Amazon Marketing Cloud's 5-Year Dataset: What It Unlocks

For years, Amazon Marketing Cloud only showed advertisers a shopper's last 13 months of purchases. A customer who bought once, then came back years later, looked to AMC like two unrelated people. Amazon Marketing Cloud's 5-year dataset closes that gap.

Amazon replaced the 13-month lookback with five years of purchase history when it launched the Amazon Retail Purchases dataset in May 2025, according to Amazon's own announcement of the dataset. Query access started as a paid subscription with a free trial attached.

Amazon made it free to run through the end of 2026 starting this June, Search Engine Journal reported. Audience creation with the dataset has been free from the start.

How the 5-year dataset changes the lifetime value math

Three of the six use cases Amazon highlights come down to the same problem: a 13-month window undercounts anything that doesn't repeat quickly. Lifetime value, new-to-brand rates and what Amazon calls gateway products all move once five years of history sit inside the query.

Consumer electronics is the example given for new-to-brand measurement. A laptop or a TV might not sell to the same household again for three or four years.

Under a 13-month window, almost every one of those buyers counted as new-to-brand, whether or not they'd bought the brand before. Stretching the window to five years fixes that distortion for any product with a repurchase cycle longer than about a year.

Gateway products work the other direction. They're the items that pull a first-time customer into a category at all. Knowing which SKU usually comes first tells a brand what to promote to new shoppers and what to hold back for people who already trust it.

What the extra history is actually for

The other three use cases are about winning customers back, not just measuring them.

Amazon's own materials describe reacquiring customers lost over a year ago through custom lookback windows, lifetime-value thresholds and product-specific segments. That beats losing them the moment they age out of a 13-month window.

Intentwise's explainer gives Prime Day as the test case. Whether a discount-driven purchase turns into a loyal repeat customer only shows up if you can compare purchases years apart, not just within the same 13-month stretch.

The sixth use case lets a brand upload hashed shopper data from its own channels and match it against Amazon purchase history, connecting a customer's lifecycle across both.

That's the first version of AMC able to see a customer's whole relationship with a brand rather than half of it, assuming the direct-to-consumer storefront collects data clean enough to match against Amazon's.

Building and refreshing six overlapping audience queries by hand doesn't scale past the first campaign. That's where automating the audience-refresh work earns back the hours AMC's console takes to query manually.

Where this fits into a broader retail media stack

Photo by Ron Lach on Pexels

AMC's numbers work best as one input into a wider growth strategy. Feeding them into ongoing marketing and growth planning, instead of pulling a report once a quarter, is what turns five years of purchase history into an actual budget decision.

Retail media keeps getting more central to how shoppers find products, and Amazon isn't the only one building around it. Platforms outside Amazon are testing AI-driven shopping search of their own. Owning clean first-party purchase data matters everywhere that's happening, not just inside AMC.

The 60-day free trial is still the cheapest way to find out whether your catalog behaves like consumer electronics, with a repurchase cycle measured in years, or like something that turns over every few months. That answer decides which of the six use cases is worth building first.

Cover photo by Miguel Á. Padriñán on Pexels