the raw, intimidating power of the 25-month AMC lookback and the practical realities of running an SMB.
Advanced agencies have already written, tested, and optimized the SQL queries required to build the Event Loyalists, the Lifecycle Upgraders, and the Long-Cycle Researchers. They maintain a library of proprietary AMC queries that can be deployed into your brand’s clean room environment almost instantly. You do not need to learn how to write a LEFT JOIN or deal with query syntax errors.
Moreover, agencies utilize DSP entity seats that allow them to aggregate spend across multiple clients. This completely bypasses the massive minimum spend requirements imposed by Amazon. An SMB can tap into the power of DSP and custom AMC audiences with a fraction of the budget, testing these advanced retargeting strategies safely and profitably.
For many advertisers, guesswork is gradually being replaced by clearer behavioral data. The era of settling for a 60-day memory is behind us. You have 25 months of pristine behavioral data waiting to be mined. Brands that start using this data to build lifecycle-driven audiences often gain an advantage over competitors that rely only on standard targeting.
Stop leaving data on the table. Access your clean room, query your history, and start building the audiences that will drive your brand’s profitability for the next two years.
For many brands, this shift will take time to fully understand. The opportunity is large, but it requires a different way of thinking about customer data. But once these audience patterns become visible, they can reshape how long-term campaigns are planned.
Would you like me to run a free audit on your current ad account to see exactly how much revenue you are losing by not retargeting your past seasonal buyers? Let’s connect and look at your data.
Amazon Marketing Cloud is a secure, privacy-safe “clean room” provided by Amazon. It allows advertisers to access and analyze event-level datasets (like impressions, clicks, and purchases) across their Amazon Ads campaigns. It removes all Personally Identifiable Information (PII) but allows you to track anonymous user IDs to build custom audiences and perform deep-dive analytics using SQL. Many advertisers initially find the interface intimidating because it relies on SQL queries rather than a traditional reporting dashboard.
Previously, AMC only allowed a 13-month lookback. This made it difficult to analyze long-term customer behavior or accurately compare Year-over-Year (YoY) performance for major seasonal events. The 25-month window gives sellers two full years of data, allowing them to track long-term replenishment cycles, measure true customer lifetime value (LTV), and retarget buyers across multiple annual events.
Yes, interacting directly with raw AMC data requires a strong proficiency in SQL (Structured Query Language). You must write code to query the database, join tables, and extract audiences. If you do not know SQL, you will need to partner with an advanced Amazon agency or use third-party software that provides pre-built queries and visual interfaces.
Standard DSP or Sponsored Display retargeting relies on pre-set, rigid lookback windows (e.g., “viewed in the last 30 days” or “purchased in the last 90 days”). AMC Custom Audiences allow you to write highly specific, complex rules using 25 months of data (e.g., “clicked an ad 8 months ago, added to cart 6 months ago, but never purchased”).
No. Currently, the custom audiences you build within Amazon Marketing Cloud can only be pushed to Amazon DSP (Demand Side Platform) or utilized within Sponsored Display campaigns. Sponsored Products remain strictly keyword- and ASIN-targeted.
If you go directly through Amazon’s managed services, they typically require a minimum commitment of $35,000 to $50,000 per month. However, if you partner with an Amazon Ads agency that has an established DSP seat, they can provide access to DSP and your AMC audiences with significantly lower, or even zero, minimum ad spend requirements.
Once you successfully run an SQL query in AMC to generate a custom audience and push it to your connected DSP account, it typically takes between 24 to 48 hours for the audience to fully populate, resolve, and become available for active targeting in your campaigns. In practice, advertisers often check audience size the following day to make sure enough users qualify before launching a campaign.
Absolutely not. AMC is a privacy-safe environment. All data is anonymized. You can track the behavior of “User_ID_98765” over 25 months to serve them highly relevant ads, but you will never know that user’s name, email address, physical address, or any other Personally Identifiable Information (PII).
No, there are prerequisites. Generally, you must have an active Amazon DSP agreement, an active Amazon Ads account, and be a registered brand owner. The exact eligibility requirements can shift, which is why many SMBs access AMC functionality through an agency partner who already has the necessary infrastructure and approvals in place.
Yes. When your AMC instance is provisioned and activated, Amazon backfills your historical advertising data. If you have been running Amazon Ads consistently, you will immediately have access to your historical event data stretching back across the 25-month lookback window, allowing you to start building long-term audiences on day one.
22 March 2026
TweetLinkedInShareEmailPrint 8 min read By Rick Wong Updated Mar 22, 2026 TL;DR Why can’t I just ...
9 March 2026
TweetLinkedInShareEmailPrint 18 min read By Rick Wong Updated Mar 10, 2026 TL;DR Why should new sellers...
27 February 2026
TweetLinkedInShareEmailPrint 8 min read By Rick Wong Updated Feb 27, 2026 TL;DR What is the optimal len...