Creator

Pattern Fusion

A Content Spy synthesizer: score 14 packaging patterns on a published video, then recommend blends to run next. Ten credits. Compare mode included. Not the Content Lab tool of the same name.

Overview

Spy Pattern Fusion sits in Content Spy under Virality & Growth. You pick a published video. The model scores 14 named patterns on the packaging (title, description, tags, duration class), then synthesizes recommended blends, a psychological stack, and structural upgrades. It is optimization, not classification. Viral Mechanics classifies what already fired. This tool tells you what to combine next.

Content Lab already has a page named Pattern Fusion at /resources/pattern-fusion (3 credits: Identification, Fusion Analysis, Innovation, Replication Blueprint). Lab reverse-engineers how a loaded video fuses formats. Spy scores packaging, then proposes new blends. Same English name. Different product, cost, and URL.

How to run

  1. 1Open Content Spy and select Pattern Fusion — or start Compare mode and add this tool.
  2. 2Choose a published (or unlisted) video. No extra form. Duration ≤ 60 seconds is YouTube Shorts; longer is YouTube Video.
  3. 3Confirm. 10 credits — the same in a single run or per video in a comparison.
  4. 4Read Blend Strength and the Dominant Cluster first (Attention, Retention, or Conversion). Then the 14 pattern scores, recommended blends, and the hook / retention / conversion upgrades.
Ten credits sits between Viral Mechanics (15) and Expected Value (8). Run Viral Mechanics when you want a 15-pattern map with transcript and thumbnail vision. Run this when you want blends to apply.

What it uses

Packaging payload. No transcript window. No thumbnail vision pass — unlike Viral Mechanics and CTR Audit.

  • Title, description (first 800 characters), up to 15 tags.
  • Duration class: YouTube Shorts if ≤ 60 seconds, otherwise YouTube Video. That switches Shorts vs long-form blending rules.
  • Channel, views, likes, comments — context only. They do not auto-inflate a pattern score.

What is not in the payload

  • Transcript — this engine does not read spoken evidence. Retention patterns are inferred from packaging, not from the body of the script.
  • Thumbnail vision — the prompt talks about thumbnail signals, but this route does not run a vision model. Face/overlay evidence is not a separate paragraph here.
  • A draft form. The video has to be in the library.
Monetization lines (CTR lift, retention lift, estimated RPV $2–$25) are model estimates, not AdSense, not YouTube Analytics RPM.

Fourteen patterns, three clusters

Every run scores all 14 names (0–100). 0 = absent. ~30 weak. ~50 moderate. ~70 strong. 90+ dominant. The Dominant Cluster is whichever group is strongest as a set: Attention, Retention, or Conversion.

Named recipes the model may recommend: Curiosity + Open Loop + Escalation (binge-watch). Fear + Specificity + Outcome (high-conversion). Identity + Validation + Tribal (community). Surprise + Escalation + Rewatch (virality). Validation, Tribal, and Rewatch are recipe names — they are not extra scored pillars in the 14.

Attention drivers

Hook
  • Curiosity Gap, Cognitive Dissonance, Surprise, Fear / Loss, Identity Hook.
  • Shorts: interrupt in the first 1–2 seconds. Long-form: the title still has to earn the click before any loop can work.

Retention drivers

Watch time
  • Open Loop, Micro-Narrative, Escalation, Progressive Revelation.
  • Shorts: new stimulus every 2–4 seconds, loops that close inside 30–60 seconds. Long-form: stacked loops, re-hooks every 60–90 seconds.

Conversion drivers

Action
  • Outcome Reinforcement, Authority, Specificity, Risk Reversal, Payoff.
  • These score whether the packaging promises a concrete, low-risk action — not whether the viewer subscribed.

What you get

Single-video runs return the blend analysis. Compare mode returns a compact pattern-weight card for the same 10-credit cost per video.

  • Pattern Scores — all 14 names, each 0–100 with a 5–15 word why.
  • Current Blend Profile — Dominant Cluster (Attention / Retention / Conversion), Blend Strength 0–100, imbalance areas.
  • Recommended Blends — named recipes with patterns, purpose, and expected impact for this video.
  • Psychological Stack — ordered triggers to deploy through the cut.
  • Structural Upgrades — Hook, Retention, Conversion.
  • Monetization Impact — estimated CTR lift, retention lift, RPV impact. Estimates, not invoices.
  • Compare mode: dominant / secondary patterns plus compact weights (curiosity, emotion, value, story, controversy, scarcity) — not the full 14-row dump.
A Blend Strength of 71 is how well the current mix works together — not a virality %, not a Lab Pattern Fusion Score, not RPM.

Case study

Composite documentation scenario — not a live client report. Scores are model judgments on packaging, not Analytics.

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). Title withholds the cause. Description lists tools, not a failed attempt. Escalation is the hole.

Context (available inputs)

  • Published library video — title, description, tags, duration class
  • No transcript, no thumbnail vision
  • Content type: YouTube Video (> 60s)
  • Cost: 10 AI credits

Tool result

Pattern scores

0–100

66avg
  • Curiosity Gap84/100
  • Identity Hook71/100
  • Specificity70/100
  • Open Loop62/100
  • Escalation41/100
Pattern Fusion — Content Spy
Curiosity Gap84 / 100
84
Identity Hook71 / 100
71
Specificity70 / 100
70
Open Loop62 / 100
62
Escalation41 / 100
41
71

Blend Strength

71 / 100

Dominant cluster: Attention. Imbalance: strong withheld title, weak escalation in the described middle. Recommended blend: Binge-Watch (Curiosity + Open Loop + Escalation).

Recommended blend

  • Patterns: Curiosity Gap, Open Loop, Escalation.
  • Purpose: keep the withheld cause alive past the settings tour.
  • Impact: fewer mid-video exits if each tool beat is more intense than the last.

Upgrades

  • Hook: keep the withheld cause; name the cost in the first line of the description.
  • Retention: one failed attempt before the reveal so Escalation has a beat.
  • Conversion: close on a single default to try, not a 12-tool stack.

Takeaways

  • Spy Fusion proposes the next mix. It does not certify that this video already fused well.
  • Lab Pattern Fusion (3 credits) is the reverse-engineer / Replication Blueprint. Link that page when you want identification, not synthesis.
  • No transcript here. If you need spoken evidence, run Viral Mechanics (15) or Retention Analysis (6) on the same video.

Frequently Asked Questions

Is this the same as Pattern Fusion in Content Lab?

No. Lab Pattern Fusion is a 3-credit research tool: Identification, Fusion Analysis, Innovation, Replication Blueprint, overall Pattern Fusion Score. Spy Pattern Fusion is a 10-credit synthesizer on library packaging, with Compare mode. URLs: /resources/pattern-fusion (Lab) vs /resources/spy-pattern-fusion (this page).

How is this different from Viral Mechanics?

Viral Mechanics classifies 15 behavioral patterns from transcript + thumbnail vision (15 credits). Pattern Fusion scores 14 packaging patterns with no transcript and no vision, then recommends blends (10 credits). Classify first, fuse second — or skip classification if you only want recipes.

Does it read the transcript or the thumbnail?

Neither. Title, description, tags, duration class, and engagement metadata. No 8,000-character transcript window. No vision paragraph.

What counts as a Short here?

Duration of 60 seconds or less — YouTube Shorts. Same line as Spy Viral Mechanics. Not Retention Analysis’s 180-second Short-form.

Are CTR lift and RPV real money?

No. They are model estimates. RPV is bounded to a realistic $2–$25 per 1K views in the prompt. Not AdSense. Not Analytics RPM.

Can I compare two videos?

Yes. Compare mode is on. 10 credits per video. The compare card is compact (dominant, secondary, six weights) — not the full 14-pattern dump.

How much does one run cost?

10 AI credits. Compare multiplies by 2 or 3.

Next Steps

Viral Mechanics

Classify 15 patterns on the same published video (transcript + thumbnail vision, 15 credits).

Pattern Fusion (Lab)

The 3-credit Content Lab tool: reverse-engineer a loaded video’s fusion and get a Replication Blueprint.

Content Spy

The rest of the spy tools, including Expected Value when you want monetization estimates next.

AI Credits

Spy Pattern Fusion costs 10 credits — the same in a single run or per video in Compare mode.