Price the evidence you ask for.
Quick signal is one credit per respondent. Deep trace is estimated before the run. Founding pilots package the panel, the reactions, and the decision brief while self-serve access is still private.
- Quick signal: 1 credit per respondent
- Deep trace: estimated before run
- Decision brief summary: 1 credit per run
Two prices exist because two jobs exist.
Product credits meter run depth. Founding pilots package a done-with-you research pass while the product is still invite-only.
Credits follow evidence depth
A quick run reads the surface and returns one grounded reaction per respondent. A deep run collects encounter evidence such as scroll, dwell, click, bounce, and what was not reached.
Pilots price the decision
Early customers are buying a sharper decision, not only raw reactions. The pilot packages include panel selection, run setup, a decision brief, and the real-user handoff.
Quick signal and deep trace
The boundary is public because it affects both price and evidence. Quick signal is the default for iteration. Deep trace is for surfaces where the path matters.
Quick signal
For cheap iteration when you need one grounded read from each respondent.
- Surface extraction
- Memory-grounded reaction
- Behavioural state movement
- Included in the decision brief summary
- No encounter trace and no attention map
Deep trace
For pages, decks, flows, and app journeys where the path changes the finding.
- Rendered encounter
- Scroll, dwell, click, and bounce signals
- Memory and state stack across the encounter
- Attention-map style evidence where available
- Explicitly not eye-tracking
Decision brief
The run-level summary that keeps individual reactions useful for product decisions.
- What landed
- What broke
- What was not observed
- What to revise next
- Which real-user profiles to validate with
Every paid run returns individual reactions, memory evidence, behavioural state, known absence, and a decision brief for what comes next.
Three public entry points
These are service-shaped wrappers around the current product while self-serve access remains private.
Quick Signal Pilot
One surface, one small respondent panel, and a fast decision brief.
- Usually 5 respondents
- One surface or message
- Quick signal reactions
- Decision brief in PDF and web view
- Best for one risky claim or price point
Reaction Sprint
A focused pass across a page, deck, or short flow before a live round.
- Usually 10 to 12 respondents
- Two to three surfaces
- Quick signal by default
- Deep trace where the path matters
- Thirty-minute readout call
Segment Contrast
A broader pass when the decision depends on income, geography, age, literacy, or behaviour.
- Usually 20 to 30 respondents
- Three to five surfaces
- Segment-level comparison
- Decision brief plus real-user handoff
- Sixty-minute readout call
Custom scopes available · invite-only while the product is being calibrated
Pricing questions worth asking early
Why not only list fixed subscription tiers?
Because the honest unit is the run evidence. Quick signal is predictable. Deep trace depends on surface length, deck pages, interaction depth, and whether the run needs rendered encounter evidence.
Is quick signal shallow?
No. Quick signal still uses respondent state, memory grounding, and behavioural movement. It is cheaper because it does not build the full encounter trace or attention-map style evidence.
Do deep runs cost more because of browsing?
Yes. Web pages, decks, and app journeys may need rendering, page traversal, and more trace events. The product estimates that cost before the run starts.
What happens after the artificial read?
Every decision brief ends with a real-user handoff: what to revise, what to rerun, and which profiles deserve live validation next.
More questions in the full FAQ.