Instinct
Sign In
AUDIENCE INTELLIGENCE

Your viewers are already telling you what to make next.

Instinct treats comments and community posts as signals — clustered, ranked, and fed into the audience-fit score behind every prediction. Not sentiment percentages. Not vanity metrics. Concrete demand.

THE SIGNAL STACK

Three sources. One picture of demand.

COMMENTS

Semantic clustering across every comment on your last 20 videos. Requests, validation, confusion, emotional response — separated, counted, and ranked.

COMMUNITY POSTS

Poll results, reaction spread, and engagement velocity read as demand signals. What your audience votes for is what Instinct weights.

CHANNEL VIEWERSHIP

Segment overlap between real subscribers and discovery traffic. Which viewer clusters actually watch, and which the algorithm sends but doesn't stick.

WHAT INSTINCT SURFACES

Named signals, not sentiment percentages.

Every signal has a type, a strength score, and a time signature. You can act on "34 explicit requests for a gear breakdown." You can't act on "68% positive sentiment."

  • · Validation spike — when a topic outperforms your baseline
  • · Audience request — explicit asks, counted and grouped
  • · Emotional pattern — recurring viewer response to specific moments
  • · Topic momentum — rising mentions of a subject or theme
  • · Confusion cluster — questions that flag a story gap
VALIDATION SPIKERolling · last 30d
Solo-adventure episodes over-index on watch-time from returning subs
92
AUDIENCE REQUESTGrowing week-over-week
34 comments across 6 videos asking for gear breakdown in a follow-up
78
EMOTIONAL PATTERNConsistent · 12 videos
Vulnerability moments drive 3.2× shares vs. channel average
84
TOPIC MOMENTUMEmerging · 14d
"Cabin / hut" mentions in comments up 4.1× since last episode
71
CONFUSION CLUSTERRecurring
Viewers repeatedly ask what happened on Day 22 — the edit skips it
58
FEEDS THE PREDICTION

Audience signals feed the Right Viewers stage of every Potential Reach score.

The audience-fit score isn't a separate report you have to remember to open. It sits inside every prediction — telling you which viewer clusters this specific cut is likely to land with, based on what your real audience has been saying.

SIGNALS
Comments · Community · Viewership
RIGHT VIEWERS
79
Stage score
POTENTIAL REACH
74
Composite
See the full pre-publish workflow
IDEA BACKLOG

Recurring signals become a prioritized backlog.

Instinct turns clustered comments, requests, and topic momentum into concrete next-episode ideas — each with a demand score derived from real audience data. Not brainstorming. Retrieval.

  • The gear that survived 400 miles (and what didn't)
    34 explicit requests + high overlap with solo-hiker segment.
    88
    DEMAND
  • Day 22: what I cut from the Patagonia edit
    Confusion cluster around a missing day; audience is asking.
    74
    DEMAND
  • Alone for 31 days: how I actually spent the time
    Emotional pattern signal + validation spike on vulnerability moments.
    82
    DEMAND
  • Finding the cabin — director's cut
    Topic momentum on 'cabin/hut' + high share rate on reveal.
    79
    DEMAND
FORMAT-AWARE

Shorts audiences are measured separately.

Your Shorts viewers and your long-form viewers rarely fully overlap. Instinct measures them independently — so a Shorts prediction scores audience-fit against your Shorts-native feed, not your long-form base.

Signals in. Ideas out.

Let your audience feed the prediction engine. See the full workflow, then run a prediction on your own cut.