CTV Performance Sequencer
Connected TV finally made television addressable and measurable — CTV upfronts now outsize primetime linear — but most brands still buy it like a linear reach spot and hope. The performance method is a sequence: prove the creative cheaply where signal is fast, scale only the winner on the big screen, then retarget the warmed audience down-screen where conversion is natural. Tap each step to see the screen, the buy, the creative, and the metric that matters.
Run CTV as four moves, not one buy. Test & learn in social to find winning videos for cheap; prove the winner against significance so you scale a fact, not a hope; scale on the big screen with the proven cut on non-skippable CTV, starting programmatic and buying direct once the data proves out; then retarget down-screen with nurture sequences and measure the whole thing by incremental lift against a holdout. The living room builds demand; the small screen captures it.
CTV performance sequence
How to use this sequencer
- Test & learn in social. Run many genuinely different creative variations cheaply in feeds — you are buying signal, not reach.
- Prove the winner. Read against statistical significance and keep only the one or two videos that actually moved response.
- Scale on the big screen. Put the proven winner on non-skippable CTV; start on curated programmatic, then buy direct once the data proves it out.
- Retarget down-screen. Re-engage the warmed audience with proven nurture sequences on smaller screens, where tapping through is natural.
- Measure incrementally. Judge the whole sequence by a holdout, not the platform’s completion tally — completion is high by default on non-skippable inventory.
RGM Expert Says
The mistake we fix most often on CTV is treating it as a brand-awareness line you cannot hold accountable. It is the opposite: the most addressable, most measurable television that has ever existed — if you run it as a sequence instead of a single reach buy. We never let an unproven video onto the expensive screen. We prove it in social where reads are cheap, scale the winner on the living-room screen for impact, and then do the part most brands skip entirely: retarget that warmed audience down-screen with proven nurture and measure the conversions against a holdout. Test in the cheap seats, perform on the big screen, convert on the small one. The brands that win CTV are not the ones spending the most on premium inventory; they are the ones who earned the premium impression with a proven creative and proved the lift before they scaled.
How the method works
Each stage has a distinct job, screen, and buy type, and they only work in order. Testing lives in social because the cost of a read is low and the volume is high, so you can find a real winner fast. The proof gate exists because scaling a lucky video onto CTV is the single most expensive error in the channel. Scaling lives in the living room because that is where a proven story lands with the most impact and the highest completion — and you start on curated programmatic to control quality, then buy direct once the data justifies locking in inventory. Retargeting lives on smaller screens because that is where a warmed viewer can actually tap through and convert. The connective tissue is measurement: attribution steers the day-to-day, but a geo or audience holdout is the only honest read of what CTV caused, because completion on non-skippable inventory is high no matter how good the ad is.
The part most brands skip
The down-screen retarget is where the sequence quietly makes its money, and it is the step almost everyone drops. A brand will spend handsomely to put a polished spot on the living-room screen, watch the completion rate come back at 95 percent, and call it a success — without ever re-engaging the people that exposure just warmed. That is leaving the conversion on the table. The warmed viewer is far more valuable than a cold one, but they will not get up from the couch to buy; you have to meet them later, on the phone or the laptop, with a message that picks up where the big-screen story left off. Run the two as one connected sequence, retarget the exposed audience with proven nurture creative, and measure the lift end to end — that is the difference between CTV as an expensive brand gesture and CTV as a channel that pays for itself.
Why CTV beats linear — if you measure it
For the first time, connected-TV upfront commitments exceed primetime linear, and a clear majority of marketers say CTV targets better than linear. The reason is structural: CTV is addressable, so you reach a defined audience rather than a daypart, and it is measurable, so you can tie exposure to outcomes. But that advantage is potential, not automatic. Only about 45 percent of CTV advertisers run incrementality testing, which means most are still grading the channel on completion rates and reach — metrics that look great on non-skippable inventory and prove almost nothing. The brands pulling ahead treat CTV as a performance channel with a brand-sized halo: they sequence creative from cheap to premium, they prove the lift with holdouts, and they reinvest against incremental results rather than against the platform’s self-reported tally.
CTV, at a glance
Reference ranges for planning a connected-TV buy. Figures vary by inventory quality, targeting, and industry; the only trustworthy read of performance is your own holdout.
| Metric | Typical range | Note |
|---|---|---|
| CTV CPM | ~$25–$45 | By industry roughly $18 (CPG) to $50 (B2B/SaaS) |
| Completion (VCR) | ~94–97% | Premium AVOD; mid-roll the highest slot — high by format |
| Pre-roll / FAST completion | ~85–93% | Lower than mid-roll, varies by inventory |
| CTV vs mobile recall | ~+20% | Higher brand recall than mobile video |
| Advertisers using incrementality | ~45% | The discipline gap that is an edge |
What CTV performance leaders emphasize
The measurement shift — from attribution to incrementality — is what separates brands who think CTV underperforms from brands who can prove it works.
Test in the cheap seats, perform on the big screen, convert on the small one — and never scale a video the data has not already proven.