Focus Groups: Perception Affect Field | Veritas Links | GEO, AEO, AI Focus Groups - analyzes
VeritasLinks

Help center

Documentation

Guides for GEO reports, AI Focus Groups, scores, sharing, and the Partner API.

Focus Groups: Perception Affect Field

On a Focus Group report, Perception Affect Field answers: how does the simulated panel feel across discussion turns? The map is built from persona replies in discussion_rounds—not from GEO assistant answers. Confidence and Idea compress the session into two percentages; Affect Field shows whether that panel is emotionally one camp or two.

Focus Group corpus — discussion turns

Affect Field on Focus Groups scores persona replies across discussion rounds (need enough turns; skipped if n < 5). Polarization here is a split in the simulated panel’s mood—not model disagreement about the company.

  • Round 2 · Security buyer

    “I’d pilot if pricing were clear — right now it feels opaque.”

  • Round 2 · Skeptical VP Eng

    “Sounds like every AI wrapper. Where’s the moat?”

  • Round 3 · Champion IC

    “The workflow cut is real — I’d push procurement.”

  • Round 3 · Risk-averse CISO

    “Without SOC2 proof I’m not putting this in the stack.”

Pipeline also backfills per-response sentiment labels on discussion turns from the same scoring pass.

What gets scored on FG: persona replies across rounds.

What the analyst should ask

  • Is the panel emotionally aligned, or do champions and skeptics form two modes?
  • Does high Idea coexist with polarized affect (enthusiasts vs blockers in the same room)?
  • Which personas / rounds feed the negative mode—pricing, trust, moat, or fit?

Where it appears

  • Focus Group results — Perception Affect Field tab after Discussion when the affect_field_analyzer step produced data.
  • Combined report — Focus group · Perception Affect Field.

Skipped when there are fewer than ~5 scorable turns (or the affect module is unavailable). The same pass can backfill per-response sentiment labels on discussion turns for reading next to the transcript.

Metrics on the tab

Affect field metrics

Valence × arousal map

Computed PNG in live reports

Valence
0.22
Arousal
0.61
Dispersion
0.48
CI radius
0.19
Dominant zone
tense / activated
Responses scored
186
Polarization index
0.72
Map frame and metric cards (same component as GEO; smaller n typical).

Split perception alert

Split perception detected

Responses cluster in two modes — not a single average. Treat this as a reputational risk signal, not neutral consensus.

Split perception = bimodal panel mood, not model disagreement.
  • Valence / Arousal — centroid of scored discussion turns.
  • Dispersion / CI radius — how tightly the panel clusters (small_n mode is common for FG).
  • Dominant zone / Responses scored — where the room sits and how many turns entered the field.
  • Polarization index — strength of split between emotional modes in the panel.
  • Verdict — short narrative of the session’s emotional field.

How to act (Focus Group–specific)

  1. If Split perception detected, open Discussion and read the two modes side by side—do not average into “mildly positive.”
  2. Reconcile with Confidence & Idea: mid Idea + polarized affect often means a blocked champion path (pricing, proof, risk) rather than a soft product.
  3. Use Risks and persona ratings to name the skeptical archetype; fix the objection in messaging before the next FG.
  4. For AI Investor Review on a pitch deck, FOCUS (FG1–FG5) later checks whether the deck preempts objections this panel raised—Affect Field is the emotional map; FOCUS is the deck cross-check.
  5. Do not confuse with GEO Affect Field: changing citation footprint may move GEO emotion without changing panel mood, and vice versa.

Related