Help center
Questions, answered plainly.
The shortest path to understanding how the product works, where the boundaries are, and what to read next if you are evaluating SaliencyLab for a real workflow.
Common questions
What does RoastIQ actually give me?
RoastIQ turns one uploaded creative into five KPI scores, benchmark context, evidence-led diagnostics, and a clearer next action before media spend is committed.
Do I need a consumer panel to use SaliencyLab?
No. RoastIQ runs on the creative itself. BuyerLens is optional and adds structured audience simulation when you want directional buyer reactions before commissioning human research.
How should I read the benchmark labels?
Benchmarks are there to interpret a score, not replace judgment. They help you see whether a result is above, around, or below the bar for a comparable format, platform, and context.
What happens after the free plan limit?
The free workspace includes 10 RoastIQ analyses per month, up to 3 per day. Pro raises that cap and unlocks the compare workflow, BuyerLens panels, and richer decision support. Access to either starts with a demo call.
Is SaliencyLab claiming to replace human research?
No. The product is designed for pre-spend screening, prioritisation, and iteration. High-risk decisions, regulated claims, and final validation can still require human research.
How is uploaded creative handled?
Uploads move through authenticated workspace flows, private storage paths, and signed media access. Public pages do not expose your raw workspace assets.
Do you publish validation coefficients on this site?
Yes. For the platform-native attribute model: +0.33 on TikTok click-through rate (n=691), held out. On the cohort its weights were tuned on the same model scores +0.43, brand-clustered 95% CI [+0.32, +0.53]. +0.40 on YouTube view counts (n=397), held out. On the cohort its weights were tuned on the same model scores +0.49, brand-clustered 95% CI [+0.37, +0.58]. Both are public outcomes (click-through and views), and they validate that model only, not the five-KPI diagnostic; the methodology page carries the full method and what we deliberately do not claim, such as sales, ROAS, or brand recall.
Where should a new team start?
Book a demo and bring one real creative. We score it live, read the five KPI scores together, and you use Compare and BuyerLens only when that first diagnostic shows a decision that needs more pressure testing.
Read next
Getting started guide
Walk through the first upload, first report, and what to do once the score lands.
Calibration notes
See how benchmarks, confidence labels, and decision framing are handled today.
Security overview
Review the current product practices for uploads, access, and vendor handling.
How the scoring works
The three scoring layers, the five KPIs, and how the verdict is derived.