Amazon Sponsored TV — the self-serve streaming ad product that opened to all sellers in late 2024 and matured through 2025 — has become the most over-pitched and least-understood ad placement in 2026. Account managers love it because the CPMs look reasonable next to traditional CTV. Brands try it because Prime Video ads went broad and the inventory feels fresh.
After running Sponsored TV campaigns for dozens of brands across 2025, here is the operator read for 2026.
Most brands doing under $250K/month on Amazon should not run Sponsored TV. The brands that should run it are doing $400K+/month, have a hero ASIN with proven CVR, and a creative they can produce in 15 and 30 second cuts without burning their P&L. Get those three boxes checked and Sponsored TV becomes one of the cheapest top-funnel levers in the ecosystem.
Get them wrong and you will spend $8K-15K to learn what your search ads already told you.
What Sponsored TV Actually Is In 2026
Sponsored TV runs across Prime Video, Amazon Freevee, Twitch, and Amazon’s network of third-party streaming publishers (IMDb TV partners, Fire TV channels, syndicated CTV inventory). It’s a self-serve, non-skippable streaming video ad purchased on a CPM basis through the Amazon ad console — the same console you use for Sponsored Products.
The 2026 version differs from launch-year STV in three ways:
- Lower minimum spend — effectively no enforced floor for self-serve (DSP-managed STV still has $50K+ minimums)
- Better targeting — lifestyle, in-market, and ASIN-remarketing audiences are now native, not workarounds
- Tighter attribution — Amazon Marketing Cloud reporting plus the new on-Amazon conversion lift report make the attribution story cleaner than any other CTV product on the market
CPMs in 2026 sit in the $28-$45 range for most categories on Prime Video inventory. Freevee runs $15-$25. Twitch runs $12-$22 but the audience is narrow.
The Qualification Math
Before any brand spends a dollar on Sponsored TV, three thresholds need to be hit.
Threshold 1: $400K+/month Amazon revenue, or DTC + Amazon combined $750K+/month. Below that, the awareness-to-conversion lag (typically 7-21 days for STV) eats too much working capital relative to the lift. Search ads still have 3-5x better short-term ROAS at that revenue stage.
Threshold 2: At least one ASIN with proven CVR > category benchmark and > 200 daily sessions. STV does not fix a bad listing. It funnels strangers to it. If the ASIN converts at 8% when the category average is 12%, STV will accelerate losses, not revenue.
Threshold 3: A 15-second and 30-second creative cut that meets Amazon’s TV ad spec. Production cost for a usable cut runs $4K-12K if you don’t have one. If you can repurpose existing video assets, $1K-3K. If you can’t get to a watchable creative under that budget, skip STV entirely — bad TV creative on premium inventory is a brand-damage outcome, not a neutral test.
Brands that hit all three thresholds and skip STV are leaving cheap reach on the table. Brands that fail one or more and run STV anyway are the ones generating the “STV doesn’t work” posts on Reddit.
Creative Requirements That Matter
Amazon’s ad spec is published. Real-world: most rejected STV creatives fail on three specific issues.
- Text-on-screen readability under 1.5 seconds — anything that requires reading kills lift on a TV ad. Voice-over or supers, not paragraph copy.
- Branding within the first 3 seconds — logo or product visible by 0:03. STV viewers don’t watch a 15-second build-up.
- Clear CTA in last 3 seconds — “Search [brand] on Amazon” or “Shop on Amazon” overlaid in the final frames. Brands that omit this lose the cross-channel attribution lift entirely.
The 15-second cut is the workhorse. 30-second cuts underperform 15s on completion rate and equal-or-worse on lift in our 2025 data. Run 15s primary, 30s as a secondary frequency control.
The format that actively wins: product-in-use opening shot, problem-or-outcome callout at 0:04-0:08, branded close at 0:11-0:15. Lifestyle openings without product-in-frame underperform by 30-40% on lower-funnel lift.
Audience Strategy: The Three Tiers That Work
After running 60+ STV campaigns, three audience structures consistently outperform.
Tier 1 — Lifestyle and in-market. Amazon’s native lifestyle audiences (e.g., “Home Cooks,” “Outdoor Enthusiasts”) and in-market audiences (e.g., “In-Market for Skincare”) are the cheapest reach in the system. Run 60-70% of STV budget here. CPMs are highest here too but the addressable population is largest, so frequency stays sane.
Tier 2 — ASIN remarketing. Audiences built off shoppers who viewed your ASIN, viewed your category, or viewed competitor ASINs in the last 30-90 days. Run 20-30% of budget here. CPMs are higher but lower-funnel lift is 2-4x lifestyle audiences.
Tier 3 — Custom audiences from AMC. Audiences built off purchase patterns or cross-category behavior in Amazon Marketing Cloud. Run 10% of budget here as a test bucket. These are powerful when they work but the build cost (analyst hours) is high.
The structures that don’t work in 2026: broad demographic targeting, single-ASIN remarketing without lookback windows, and “Amazon shopper” generic audiences. All three burn budget on people who were going to buy anyway.
Attribution: What To Actually Measure
The wrong way to measure Sponsored TV is by ROAS in the ad console. The console reports last-touch on-Amazon attribution. STV is upper-funnel. Last-touch will always look weak.
The right way:
- Brand search lift — measure branded keyword search volume in Brand Analytics during the campaign vs. the prior 4-week baseline. Expect +15-40% lift on a properly-running STV campaign.
- New-to-brand percentage — STV should drive NTB rates of 60-80% on the impacted ASINs. Below 50% and the audience is over-targeted to existing buyers.
- Detail page view lift — DPV on STV-promoted ASINs should rise 20-50% during the campaign window.
- AMC conversion lift report — the most defensible attribution. Run it 30 days into the campaign minimum. Below 10% incremental lift, the campaign is washing.
Brands that measure on console ROAS only will kill working STV campaigns. Brands that measure lift will scale them.
Common Failure Modes
Five patterns we see fail repeatedly.
- Running STV before search is saturated. If your branded SP campaigns are still scaling efficiently, you have cheaper reach available. STV is a saturation play, not an early-stage play.
- Single-ASIN STV. Running STV to one ASIN concentrates risk. The campaign lives or dies on that ASIN’s CVR. Run STV to a brand-level audience that lands on a curated storefront page instead.
- Premature kill. STV needs 21-30 days to stabilize. Brands that pull plugs at day 10 because ROAS looks bad are reading the wrong scoreboard.
- Reusing DTC video without reformatting. A 30-second YouTube ad with text-heavy supers is a bad STV ad. The format, pacing, and supers all need recutting.
- Frequency capping at default. Default frequency caps create over-served audiences. Pull the cap to 3-4 impressions per user per week and reach widens significantly.
Sponsored TV vs. Sponsored Brand Video: Which First?
Brands that don’t yet run Sponsored Brand Video should run SBV first, then STV. SBV is on-platform, last-touch attributable, and converts in days, not weeks. SBV is also cheaper to test — same creative requirements, lower CPMs, faster feedback loop.
If SBV is already running profitably and your hero ASIN is healthy, STV is the next layer. If SBV is not running or not profitable, STV will not save you.
FAQ
What’s the minimum monthly STV budget that’s actually viable?
$8K-12K/month minimum for self-serve. Below that, frequency falls under 2 impressions per user per month and lift becomes statistically invisible. The brands seeing the cleanest lift are spending $20K-50K/month on STV.
Does Sponsored TV work for brands without a national presence?
Yes — STV supports geo-targeting at the DMA level. Regional brands can run STV in their served markets and avoid wasted reach. This is one of STV’s underrated advantages over linear TV.
How long until I see results from STV?
Day 21 minimum. Day 30 for a clean read. Brand search lift is the first signal (week 2). DPV lift follows (week 2-3). NTB orders catch up (week 3-4). Faster reads will mislead.
Can I run Sponsored TV without DSP?
Yes — that’s the entire point of self-serve STV. You do not need an Amazon DSP seat. DSP unlocks more advanced audience tools and lower minimums on managed-service buys, but the self-serve console handles 80% of brand use cases.
Is Sponsored TV worth it for a brand doing $150K/month?
No. Run Sponsored Brand Video, fix your hero image, scale Sponsored Products, and revisit STV at $400K+/month. Brands at $150K/month running STV are usually doing it because their agency wants to expand the retainer scope, not because the math supports it.
If you’re looking for a team that manages every lever — creative, advertising, and operations — Velocity Sellers works with brands doing $100K+/month on Amazon. Contact us for a free account audit.
