Viral Mechanics
A Content Spy classifier that maps a published video onto a 15-pattern viral taxonomy — attention, retention, emotion, amplification — from title, transcript, and thumbnail vision. Fifteen credits. Compare mode included.
Overview
Viral Mechanics sits in Content Spy under Virality & Growth. You pick a published video from your library. The model reads title, description, tags, transcript, duration, engagement metadata, and a thumbnail vision pass — then scores which of 15 behavioral patterns are actually present. It does not predict views, does not guarantee a viral hit, and does not read YouTube’s “viral” badge.
How to run
- 1Open Content Spy and select Viral Mechanics — or start Compare mode and add this tool to the run.
- 2Choose a published (or unlisted) video. There is no extra form. Duration ≤ 60 seconds is labeled YouTube Shorts; longer is labeled YouTube Video.
- 3Confirm the run. It costs 15 credits — the same in a single analysis or per video in a comparison.
- 4Read the Dominant Pattern first, then only the patterns that scored above 0. Absence is a finding: a zero is omitted, not shown as a 0 bar.
What it uses
Published-video payload. No draft form. Thumbnail vision is on — unlike Retention Analysis.
- Title, description (first 1,000 characters), and up to 15 tags.
- Transcript — first 8,000 characters. Same window as Retention Analysis; hold and pattern evidence live in the body.
- Thumbnail vision — one paragraph on face, contrast, overlay, clutter. Used especially for Attention Acquisition (patterns 1–4). Same class of signal as CTR Audit, different job.
- Duration, channel, category, views, likes, comments, subscribers — context only. High views do not raise a pattern score by themselves.
What is missing, and what that does
- No transcript: the model scores conservatively and should say so. Retention Engine and Rewatch get weaker evidence.
- No thumbnail vision (key missing or image unavailable): Attention Acquisition is scored from title and copy only.
- A draft script. Unlike Retention Prediction, you cannot type a plan here. The video has to be in the library.
The 15-pattern taxonomy
Four groups. Each pattern is 0–100 when present. Score 0 is omitted from the list. Dominant Pattern is the single highest score among those listed.
I. Attention Acquisition
CTR drivers- Cognitive Dissonance — contrarian claim that breaks an expected belief.
- Curiosity Gap — withheld information that forces the click.
- Outcome-First — result first, process later (“$0 to $50K”).
- Identity & Ego — the viewer recognizes themselves in the address.
II. Retention Engine
Watch-time drivers- Open Loop — unanswered question (Zeigarnik).
- Micro-Narrative — protagonist, conflict, resolution.
- Escalation — each beat more intense than the last.
- Pattern Interrupt — a mid-video reset of tone, visual, or topic.
III. Emotional Activation
Feel, then share- Fear/Loss — loss aversion (“you’re losing this every day”).
- Surprise — prediction error, a reveal.
- Validation — “that’s me” relatability.
- Status — social comparison, aspiration (“how the top 1% think”).
IV. Virality Amplification
Comments and shares- Debate — polarization that pulls comments.
- Tribal — in-group belonging (“only real [identity] will get this”).
- Rewatch — density high enough that one pass is not enough.
What you get
Single-video runs return the taxonomy. Compare mode returns a compact side-by-side viral score for the same 15-credit cost per video.
- Dominant Pattern — name plus a one-to-two sentence why. Must be the highest listed pattern score.
- Pattern Scores — only patterns > 0. Each line: name, 0–100, specific evidence.
- Viral Levers, Weak Points, Optimization Suggestions — grounded in this video, not generic growth tips.
- Engagement Blueprint — Hook Strategy, Retention Mechanism, Emotional Stacking, Amplification Strategy.
- Compare mode also surfaces a 0–100 viralScore with triggers and shareability factors — a compact view, not a second taxonomy.
Case study
Scenario
A competitor’s published “I found out why my workflow was broken — and it wasn’t the tools” video. Long-form (over 60 seconds). Strong withheld reveal in the title. Mid-video settings tour with one tone shift. Comments are mixed, not polarized.
Context (available inputs)
- Published library video + transcript (8,000-character window) + thumbnail vision
- Content type: YouTube Video (duration > 60s)
- Views / likes used as context, not as a score bonus
- Cost: 15 AI credits
Tool result
Detected patterns
0–100
- Curiosity Gap82/100
- Open Loop74/100
- Outcome-First68/100
- Pattern Interrupt55/100
- Debate48/100
Dominant — Curiosity Gap
82 / 100
Title withholds the cause (“it wasn’t the tools”). Thumbnail overlay is four words. Transcript plants “the real reason” before the mid-roll. Debate is present but not the engine.
Viral levers
- Withheld cause in the title forces the click.
- Open loop (“the real reason”) holds the first third.
Weak points
- Settings tour has only one interrupt — Escalation is omitted (0).
- No tribal in-group address. Rewatch density is low after the reveal.
Suggestions
- Name one failed attempt before the reveal so Escalation has a beat.
- Close with a belonging line if you want Tribal — or cut Debate and stay in Curiosity Gap.
Takeaways
- The output is a map of which patterns fired — not a forecast of views.
- Zeros are omitted. If Escalation is missing, that is the finding.
- Content Lab’s Viral Mechanics group is a different product (3-credit research tools). This Spy classifier is 15 credits and Compare-capable.
Frequently Asked Questions
Will this tell me if the video will go viral?
No. It classifies which of 15 behavioral patterns are evidenced in this cut. It does not predict views, shares, or a YouTube viral badge.
Is this the same as Viral Mechanics in Content Lab?
No. In the Lab, “Viral Mechanics” is a menu group: Why This Works, Hook & Structure, Content DNA, Pattern Fusion — 3 credits each. Spy Viral Mechanics is one 15-credit taxonomy tool with Compare mode.
How is this different from Pattern Fusion?
Viral Mechanics classifies the patterns already in one published video. Spy Pattern Fusion (10 credits) blends patterns into a new strategy. Lab Pattern Fusion is a third, 3-credit research tool. Do not mix the three.
What counts as a Short here?
Duration of 60 seconds or less — YouTube Shorts. That is not Retention Analysis’s 180-second Short-form line.
Does it look at the thumbnail?
Yes. A vision pass describes face, contrast, overlay, and clutter. That evidence is aimed at Attention Acquisition (patterns 1–4). Retention Analysis does not run vision.
Why are some patterns missing from the result?
Score 0 is omitted. The list is detections, not a 15-row report card. A missing pattern means no evidence, not a UI bug.
Can I compare two videos?
Yes. Compare mode is on. Cost is 15 credits per video in the run. The compare view is a compact viralScore, not the full 15-pattern dump per card.
How much does one run cost?
15 AI credits — the highest single-tool cost in Content Spy. Compare multiplies by 2 or 3.
Next Steps
The packaging sibling. Same thumbnail-vision class of signal, scored as click strength — not taxonomy.
Score hold design on the same published video (five equal pillars, 6 credits).
The rest of the spy tools, including Pattern Fusion when you want to blend patterns instead of classify them.
Viral Mechanics costs 15 credits — the same in a single run or per video in Compare mode.