How to Measure Streaming TV When There Is No Click
Non-skippable streaming video doesn't produce a click to attribute, so the honest measurement stack is completion rate, branded search lift, and Amazon Marketing Cloud's view of downstream Sponsored Products engagement — not a last-click conversions column. Our own practice sets streaming a return target near 1.0x rather than a retargeting-level bar, precisely because standard attribution structurally undercounts what this inventory actually does.
What this looks like in a real account
Why last-click attribution fails this format specifically
A search ad gets clicked, the click gets tracked, the sale that follows within an attribution window gets credited. Full-screen streaming video has none of that mechanism — a viewer watches, doesn't click anything, and may not search or buy for days, if the influence shows up in a trackable way at all. Amazon's own shoppable and interactive features on this inventory generally require a logged-in Prime member to register an action, which means a meaningful share of genuine influence happens completely outside what any dashboard can attribute directly. Judging this inventory on the same conversions column used for search is measuring it with the wrong instrument, not measuring it honestly.
What to measure instead
Completion rate is the first, fastest signal — it tells you whether the creative is holding attention before the sales data is even trustworthy enough to read. Branded search lift is the second: Amazon's own published Fellow case study showed a 500% lift in branded search after adding streaming TV to an existing online-video strategy, with 90% of streaming-influenced purchasers new to the brand, a 75% higher new-to-brand rate than online video alone produced. Amazon Marketing Cloud is the third and deepest layer — a clean-room environment that can join DSP exposure data with downstream Sponsored Products engagement and purchase signals, supporting linear, time-decay or position-based attribution models instead of forcing everything through last-click.
Amazon's own reporting on Interactive Video Ads specifically adds a fourth layer where those formats are eligible: 6x higher brand searches, 4x higher product page views, 4x higher add-to-cart actions and 5x higher purchase rates compared with standard streaming TV campaigns carrying the same objective. Where interactive formats are available and in budget, they generate genuinely richer engagement signal than static or standard video, precisely because the interaction itself — an add-to-cart inside the viewing experience — is a trackable event that standard video never produces.
A worked example
Set a streaming line item a return target of roughly 1.0x rather than the 4–6x a retargeting campaign might carry — a practice we apply across our own managed DSP accounts specifically because last-click attribution structurally undercounts this inventory. Pair that target with two supporting reads: completion rate against the industry benchmark (premium streaming and connected-TV video generally runs 90%+, with non-skippable pre-roll on some ad-supported CTV inventory reported as high as 98.6%), and branded search volume in the two weeks following the flight, compared against a pre-flight baseline. A campaign that hits 0.8x on last-click attribution, 92% completion rate, and a genuine lift in branded search is working — the ROAS number alone would have said otherwise.
What to do when the numbers still look bad
If completion rate is weak, that's a real, trustworthy signal — fix the creative before touching anything else, since a low completion rate means the ad isn't being watched regardless of what any downstream metric later shows. If completion rate is strong but branded search shows no movement at all, give it more time before concluding the campaign failed. On one account we watched, organic order share rose from 13.5% before video to 27.9% after, and was still climbing four weeks after spend had actually stopped — one account, not a guarantee, but a reminder that the lift window can run longer than a two-week read allows for. If Amazon Marketing Cloud shows genuine downstream Sponsored Products engagement but the DSP dashboard still reports a weak ROAS, trust the AMC read over the DSP dashboard — that's exactly the gap this measurement stack exists to close, and it's the reason a single-source view of this inventory is incomplete by design, not by mistake.
The common mistake
The mistake is pulling a streaming line item after two weeks because the DSP dashboard's attributed ROAS looks bad, without ever checking completion rate or branded search. It's an easy mistake because the dashboard number is right there and everything else takes extra work to pull. reMKTR builds the completion-rate and branded-search check into the standard weekly report for every streaming line item from day one, specifically so the DSP dashboard's incomplete number never becomes the only number a client or stakeholder sees.
The reverse mistake exists too, and it's worth naming honestly: treating every weak-ROAS streaming line item as automatically fine because "attribution undercounts it" is a convenient excuse that can shield genuinely broken creative or badly mistargeted audiences from real scrutiny. The measurement stack above exists to give a fuller picture, not an automatic pass — a campaign with weak ROAS, weak completion rate and no branded search movement is not being undercounted, it's underperforming.
| Signal | What it tells you | When to trust it |
|---|---|---|
| Completion rate | Whether the creative is holding attention | Immediately — available from day one |
| Branded search lift | Whether the campaign is building demand | After 1–2 weeks, against a pre-flight baseline |
| Amazon Marketing Cloud | Downstream Sponsored Products engagement, custom attribution | Once enough volume has accumulated for a reliable read |
| Last-click DSP ROAS | Only the sales this inventory can directly claim | Directional at best — expect it to understate real influence |
Which one you should actually pick
This measurement approach suits any brand running non-skippable streaming inventory and reporting to a stakeholder who expects a same-week sales number by default. A brand that only cares about last-click-attributable sales, and has no interest in top-of-funnel reach, is better served putting the budget into a clickable, directly-attributable format instead — streaming was never built to satisfy that specific question.
Shortlist on the job, not the feature grid. Pull your search-term report for the last 90 days and total the spend against terms that produced no orders — 33.6% on the account above. Then ask each vendor on your list what they would do about it in week one, and see who answers with a process rather than a screenshot.
Common questions
What's a good completion rate for streaming TV creative?
Industry benchmarks put premium streaming and connected-TV completion rates at 90% or higher, with some ad-supported CTV pre-roll reported as high as 98.6% — well above skippable, feed-based video, where completion typically runs far lower.
How long should I wait before judging a streaming campaign's branded search impact?
At least two weeks, and ideally longer — on one account we've watched, organic lift from video exposure kept building for weeks after spend even stopped, which is a reason not to draw a conclusion from the first few days.
Do I need Amazon Marketing Cloud to measure streaming TV properly?
Not to get a directional read — completion rate and branded search lift are available without it — but AMC is the only environment that can join DSP exposure with downstream Sponsored Products engagement and apply attribution models beyond last-click, so it's worth setting up once volume justifies it.
Why does reMKTR use a 1.0x return target for streaming instead of a higher bar?
Because last-click attribution structurally undercounts non-clickable, top-of-funnel inventory — a higher target would judge the format against a measurement system that was never going to capture most of what it actually does, and would lead to cutting working campaigns based on an incomplete read.
Does this measurement approach apply to Fire TV and Twitch the same way it applies to Prime Video?
Yes, in principle — none of Amazon's non-skippable streaming inventory produces a native click, so the same completion-rate, branded-search and Amazon Marketing Cloud stack applies across supply sources, though interactive-format benchmarks specifically are tied to the formats where interactivity is actually available.
We show the method before the number.
Claim the free auditRead next
- Skai Pricing: Five Published Tiers, Read CloselyPricing · skai pricing
- Podean Reviews: Evaluating a Partner Without StarsReview · podean reviews
- Pacvue Pricing: No Public Number — What to AskPricing · pacvue pricing
- Skai vs Pacvue: Contracts, Not Feature GridsHead to head · skai vs pacvue