Studies

Every study in the archive, scored with the validated method. Pick one to open its winners board.

Spot Library

0.0s / 0.0s

All spots

Compare

Every spot in the study, every segment, one table. Click a column to re-rank; click a spot to open it.

Race war board

Races

Every screened ad in one race, ranked by the race's own polling mix — and the mover issues nothing in the library touches yet.

Seven years · four cycles

Baselines

Any spot against the whole archive. Production studies only — test respondents never enter a baseline.

min n

Spot	Study	Client	Year	n	Avg Net ▾

War room · recorded studies

Live Survey

Watch a respondent link field in real time — every number below moves as sessions, presses, and answers arrive.

Synthetic-audience substrate · v0

Audience Model

What seven years of real reactions say moves the persuadable-right voter — and a pre-read that scores a script against it. This is ground truth, not an AI guessing: every number is the audience's own measured response, with a validated hit rate.

Loses the room

Holds the room

Script pre-read

Paste a script or concept. It's scored line by line against the audience's learned response — before a camera rolls.

New Test

Score or field a spot

Two live paths onto the same validated engine: score dial data you already have, or collect fresh reactions with the dial widget.

1 · Score a dial file against the norm

Drop in a raw dial export — from any test, any era — and it's scored on the same axis as seven years of archive, with a G2 Index against the standard cohort, in seconds.

Drop a dial CSV here or click to browse

aid, VideoID, Time, Positive Reaction, Negative Reaction · one row per press

Spot name

Client (optional)

Duration (optional)

Pick a CSV and give the spot a name to enable scoring. Validated G2 method v1 — the same engine behind every readout in this library. The uploaded spot is measured against the standard cohort; it does not join the norm.

2 · Run any ad through the synthetic audience

Paste any public YouTube ad and the synthetic audience reads it — the predicted dial trace with its confidence band, GOP / Independent / all-voter reads modeled from real segment splits, expected trouble moments, and the real archived spots it reasoned from. A prediction, clearly labelled — never a measurement, never in the norm.

YouTube link

Spot name

Script (paste from YouTube's transcript panel)

Creates the spot in the quarantined Pre-reads project — it can never enter baselines or the Standard.

3 · Field a fresh test

Collect new reactions live: open any spot, use its Field panel to mint a dial-widget embed for Alchemer, email, or a public link — then watch presses arrive and Close & Score. Pick a spot from the Studies list to start.