WayToGlobal Insights · Emerging Media

WayToGlobal Insights · 新兴媒体

Roku + Amazon DSP: Shoppable CTV Reaches 80M Households (2026 Guide)

Roku × Amazon DSP:8000 万家庭的可购物 CTV

2026-10-29 · 9 min read · checked against primary sources

2026-10-29 · 约 9 分钟读完 · 逐条核对原始来源

All insights全部文章

30-second read

  • Amazon DSP buys Roku with logged-in targeting and closed-loop measurement across ~80M US households.
  • The first CTV buy that behaves like performance marketing — if you sell on Amazon.
  • Brand reach and attributable conversions in one line item; entry via DSP bidding.

Connected-TV advertising used to be a big-brand game. The 2026 change is the Amazon Ads x Roku deal going live: Roku's CTV inventory is now buyable inside Amazon's DSP, where logged-in identification means the platform knows who is watching — an estimated combined reach of ~80 million US households, about 80% of CTV homes. For Chinese DTC brands selling into the US — especially those with an Amazon storefront — this is the first time big-screen reach and ecommerce-grade attribution come in one DSP.

Why logged-in matters

Linear TV's three fatal gaps: no knowledge of who watched, whether they finished, or whether they bought. The Amazon-Roku integration addresses each: logged-in devices carry Amazon shopping profiles; frequency and completion are visible in the DSP; and when the viewer has Amazon purchase history, the exposure-to-order loop actually closes. Industry data backs the performance direction — Amazon Ads has reported interactive video driving 6x brand search and 4x product-detail-page views.

Why now

Two reasons. The integration landed in Q4 2025, inventory consolidation is still settling, and reported pricing undercuts what the inventory fetched separately — early buyers arrive before price discovery completes. And CTV is at a shoppable inflection: Samsung launched its own shoppable CTV experience in June 2026, and the industry is reclassifying "TV advertising" from brand spend into measurable media while the rules are unwritten. Poor fits: low-AOV categories that can't carry CPMs (big-screen units price above feeds), and teams whose US presence or attribution chain (Amazon store or site) isn't wired up.

Three TV gaps, closed
Three TV gaps, closed

Getting started

Qualification first: brands with an Amazon store or Brand Registry get the closed loop directly; pure-site brands can buy, but attribution runs through Amazon's purchase-signal mapping at lower precision. DSP onboarding goes through Amazon Ads or certified partners, with budget floors (industry convention: several thousand dollars monthly). Targeting trio: logged-in category browsers, owned audiences (your Amazon followers and buyers for retargeting), and content-context dayparts. Creative to big-screen standard: 1920x1080 at 15/30 seconds, brand and benefit inside five seconds, an explicit actionable at the end. Budget and cadence: a few hundred to a thousand dollars a week across 3-4 weeks, aligned to promo beats.

Risks and measurement

Three risks. Attribution stratifies: audiences with Amazon purchase signals close the loop; site-based attribution is inferred mapping — keep them in separate buckets. The scale trap: CTV delivers big reach at low per-exposure value; judged by feed-CPA standards it always "looks expensive" — judge it on incremental households reached times branded-search lift instead. Creative cost: one proper 30-second spot is a real investment; TV-grade templates over product footage cut the bill. Measure with a pre-launch baseline on branded search and direct traffic, in-flight reach/frequency/completion in the DSP, and post-launch pulses on Amazon store visits and branded search — the timeline has to line up across all three. Set expectations at quarterly brand investment, not a monthly performance channel.


Sources: eMarketer / Modern Retail (deal announced June 2025, live Q4 2025); Advanced Television (Samsung, June 2026; Amazon interactive-video figures). Verified 2026-10-29.

Official numbers: the authenticated-audience scale

Official case page screenshot
Screenshot of the official case page, taken 2026-10-29

Roku's own ad chief put the Amazon partnership authenticated reach at ~80% of the US CTV footprint (80M+ logged-in households, Beet.TV, December 2025), and Roku's advertising site publishes DTC-focused case resources on using streaming TV alongside performance marketing with ACR viewing data. Named-brand case pages are thinner here than on social platforms — a sign of the channel enterprise roots, and of the whitespace.

Sources: advertising.roku.com; Beet.TV interview with Roku Jeff Katz (December 2025).

Series: all ten emerging-media platforms compared on one page — Emerging Media Buying Guide 2026. 系列:十个平台一张对照表——《2026 新兴媒体投放总览》。

Selling into the US? Let's map your CTV entry.

Need an Amazon DSP onboarding partner — store linkage, CTV setup and the first closed-loop campaign? We do this end to end. Leave your product category, price point and current channels — the account comes next.

30 秒看完

  • Amazon DSP 买 Roku:登录态定向+闭环归因,覆盖约 8000 万美国家庭。
  • 第一种「按效果买」的 CTV——前提是你在 Amazon 有店。
  • 品牌触达和可归因转化合成一条投放;走 DSP 竞价进场。

CTV(联网电视)广告过去是 big brand 的游戏,2026 年的格局变化是 Amazon Ads 和 Roku 的合作落地:Roku 的 CTV 库存进了 Amazon DSP,Amazon 的登录态识别能认出谁在看——预估合并触达约 8000 万美国家庭,占美国 CTV 家庭的八成。对做美国市场的中国 DTC 品牌(尤其有 Amazon 店的),这是第一次「大屏曝光+电商级归因」能在一个 DSP 里同时买到。

为什么登录态是关键

传统电视广告的三大死穴:不知道谁在看、不知道看没看完、不知道看了买没买。Amazon DSP 买 Roku 库存的三件事逐一解掉:登录态意味着设备后面是个有购物画像的具体人(Amazon 账号体系);频次与完播数据在 DSP 里可见;最关键的——如果这个人在 Amazon 上有购物历史,广告曝光到下单的归因闭环能对上。行业数据也在给 CTV 效果化背书:Amazon Ads 公布过互动视频带来 6 倍品牌搜索和 4 倍商品详情页浏览的增量口径。

现在进场的理由

两条。一是合作刚落地(2025 年 Q4 起),库存整合期的价格据报道低于各自单独售卖——早期买家在价格发现完成前进场。二是恰逢 CTV 购物化拐点:Samsung 等厂商 2026 年也在推可购物 CTV,行业正把「电视广告」从品牌支出重分类为可测媒介,规则未定型,先手有定义权。不适合的:客单价低、毛利率撑不起 CPM 的品类(大屏曝光单价天然高于信息流);以及美国市场占比很低、归因链路(Amazon 店或独立站)没铺好的团队。

大屏广告三个死穴的解法
大屏广告三个死穴的解法

怎么起步

第一步,确认资质路径:有 Amazon 店铺/品牌备案的最顺——Amazon DSP 的归因闭环直接吃到;纯独立站品牌也可投,但归因走 Amazon 的购买信号映射,精度打折。DSP 开户通常走 Amazon Ads 官方或认证代理,有预算门槛(行业行情约每月数千美金起)。第二步,定向三件套:登录态人群(浏览过你品类的人)、自有人群(你的 Amazon 关注者/购买者做再营销)、兴趣场景(品类相关内容时段)。第三步,素材按大屏做:1920×1080 的 15/30 秒视频,前 5 秒品牌+利益点双出,结尾放明确的「可行动」信息(Prime 会员可购/官网直达)。第四步,预算与周期:周预算几百到一千美金级,覆盖 3-4 周一个完整测试;和促销档期对齐。

风险与验数

三个风险。一是归因精度分层:有 Amazon 购买信号的人群闭环准,独立站归因是映射推断——分开记,别混。二是量级陷阱:CTV 的曝光量大、单次点击价值低,用信息流的 CPA 标准评它会「看起来很贵」,用「新增触达家庭×品牌搜索增量」评才对。三是素材成本:大屏素材制作门槛高,一条 30 秒的专业视频是硬投入,用电视级模板+产品实拍混剪可以压成本。验数:投前记品牌词搜索与直接访问基线,投中看 DSP 内的 reach/频次/完播,投后看 Amazon 店铺访问(如有)与品牌搜索的脉冲——三段对上时间线才算数。定位预期:季度级品牌投资,不是月度效果渠道。数据出处:eMarketer/Modern Retail(合作 2025-06 宣布、Q4 2025 落地)、Advanced Television(Samsung 2026-06;Amazon 互动视频口径)。主会话 2026-10-29 查证。

官方口径:登录态触达的规模

官方案例页截图
官方案例页实拍(2026-10-29 截图)

Roku 广告负责人亲口把 Amazon 合作的登录态触达定为美国 CTV 家庭的 ~80%(8000 万+登录家庭,Beet.TV 2025-12);Roku 官方广告站发布面向 DTC 的案例资源——用流媒体电视配合效果营销、以 ACR 观看数据定向。具名品牌案例页比社交平台少——这既是渠道的企业级出身,也是空白所在。

出处:advertising.roku.com;Beet.TV 访谈 Roku Jeff Katz(2025-12)。

Series: all ten emerging-media platforms compared on one page — Emerging Media Buying Guide 2026. 系列:十个平台一张对照表——《2026 新兴媒体投放总览》。

在做美国市场?我们帮你规划 CTV 进场。

需要 Amazon DSP 代开户与陪跑(店铺打通、CTV 搭建、首个闭环 campaign)?全流程我们做。 留下品类、客单价和当前渠道,账户随后就开。