SaliencyLab vs System1 Test Your Ad.
Two methodologies, two different moments in the workflow.
System1 built one of the most important academic stories in modern advertising, Les Binet and Peter Field on the long and the short of it, fluent devices, emotional priming. SaliencyLab is the tool for the decision before the Star Rating is commissioned. Here is what each does, written by someone who spent six years inside the buying side.
System1 Test Your Ad and SaliencyLab solve different parts of the same problem. System1 measures emotional response on a small consumer panel using FaceTrace, then maps that response to a 1.0–5.9 Star Rating calibrated against the IPA effectiveness dataset. The Star Rating predicts long-term brand effects. SaliencyLab is a 90-second pre-spend diagnostic that scores five frozen KPIs against public engagement and click-intent outcomes, and issues a Scale, Sharpen, or Rebuild verdict.
System1 measures emotion on real people. SaliencyLab predicts engagement from the creative itself. If you need an emotional readout backed by panelist faces, you need System1. If you need a fast verdict between edit rounds, you need SaliencyLab. Many teams run both.
The comparison, on what actually moves a decision.
Where System1 is stronger, the table says so. Where SaliencyLab is stronger, it says so. Where the answer is "they do different things", that is what it says.
| Dimension | SaliencyLab | System1 Test Your Ad |
|---|---|---|
| Decision moment | Pre-spendRun between edit rounds, before the media buy | Pre-launchValidate after the creative is largely locked |
| Time to verdict | ~90 seconds per ad | ~24–48 hours (rapid panel field + emotional coding) |
| Price per ad | Included in Pro plan, no per-test charge | ~$5,000–$25,000+ per ad (industry estimates; not publicly listed) |
| Sample / cohort | 1,200+ ad calibration pool with public outcome data | ~150-person consumer panel per test, IPA-effectiveness-calibrated cohort behind the scale |
| Emotional measurement | NoNot measured; not claimed | YesFaceTrace coding of seven emotions on panelist faces |
| Headline output | Composite 0–100 → Scale / Sharpen / Rebuild verdict | 1.0–5.9 Star Rating (long-term brand effect prediction) |
| What it predicts | Engagement + click intent (TikTok, YouTube) | Long-term brand effects (calibrated on IPA dataset) |
| Held-out OOS performance | Spearman ρ +0.30 to +0.32 (engagement, click intent) | Published validation against IPA Effectiveness Awards dataset |
| Methodology transparency | Five frozen KPIs, fixed weights, mechanical verdict logic published | Methodology published in academic papers and books; scoring engine proprietary |
| Sales / ROAS prediction | NoExplicitly out of scope | IndirectStar Rating maps to long-term effects, not direct sales |
| Iteration cycles | Designed for re-running after every cut | Round-trip per test; rapid but not real-time |
| Self-serve | YesUpload, 90 seconds, verdict | HybridSelf-serve portal + managed engagements |
| Best for | DTC, performance teams, agencies iterating between rounds | Brand teams validating emotional craft before a hero launch |
Why this is not a stranger's opinion.
EEAT, Experience, Expertise, Authoritativeness, Trustworthiness, on creative testing. Here is the receipt for this one.
Oussama Nakhil
I have read more System1 reports than I can count. The 5.9-star scale was a fixture of the L'Oréal long-term-effects conversation, sitting alongside Kantar Link AI on the wall of pre-launch evidence. I respect that work. It is grounded in academic research most marketers have never actually read.
This comparison is not "System1 is wrong". It is "System1 measures something specific, beautifully, on a real panel, in 24–48 hours, and pre-spend creative iteration happens in 90 seconds, on every cut, before any panel is commissioned". Different tools, different decisions.
Two tools. Two different stages of the same workflow.
SaliencyLab between every edit round. System1 once, before the hero asset launches. That is how a serious brand team uses both.
Use SaliencyLab when
- You are iterating between cuts and need a verdict in 90 seconds.
- You have ten variants to triage before sending one to a panel.
- You want a deterministic Scale / Sharpen / Rebuild call grounded in five frozen KPIs.
- You need pre-spend evidence on creatives whose budget cannot justify a panel test.
- You want held-out OOS validation numbers you can point at on the page.
Use System1 Test Your Ad when
- You need emotional response measured on panelist faces, not predicted by a model.
- The story you are telling the brand team is about long-term effects (Binet/Field).
- The hero asset budget justifies a $5k–$25k+ test before launch.
- The 5.9-star scale is the language the room already speaks.
- You want a calibration against the IPA Effectiveness Awards dataset.
What we have validated, and what we have not.
Both companies publish numbers. Here are ours, in the same format you would expect from any serious insights vendor.
What SaliencyLab does not claim
We do not measure emotion. We do not record viewers. We do not predict long-term brand effects in the Binet/Field sense. We do not predict sales lift, ROAS, attributed conversion, or brand recall. Heatmaps are predicted visual attention, not measured eye tracking. Synthetic interviews via BuyerLens are scenario simulation, not a real consumer panel. We say this on every page that produces a score.
System1 does measure emotion on real panelists, does report the Star Rating, and does calibrate against the IPA dataset. Those are real, useful, validated outputs. They take longer to produce, cost more per test, and answer a different question. Both can be true.
Questions teams ask before they choose a tool.
Is SaliencyLab a replacement for System1 Test Your Ad?
Does SaliencyLab measure emotion like System1?
How is the SaliencyLab composite different from the System1 Star Rating?
How much does each cost?
Why should I trust the SaliencyLab score?
Can I run both?
Score your next cut before you commission a panel test.
SaliencyLab returns a verdict in 90 seconds. Three free runs, no card required. Decide which variants are worth a Star Rating in the first place.