PHASE 5: UNIFIED STRUCTURAL FUNCTION TAXONOMY

Direct-Response Video Ads -- Final Deliverable

Merged from 47 Reference Codes + 63 BR Clusters (316 raw annotations, 25 ads)

Generated: 2026-02-19

SECTION 1: UNIFIED TAXONOMY TABLE

Merge Methodology

The taxonomy below merges two independent sources:

1. Reference Vocabulary -- 47 codes drawn from English-language direct-response copywriting frameworks (codes 1-47 in the Reference Vocabulary list).
2. BR Clusters -- 63 clusters (C01-C63) extracted bottom-up from 316 Brazilian Portuguese copywriter annotations across 25 real-world video ads.

Merge rules applied:

• Where a BR cluster and a reference code describe the same structural function, they are merged into a single unified code (source = BOTH). The more descriptive name is preferred.
• Where a BR cluster has no reference equivalent, it is added as BR_ONLY.
• Where a reference code has no BR equivalent, it is added as REFERENCE_ONLY.
• C10 (CTA_COMPOUND) is flagged as a structural annotation artifact. Its component functions already map to other codes. It is not given a unified code but is documented in the notes.
• C07 (PRESENTATION_SOLUTION_MECHANISM) is kept as BR_ONLY -- it captures a distinct step (formal introduction) between tease and explanation with no direct reference equivalent.
• C27 (NEGATIVE_FUTURE_PACING) is kept as BR_ONLY -- it is the inverse of FUTURE_PACE with no reference equivalent.

Legend

ATTENTION Category

QUALIFICATION Category

PROBLEM Category

INVALIDATION Category

SOLUTION Category

PROOF Category

EMOTION Category

CTA Category

SCARCITY Category

PACING Category

META Category

No codes from either source were assigned to META during the merge. The META category exists in the hierarchy as a placeholder for future annotation-level metadata (e.g., annotation quality flags, structural position markers).

Merge Notes

1:1 Merges (BR cluster = Reference code)

Split/Multi-Map Decisions

1. OBJECTION_HANDLING (C37) vs. SKEPTICISM_HANDLING (C38) vs. SKEPTICISM_DISARM (#20): The reference code SKEPTICISM_DISARM covers both objection handling and skepticism management. The BR data clearly separates these into two functions: C37 handles specific objections (safety, price, surgery fear), while C38 handles general skepticism ("I know this sounds crazy"). Decision: Keep both. Map SKEPTICISM_DISARM to BOTH C37 (as OBJECTION_HANDLING) and C38 (as SKEPTICISM_HANDLING).

2. PROOF_AUTHORITY (#23) vs. MEDICAL_AUTHORITY (C17): C17 is a sub-type of PROOF_AUTHORITY specific to medical credentials. Both are kept. PROOF_AUTHORITY is mapped as BOTH (broader); MEDICAL_AUTHORITY is kept as a specialized variant. In practice, the same annotations often serve both functions; we note the overlap but preserve the distinction for analytical precision.

3. PAIN_ARTICULATE (#5) + PAIN_AGITATE (#6) vs. PAIN_DEEPENING (C22): The BR data does not separate the act of naming pain from escalating it. C22 covers both. Decision: Map C22 to PAIN_ARTICULATE (the naming function) as the primary match, and note that PAIN_AGITATE is a REFERENCE_ONLY code representing a distinct escalation step that BR annotators did not separate.

4. C10 (CTA_COMPOUND): This is a structural annotation artifact, not a distinct function. Each CTA_COMPOUND instance decomposes into CTA_DIRECT + another code (SCARCITY, SOCIAL_PROOF, BENEFITS, etc.). No unified code is created. The 14 occurrences are redistributed to their component codes.

Orphan Disposition

SECTION 2: CATEGORY HIERARCHY

UNIFIED TAXONOMY HIERARCHY
============================
78 codes across 11 categories (10 populated + 1 placeholder)

ATTENTION (hooks, openers, pattern interrupts)
|
+-- SCROLL_STOP [#1] (REFERENCE_ONLY)
|
+-- CURIOSITY_HOOK [#2] (BOTH)
|   +-- PARADOXICAL_QUESTION [#3] (BOTH, sub-type: question-based curiosity)
|   +-- CONTRARIAN_HOOK [#4] (BR_ONLY, sub-type: contrarian-based curiosity)
|
+-- BELIEF_DISRUPTION [#5] (BOTH)
|
+-- THE_ONE_THING [#6] (BR_ONLY)
|
+-- STORY_EMOTIONAL [#7] (BR_ONLY)
|
+-- CURIOSITY_AMPLIFIER [#8] (REFERENCE_ONLY)


QUALIFICATION (who is this for)
|
+-- SELF_SELECT [#9] (BOTH)
|   +-- AUDIENCE_EXPANSION [#11] (BR_ONLY, broadens qualified audience)
|
+-- UNIVERSAL_APPLICABILITY [#10] (BOTH)
|
+-- QUALIFY [#12] (REFERENCE_ONLY, late-ad requalification)


PROBLEM (establishing what is wrong)
|
+-- PAIN_ARTICULATE [#13] (BOTH)
|   +-- PAIN_AGITATE [#14] (REFERENCE_ONLY, escalation of articulated pain)
|
+-- SPOILER_PROBLEM_MECHANISM [#15] (BOTH)
|   +-- EXPLANATION_PROBLEM_MECHANISM [#16] (BOTH, full reveal -- child of tease)
|       +-- ANALOGY [#77] (BR_ONLY, rhetorical device within explanation)
|
+-- ROOT_CAUSE_REVEAL [#17] (BOTH)
|
+-- CONSEQUENCES_OF_PROBLEM [#18] (BR_ONLY)
|
+-- COMMON_MISTAKE [#19] (BOTH)
|
+-- REASON_WHY [#20] (BOTH)


INVALIDATION (why other solutions fail)
|
+-- INVALIDATE_SOLUTIONS [#21] (BOTH)
|   +-- INVALIDATE_ON_PRICE [#22] (BOTH, specific: too expensive)
|
+-- SOLUTION_2_0_BRIDGE [#23] (BR_ONLY, transition from invalidated 1.0 to 2.0)


SOLUTION (the method/product)
|
+-- SPOILER_SOLUTION_MECHANISM [#24] (BOTH)
|   +-- PRESENTATION_SOLUTION_MECHANISM [#25] (BR_ONLY, formal introduction)
|       +-- EXPLANATION_SOLUTION_MECHANISM [#26] (BOTH, full reveal)
|
+-- METHOD_SIMPLICITY [#27] (BOTH)
|   +-- RAPID_RESULT [#29] (BR_ONLY, speed of results -- child of simplicity)
|
+-- METHOD_SAFETY [#28] (BOTH)
|
+-- PROMISE [#30] (BOTH)
|   +-- RESULTADO_FINAL [#37] (BR_ONLY, concrete end-state image)
|
+-- BENEFITS [#31] (BR_ONLY)
|   +-- PRICE_BENEFIT [#32] (BR_ONLY, cost advantage)
|
+-- OBJECTION_HANDLING [#33] (BOTH)
|
+-- DIFFERENTIATION [#34] (BOTH)
|
+-- PRODUCT_BUILDING [#35] (BR_ONLY)
|
+-- MARKETING_THESIS [#36] (BR_ONLY)


PROOF (evidence and credibility)
|
+-- SOCIAL_PROOF [#38] (BOTH)
|   +-- TESTIMONIAL_PROOF [#42] (BR_ONLY, named individuals)
|
+-- PERSONAL_PROOF [#39] (BR_ONLY)
|
+-- DEMONSTRATIVE_PROOF [#40] (BR_ONLY)
|
+-- SCIENTIFIC_PROOF [#41] (BR_ONLY)
|   +-- SCIENTIFIC_DISCOVERY [#47] (BR_ONLY, framing as breakthrough)
|
+-- EXPERT_PRESENTATION [#43] (BOTH)
|   +-- MEDICAL_AUTHORITY [#44] (BOTH, medical credentialing)
|   +-- PROOF_AUTHORITY [#45] (BOTH, broader authority signals)
|
+-- PROOF_SPECIFICITY [#46] (REFERENCE_ONLY)
|
+-- SUPERSTRUCTURE [#48] (BOTH)
|
+-- PROOF_THAT_SOLUTION_WORKS [#49] (BR_ONLY)
|
+-- CREDIBILITY_SEED [#50] (REFERENCE_ONLY)


EMOTION (psychological levers)
|
+-- FEAR_DEEPENING [#51] (BOTH)
|
+-- COMMON_ENEMY [#52] (BOTH)
|   +-- CONSPIRACY_LEAN [#53] (REFERENCE_ONLY, stronger version)
|
+-- IDENTITY_SPEAK [#54] (REFERENCE_ONLY)
|
+-- FUTURE_PACING [#55] (BOTH)
|   +-- NEGATIVE_FUTURE_PACING [#56] (BR_ONLY, inverse)
|
+-- DESIRE_ARTICULATION [#57] (BOTH)
|
+-- EXCLUSIVITY_UNIQUENESS [#58] (BR_ONLY)
|
+-- EMOTIONAL_MOBILIZATION [#59] (REFERENCE_ONLY)
|
+-- TRIBAL_MOBILIZATION [#60] (REFERENCE_ONLY)
|
+-- VIRAL_TRIGGER [#61] (REFERENCE_ONLY)


CTA (call to action)
|
+-- CTA_BUILDING [#62] (BOTH)
|   +-- MOMENT_OF_TRUTH [#65] (BR_ONLY, dramatic transition sub-type)
|
+-- CTA_DIRECT [#63] (BOTH)
|
+-- CTA_EMOTIONAL [#64] (REFERENCE_ONLY)
|
+-- OFFER [#66] (BR_ONLY)
|
+-- IDENTITY_CLOSE [#67] (REFERENCE_ONLY)


SCARCITY (urgency)
|
+-- SCARCITY_OFFER [#68] (BOTH)
|   +-- SCARCITY_INFO [#69] (BOTH, sub-type: info itself is scarce)


PACING (structural/flow devices)
|
+-- SKEPTICISM_HANDLING [#70] (BOTH)
|
+-- HOPE_BRIDGE [#71] (BOTH)
|
+-- TRANSITION_PIVOT [#72] (REFERENCE_ONLY)
|
+-- BULLETS_LIST [#73] (BR_ONLY)
|
+-- GOSSIP_INTIMATE_TONE [#74] (BR_ONLY)
|
+-- STEALTH_SELLING [#75] (BR_ONLY)
|
+-- GAP_TO_VSL [#76] (BR_ONLY)
|
+-- ANALOGY [#77] (BR_ONLY)
|
+-- PAUSE_FOR_EMPHASIS [#78] (REFERENCE_ONLY)


META (annotation-level metadata)
|
+-- (no codes currently assigned)

SECTION 3: FREQUENCY RANKING

All 78 unified codes ranked by BR corpus frequency (highest to lowest). Reference-only codes appear at the bottom with frequency 0.

SECTION 4: CO-OCCURRENCE PATTERNS

Based on the compound tag analysis from Phase 3 (152 compound annotations decomposed into atomic components), the following pairs of functions most frequently appear fused within a single annotation, indicating they are deployed simultaneously by the copywriter.

Top 15 Co-Occurrence Pairs

Co-Occurrence Interpretation

These pairings reveal the fusion logic of Brazilian direct-response copywriting:

• CTA is almost never alone. It co-occurs with SCARCITY (urgency compress), SOCIAL_PROOF (credibility at point of action), BENEFITS (restating why), and UNIVERSAL_APPLICABILITY (removing last-second exclusion worry).
• INVALIDATION and MECHANISM SPOILERS are deployed together. When disqualifying old solutions, copywriters simultaneously tease the new mechanism -- invalidation creates the vacuum, the spoiler fills it.
• EXPERT_PRESENTATION always carries AUTHORITY. The expert is never just named; credentials arrive simultaneously.
• FEAR_DEEPENING triggers SOLUTION TEASE. Fear is not left hanging -- it is immediately paired with a solution hint to prevent viewer dropout.
• SUPERSTRUCTURE + COMMON_ENEMY is a power pairing: celebrity association elevates the speaker while the enemy frame creates an us-vs-them dynamic.

SECTION 5: SEQUENCE MAP

Canonical Phase Sequence for Brazilian Direct-Response Video Ads

Based on analysis of structural position data across the 25 annotated ads, the following canonical sequence emerges. This is not rigid -- ads branch and loop -- but represents the dominant pattern.

PHASE 1: HOOK (0-5 seconds)
============================================================
Primary Functions:
  SCROLL_STOP ........... Visual/auditory pattern interrupt
  CURIOSITY_HOOK ........ Open loop (strange ingredient, paradox)
  PARADOXICAL_QUESTION .. Cognitive dissonance question
  CONTRARIAN_HOOK ....... "Everything you know is wrong"
  SELF_SELECT ........... "If you are X, keep watching"
  THE_ONE_THING ......... Single crucial insight frame

Recurrence: Appears ONCE at the start.
Notes: BR ads favor CURIOSITY_HOOK + SELF_SELECT as the dominant
       opener pattern. SCROLL_STOP is implicit in format (visual
       disruption) rather than explicitly annotated.


PHASE 2: PROBLEM ESTABLISHMENT (5-30 seconds)
============================================================
Primary Functions:
  PAIN_ARTICULATE ....... Name the audience's pain
  STORY_EMOTIONAL ....... Personal narrative to create connection
  SPOILER_PROBLEM_MECHANISM .. Tease the MUP
  COMMON_MISTAKE ........ "The error most people make"
  FEAR_DEEPENING ........ Escalate anxiety about consequences

Recurrence: Single pass, but FEAR_DEEPENING can recur later.
Branching: Some ads skip STORY_EMOTIONAL and go straight to
           mechanism teasing. Emotional vs. logical entry point.


PHASE 3: INVALIDATION (30-60 seconds)
============================================================
Primary Functions:
  INVALIDATE_SOLUTIONS .. Disqualify competing methods
  INVALIDATE_ON_PRICE ... "They cost too much"
  COMMON_ENEMY .......... Name the external villain
  BELIEF_DISRUPTION ..... Break a common assumption
  SCARCITY_INFO ......... "This info is being suppressed"

Recurrence: INVALIDATION appears 2-3 times across the ad.
            Initial invalidation here; reinforcement near CTA.
Branching: Some ads merge invalidation with ENEMY_FRAME
           for a conspiracy-adjacent tone. Others keep it
           clinical and price-focused.


PHASE 4: MECHANISM REVEAL (60-120 seconds)
============================================================
Primary Functions:
  ROOT_CAUSE_REVEAL ........... "The real reason is..."
  EXPLANATION_PROBLEM_MECHANISM  Full MUP explanation
  ANALOGY ..................... Metaphor for complex mechanism
  HOPE_BRIDGE ................. Pivot from problem to solution
  SPOILER_SOLUTION_MECHANISM .. Tease the MSOL
  PRESENTATION_SOLUTION_MECHANISM  Formally introduce the MSOL
  EXPLANATION_SOLUTION_MECHANISM   Full MSOL explanation
  REASON_WHY .................. Logical justification
  SCIENTIFIC_DISCOVERY ........ Breakthrough framing

Recurrence: Core mechanism reveal appears ONCE. But SPOILER_MSOL
            may have been teased earlier (Phase 2) and is now
            formally revealed.
Branching: The 3-step MSOL sequence (tease -> present -> explain)
           is a BR-specific pattern. Some ads compress to 2 steps.
           HOPE_BRIDGE is the structural hinge between problem
           and solution phases.


PHASE 5: PROOF STACK (120-180 seconds)
============================================================
Primary Functions:
  EXPERT_PRESENTATION ... Introduce the authority figure
  MEDICAL_AUTHORITY ..... Medical credentialing
  PROOF_AUTHORITY ....... Broader authority signals
  SCIENTIFIC_PROOF ...... Studies, clinical data
  SOCIAL_PROOF .......... Crowd/statistics
  PERSONAL_PROOF ........ Speaker's own experience
  DEMONSTRATIVE_PROOF ... Visual demonstration
  TESTIMONIAL_PROOF ..... Named testimonials
  SUPERSTRUCTURE ........ Celebrity/TV association
  PROOF_THAT_SOLUTION_WORKS  Direct evidence of MSOL

Recurrence: SOCIAL_PROOF recurs 2-4 times throughout the ad.
            Other proof types typically appear once but are
            stacked in rapid succession in this phase.
Branching: Some ads front-load EXPERT_PRESENTATION (Phase 2)
           and use this phase for SOCIAL_PROOF only. Others
           save all proof for a concentrated "proof barrage."


PHASE 6: SOLUTION AMPLIFICATION (180-240 seconds)
============================================================
Primary Functions:
  METHOD_SIMPLICITY ..... "It's easy, anyone can do it"
  METHOD_SAFETY ......... "It's safe, natural"
  RAPID_RESULT .......... "Results in 7 days"
  BENEFITS .............. Enumerate positive outcomes
  PRICE_BENEFIT ......... "Costs less than alternatives"
  UNIVERSAL_APPLICABILITY  "Works for everyone"
  DIFFERENTIATION ....... "This is not like X or Y"
  OBJECTION_HANDLING .... Address remaining doubts
  FUTURE_PACING ......... "Imagine your life when..."
  NEGATIVE_FUTURE_PACING  "If you don't act..."
  PROMISE ............... Core transformation claim
  RESULTADO_FINAL ....... Concrete end-state image

Recurrence: BENEFITS and METHOD_SIMPLICITY recur throughout.
            FUTURE_PACING typically appears once in this phase
            and once at the CTA.
Branching: Rational ads emphasize METHOD_SIMPLICITY + PRICE.
           Emotional ads emphasize FUTURE_PACING + PROMISE.


PHASE 7: CLOSE / CTA (final 30-60 seconds)
============================================================
Primary Functions:
  CTA_BUILDING .......... Prepare the viewer for the ask
  MOMENT_OF_TRUTH ....... "This is your moment"
  CTA_DIRECT ............ "Click the button below"
  SCARCITY_OFFER ........ Limited time/slots
  SCARCITY_INFO ......... "This video may be taken down"
  COMMON_ENEMY .......... (reinforcement) Enemy reminder
  INVALIDATE_SOLUTIONS .. (reinforcement) "Don't go back to..."
  SOCIAL_PROOF .......... (reinforcement) Final social proof
  QUALIFY ............... "If you're serious about..."

Recurrence: CTA_DIRECT appears 2-3 times (early soft CTA,
            mid-ad CTA, final hard CTA). SCARCITY intensifies
            with each repetition.
Notes: The "triple CTA" pattern is dominant in BR ads:
       - CTA_1 (soft, after proof) ~60% through
       - CTA_2 (medium, after benefits) ~80% through
       - CTA_3 (hard, with scarcity) ~95% through

Recurrence Summary Table

Structural Branching Points

1. After Hook: Emotional path (STORY_EMOTIONAL -> PAIN_ARTICULATE) vs. Logical path (SPOILER_MUP -> COMMON_MISTAKE)
2. After Invalidation: Deep mechanism path (ROOT_CAUSE_REVEAL -> full MUP) vs. Surface path (skip MUP -> HOPE_BRIDGE -> MSOL tease)
3. Proof Phase: Authority-led (EXPERT_PRESENTATION first) vs. Social-led (SOCIAL_PROOF first) vs. Science-led (SCIENTIFIC_DISCOVERY first)
4. CTA approach: Scarcity-driven (SCARCITY_OFFER dominant) vs. Emotion-driven (FUTURE_PACING + NEGATIVE_FUTURE_PACING dominant)

SECTION 6: VIEWER PSYCHOLOGY LAYER

Dual-Layer Map: Strategic Function <-> Psychological Response

The 6 psychological tags from Ad 1 provide a rare window into the viewer's internal monologue at specific structural points. Below, each psychological response is mapped to the strategic function that triggers it, creating a dual-layer view of the persuasion process.

Layer Mapping

The Persuasion Cascade

Reading the six responses in sequence reveals a complete persuasion cascade:

RECOGNITION        ->  "I know this person"
    |
CREDIBILITY        ->  "They seem legitimate"
    |
SURPRISE           ->  "Wait, I didn't know that"
    |
ACCEPTANCE         ->  "OK, that makes sense"
    |
ALARM              ->  "This is worse than I thought"
    |
CURIOSITY VACUUM   ->  "So what's the answer?"
    |
    v
[VIEWER IS NOW MAXIMALLY RECEPTIVE TO THE SOLUTION]

This cascade maps to the structural phases:

• Recognition + Credibility = HOOK / QUALIFICATION phase
• Surprise = PROBLEM ESTABLISHMENT phase (MUP tease)
• Acceptance = MECHANISM REVEAL phase (full explanation)
• Alarm = FEAR DEEPENING (emotion escalation)
• Curiosity Vacuum = INVALIDATION -> HOPE_BRIDGE (transition to solution)

Implications for Taxonomy Use

The psychological layer suggests that each strategic function has a target internal state it aims to produce in the viewer. Future annotation projects could track both layers simultaneously:

SECTION 7: COVERAGE ANALYSIS

Summary Statistics

Frequency Extremes

Category Distribution

Source Coverage Analysis

What BR Copywriters Use That the Reference Vocabulary Missed

The 32 BR_ONLY codes reveal functions that grounded practice in Brazilian direct-response ads employs but that are not captured in the English-language reference framework. Key gaps:

1. 3-step MSOL reveal (Tease -> Present -> Explain): The reference vocabulary has only Tease and Explain. BR annotators consistently identified a middle step -- PRESENTATION_SOLUTION_MECHANISM -- where the solution is formally named and positioned before being explained.

2. Proof type granularity: BR annotations distinguish PERSONAL_PROOF, DEMONSTRATIVE_PROOF, SCIENTIFIC_PROOF, and TESTIMONIAL_PROOF as separate functions, while the reference collapses these into PROOF_SOCIAL and PROOF_AUTHORITY.

3. Negative Future Pacing: The reference has FUTURE_PACE (positive) but no inverse. BR ads consistently use "what happens if you DON'T act" as a separate function.

4. Price as persuasion: PRICE_BENEFIT is a standalone function in BR ads (8 occurrences), not just a sub-type of invalidation or offer.

5. Structural/flow devices: BULLETS_LIST, GOSSIP_INTIMATE_TONE, STEALTH_SELLING, GAP_TO_VSL, and ANALOGY are all pacing/structural devices that the reference vocabulary does not address.

6. Rapid Result: Speed of results is a distinct function (10 occurrences) in BR ads, not merely an attribute of PROMISE.

What the Reference Vocabulary Has That BR Copywriters Did Not Annotate

The 15 REFERENCE_ONLY codes represent functions that either:

• Are theoretically valid but were not observed in the BR corpus (e.g., CONSPIRACY_LEAN, TRIBAL_MOBILIZATION, VIRAL_TRIGGER)
• Are implicit rather than annotated (e.g., SCROLL_STOP is achieved visually but not tagged as a copywriting function)
• Represent finer distinctions not made by BR annotators (e.g., PAIN_AGITATE as separate from PAIN_ARTICULATE)

This does not mean these functions are absent from BR ads -- only that the annotators did not label them as distinct structural elements.

Completeness Assessment

The unified taxonomy of 78 codes provides comprehensive coverage of direct-response video ad structure. Every one of the 316 raw annotations from the BR corpus maps to at least one unified code. The reference vocabulary's 47 codes are all represented. The taxonomy is ready for use as an annotation framework, a structural analysis tool, or a copywriting blueprint.

End of Phase 5: Unified Structural Function Taxonomy Total unified codes: 78 Source coverage: 47 reference + 63 BR clusters -> 78 merged codes Corpus: 316 annotations from 25 Brazilian direct-response video ads Generated: 2026-02-19