Attain purchase data lands inside Amazon Ads DVA+

Attain, a purchase data company, says its audiences can now be built and activated natively inside Amazon Ads DVA+. For brand owners buying display and streaming TV, that means reaching people who bought a rival product at another retailer, online or in a physical store, from a panel Attain puts at over 14 million consumers.

What Attain connected to DVA+

Attain said on November 5, 2026 that advertisers can now build and activate its purchase-based audiences directly inside Amazon Ads DVA+, the display, video and audio buying product that took over from the Amazon DSP name in September. A demand-side platform is the software that bids on individual ad impressions for an advertiser. DVA+ is the campaign flow built on top of that machinery, and Attain says the integration was live from the day of the announcement and reaches the full range of inventory of Amazon Ads DVA+.

The segment builder combines three variables. Attain says advertisers can segment consumers by retailer or combination of retailers, purchase category, and shopping channel, meaning online or in store. So a segment can say something like bought this category at these two chains, in a physical store, recently. The supply behind it is a consumer panel. Attain says it connects directly with over 14 million consumers through owned and operated apps spanning financial wellness and shopping rewards, built from permissioned, organic spending across tens of thousands of merchants and brands, with signals refreshed daily.

The same data was already inside Amazon, at the far end of the campaign. Attain says it became one of the first automated offline sales lift measurement partners of the former Amazon DSP. Sales lift measurement compares what an exposed group bought against a comparable unexposed group and attributes the gap to the ads. Records that used to grade a finished campaign now pick who the campaign reaches.

What you can inspect and what you have to trust

Pull the feature apart into inputs, decisions and outputs. The decisions and the outputs sit in front of you in the buying tool, with the segment you assembled, the inventory it ran against, and the delivery and sales that followed. The inputs are where the questions start. The announcement names no app, gives no pricing, no segment count and no list of the retailers in its taxonomy, and it does not say whether transactions come from linked card or bank accounts, scanned receipts, connected email or a mix. The word organic goes undefined. Attain's own description wobbles on speed, calling signals daily in the body and real-time in the company description at the end.

That gap matters more here than with an Amazon-built audience, because what you are buying is coverage you have no way to verify. A segment of people who bought a competitor's product at a particular grocery chain is only as true as the capture method behind it, and the capture method is the part you are asked to take on faith. Vendor-side evidence has the same shape. A Trade Desk post published on October 5, 2026 reported that 90% of campaigns beat Attain's own benchmarks for new-buyer ROAS by an average of 122%, return on ad spend from first-time buyers, although that post gave no study period, holdout design or campaign count. Treat it as a vendor benchmark until someone publishes the holdout.

A read on who this serves

Amazon's gain here reads as pulling more spend into its own inventory, and specifically money from advertisers who sell nothing on Amazon. Attain pitches the integration at QSR, travel, financial services, entertainment and automotive brands, categories whose transactions Amazon's retail data never sees. Those advertisers have always been able to buy Prime Video inventory. What they lacked was a defensible way to choose which viewers. Native third-party purchase audiences supply one, without anyone carrying data in from a separate system. That read comes from the timing and from Attain's stated target list of advertisers running campaigns across third-party inventory or Amazon's owned channels. Amazon has said nothing about its own intent, and no Amazon Ads executive is quoted in the announcement.

For brands that do sell on Amazon, Attain's case is competitor visibility. The company says the integration unlocks more visibility, allowing advertisers to discover and target audiences based on a broader view of shopping habits, competitor dynamics, and customer loyalty. Put plainly, Amazon's purchase signals stop at the edge of its own store. A shopper who refills a rival brand every six weeks at a drugstore leaves a trail Amazon has no way to sell you.

Who should actually test this

The fit is narrow, and worth being honest about. If your category's rival purchases mostly happen off Amazon, in grocery, beauty, household CPG or anything with heavy shelf presence, a conquesting audience built on off-Amazon purchases gives you something Amazon's own signals cannot construct. If your category lives almost entirely on Amazon, this adds a third-party dependency and a narrower pool in exchange for a signal you may already approximate with on-platform purchase and view audiences. Attain published no pricing with the announcement, so the cost side of that trade is not something anyone outside the negotiation can size yet.

There is a published Sellrbox case study worth naming here for what it shows about one account, not as a rule for yours. Over 12 months on a 10,529-SKU industrial equipment catalogue in the US marketplace, the Astra-run campaigns held 17.8% ACOS while the owner-run campaigns in the same account ran at 48.6%, advertising cost of sales as reported there, with the owner hands-on about 12 minutes a week. One account, one category, two ways of running the same spend, and a wide distance between the halves. That says nothing about what an off-Amazon purchase audience would do for anybody. It is a reminder that a controlled comparison inside one account tells you more than an account average does, which is the shape the Attain test needs too. If you want to run that kind of side-by-side yourself, the self-service setup is where that work sits.

The grading problem to name before you buy

If you select the audience with Attain segments and then measure offline sales with an Attain study, the campaign would be graded by the same source that chose who saw the ads. Nothing in the announcement says advertisers must run it that way, and the same question already applies to retail media networks that sell an audience and then report the sales that audience produced. The fix is procedural. Decide now which source grades the test, and make it a different one from the source that built the audience.

A four-week test, and what to watch

Treat the first run as a narrowing test with a budget you can afford to see underspent. Pick one line where you know the offline competitor by name, build a single Attain segment for buyers of that competitor at one or two retailers, and run it against a comparable Amazon-built in-market or purchase audience on the same creative and the same inventory type. Compare cost per new-to-brand order over four weeks, and watch delivery volume as closely as you watch efficiency.

My read, taken from the three-variable segment structure Attain describes rather than from any reach figure it published, is that stacking retailer, category and channel on a 14 million person panel will narrow addressable reach quickly. Attain gave no reach numbers, so check the forecast in the buying tool before you commit a budget to it, and expect a segment that cannot spend to teach you very little either way.

Two things to confirm before you plan around it. PPC Land reports that which DVA+ tier can use them is unclear, so ask your account team whether your buying tier has the segments at all. Then watch the overlap with Amazon's own offline work, since on July 7, 2026 Amazon began breaking offline sales down by the individual retailer where they happened at study and campaign level in the United States. For the wider context on how DVA+ and Amazon's agent-driven ad tooling are being assembled, our deep dive on Amazon unBoxed 2025 walks through the full-funnel direction this integration slots into. If you would rather have someone else run the test and the reporting, the managed service is built for that.