A validated SOP to create exceptional ads for performance marketing — performance-first, Meta-native, judgment-led.
Performance-first · Meta-native · Judgment-led. Read this as a system, not as advice.
This document is written so that nothing is assumed, taste is enforced structurally, bad ads are impossible to ship, new hires can execute without “Yuvi intuition,” and clients cannot derail quality.
“Nothing ships unless it survives every gate in this system.”
No creativity allowed here. Step 1 — build the Product / Truth File: the single source of truth every later section pulls from.
What is the main job the product is hired to do?
Supporting benefits only:
One sentence only: “Unlike ___, this product ___.”
Each must be concrete, provable, specific.
❌ “High quality”, “next-gen”, “AI powered” — banned.
Explicit exclusions to protect positioning:
For each alternative, document where it breaks and what it fails to guarantee. Feeds anti-pattern angles.
Every feature must map to a functional benefit and an emotional/business outcome. If it can’t → it’s irrelevant for ads.
List all proof and score each 0–5:
Explicit banned claims & words.
To decode what the market is rewarding right now, what’s overused, what’s missing, and how audiences at different temperatures respond — so strategy isn’t guessing.
Required fields for each competitor: brand name, website URL, price positioning (cheap/mid/premium), differentiation claim (what they say), funnel style (hard sell vs education vs brand).
Capture these variables: creative type (image / video / carousel), creator type (founder / customer / model / voiceover / meme), production style (UGC / studio / b-roll / motion graphics), length (if video), hook style category.
For each ad: assign funnel stage, assign audience temperature, and note CTA intensity — Soft (learn more, see how), Medium (get yours, try now), Hard (buy now, subscribe).
Quick rules to tag correctlyBuild six pattern maps:
List every angle used, count frequency across ads. Output: top 5 angles, emerging angles, rare angles.
Count UGC testimonial, founder POV, demo, before/after, anti-pattern, comment reply, voiceover + b-roll, skit/meme. Output top 3 formats by category.
Classify: contrarian, pain mirror, result-led, curiosity, authority, “stop doing X”, social proof. Output top 5 hook styles + most overused hook style.
Reviews screenshots, numbers (CPA/ROAS/results), credentials, “as seen in”, case study references, visual demonstration. Output which proof type dominates / is missing.
PAS / BAB / PPP / objection-led / mechanism-led. CTA styles (soft vs hard). Language tone (aggressive vs calm vs humorous).
Talent type, shot style, typography density, color palette trends, product prominence, visual clutter level.
Output: “Overused list” (top 5)
Output: “Whitespace list” (top 5)
From competitor ad comments (if accessible), YouTube comments, Reddit threads, Trustpilot/Amazon — extract:
Output: 50 raw lines of customer language (verbatim) + objection frequency ranking.
Market Reality Summary must include:
At Yuvichaar Funnels, personas exist to make ads believable — not academic. If a persona doesn’t help casting, it doesn’t belong in the SOP.
To define a clear, shared mental picture of the person this ad is speaking to, the kind of messenger they will trust, and the tone, credibility, and delivery required to avoid disbelief. If this section is weak: casting will be wrong, scripts will feel “off,” ads will look fake even if copy is good.
A persona cannot be imagined. It must be assembled from reality.
Required sources (minimum 3)Deliverable: minimum 30–50 verbatim user statements, raw language only (no paraphrasing).
Every persona must follow this exact structure.
Give the persona a real human name. Naming matters — teams argue better with “Arjun” than with “Target Audience Segment B”.
Answers: “Who is this ad for?”
Demographics exist only to guide casting, accent, wardrobe, setting, cultural references.
Answers: “Why would they care?”
List 3–5 pains they actively complain about. Rules: must appear in real user language, must involve money/risk/reputation, must feel unresolved. These are what they rant about, not what sales decks say.
This section defines what blocks conversion.
This section exists to prevent robotic copy. Include 6–10 exact phrases they use, slang / tone / sarcasm, words they trust, words they hate. Directly informs script phrasing, creator delivery, on-camera improvisation.
This is where personas become useful for performance. Define who they trust (peer like them? someone ahead of them? founder? expert? system/process?), credibility threshold (low / medium / high), what builds trust fast (proof types, tone, boundaries — “we don’t do X”), and what kills trust instantly (over-polished creators, hype language, big promises without proof).
Answers: “Who should say this message?”
The main reason personas exist at Yuvichaar Funnels. Define explicitly: on-camera messenger (UGC peer / founder / expert / VO), age & vibe (younger / same age / older than persona), delivery style (calm / assertive / empathetic / skeptical + energy + confidence level), visual setting (office / home / studio / screen recording).
Each persona must be tagged with primary funnel stage (Cold / Warm / Hot), audience temperature (Curious / Skeptical / Burned / Ready), CTA tolerance (Soft / Medium / Hard) — preventing aggressive CTAs to cold audiences and soft CTAs to ready buyers.
Each persona gets one defensive belief. Rules: must appear in their language, must block action, must be counterable by product truth.
Write it as: “I believe ___, so I don’t take action.”
This belief feeds: angles, hooks, proof strategy.
Founder / performance marketer running a D2C or service business. Has already spent money on Meta ads. Results are inconsistent. Scaling feels risky.
What single promise are we betting money on? Everything else is just creative decoration.
“If it’s not the promise you’d risk money on, it’s not an angle. Everything else is just creative decoration.”
To decide the ONE benefit worth betting serious money on, and to frame it in the most persuasive, belief-winning way for a specific persona and funnel stage. This section exists to answer one question:
“What is the single promise of this product that, if believed, makes action obvious?”
Before deciding what to say, decide how hard you’re allowed to say it.
Belief-locked or unaware → angle must educate or reframe
Problem-aware, solution-curious → angle must differentiate & build trust
Solution-aware, near decision → angle must remove final friction
One funnel stage per creative batch. No exceptions.
Before locking the angle, confirm persona name, trust requirement, credibility threshold, CTA tolerance.
This is the most important step in the entire SOP.
A marketing angle is the single most attractive benefit of a product, framed in a way that challenges an existing belief and makes this outcome feel uniquely achievable, credible, and worth acting on.
Short version: the angle is the one benefit you’d bet all your money on.
“If this product could only promise ONE thing better than anything else, what would it be?”
If you feel the urge to add more → you haven’t found the angle yet.
Examples of NOT angles: “High ROAS”, “AI ads”, “Save time”, “Better creatives” — these are inputs, not bets.
If any is missing → the angle is weak.
If you can’t write the angle cleanly in one sentence → it’s not clear enough to scale.
An angle only works if it attacks a real belief.
Definition: the blocking belief is what makes the customer dismiss this benefit today. Write it in first person: “I believe ___, so I don’t take action.”
You are allowed 1–3 framings of the SAME angle. The benefit stays the same — only the belief attack or explanation changes.
Example (same angle, different framings):
Explore / understand
Try / see / get access
Start / buy / subscribe
CTA intent must match funnel stage, belief resistance, proof strength.
If any of these are missing → no formats, no scripts, no shoots.
Every angle is evaluated on two axes only: Degree of Pain it Solves × Degree of Transformation it Promises.
Pain is ranked from strongest to weakest. This is hierarchical, not subjective.
“If this doesn’t change, I decline.”
Any 3 required: automatic worsening over time, loss of control/agency, threat to survival/relevance/competence, fear of irreversible failure.
e.g. Business won’t scale · Money is being gambled · I don’t know what works anymore
“I’m bleeding control or money and it’s unacceptable.”
Any 2 required: consistent financial loss, repeated failed attempts, decision paralysis, high anxiety around outcomes.
“Others are moving ahead while I’m not.”
Any 2 required: missing upside, competitive disadvantage, time-sensitive opportunity.
“This is slow, annoying, or inefficient.”
Any 1 required: time waste, manual effort, minor inefficiency.
“This would be nice.”
e.g. Easier · Cleaner · Slightly better UX.
⚠️ Never bet-worthy alone.
Pain score ≥ 2 → allowed to proceed.
Transformation is not improvement. It’s state change.
“Who I am or whether I survive changes.”
e.g. From failing → surviving · From irrelevant → competitive · From powerless → in control.
“How I see myself as an operator/person changes.”
e.g. From guessing → deciding · From amateur → serious · From reactive → deliberate.
“I gain clarity and repeatability.”
e.g. Predictable outcomes · Structured decision-making · Reduced uncertainty.
“I can do more, faster.”
e.g. Automation · New skills · Faster execution.
“Same life, slightly better.”
e.g. Faster · Cheaper · Easier.
Transformation score ≥ 2 → allowed to proceed.
Outcome Delight = Pain Score × Transformation Score
Just: how much does this matter × how much does life change.
“Confidence to scale ads without gambling on creatives”
Pain: Existential (5) · Transformation: Identity (4–5)
👉 Outcome Delight = 20–25 → 🚀 Scale-worthy
“Become the strongest version of yourself”
Pain: Aspirational identity pain (3) · Transformation: Identity (4)
👉 Outcome Delight = 12 → ✅ Strong
“Save 3 hours a week on reporting”
Pain: Efficiency (2) · Transformation: Capability (2)
👉 Outcome Delight = 4 → ❌ Kill
How the angle is expressed, not what it is. To select the ad format that maximizes believability, speed of comprehension, platform-native consumption, and conversion efficiency.
“Given this angle, what is the most believable way to deliver it?”
If the format weakens belief → it is killed, even if the angle is strong.
✅ Format = the structural container through which the message is delivered. e.g. “Founder talking to camera”, “UGC testimonial”, “VO + B-roll explainer” are all formats.
Before selecting a format, lock these inputs:
If any input is missing → STOP.
Formats are selected based on what needs to be proven.
Best for existential / high-stakes pain
Best for aspirational + identity transformation
Best for control, opportunity cost, logic-driven angles
Best for TOF, scroll-stopping
⚠️ Hard rule: must still deliver the SAME angle and cannot dilute transformation.
Best for opportunity cost & switching decisions
Before approval, answer:
If any answer is NO → KILL FORMAT.
No scripts. No hooks. No visuals yet.
Master rule — “Strong angles die in weak formats. Weak angles cannot be saved by strong formats.”
The final make-or-break layer. To transform a validated SMAB into a message so clear, believable, and compelling that action feels like the natural next step.
Everything before this section only creates inputs. This section decides whether the ad earns belief and money. If this is weak, nothing else matters.
First principle — “A script is not a collection of lines. It is the precise order in which belief, emotion, and trust are constructed. Words don’t persuade by existing. They persuade by arriving in the right sequence.”
Scripts are not “good” or “bad”. They are inevitable or forgettable.
“Did we deliver the SMAB in a way that makes belief, desire, and action feel obvious?”
Agreement is not success. Nodding is not success. Momentum is success.
Forget rigid frameworks (PAS, AIDA, BAB) — they are references, not rules. Every winning script must still achieve these four outcomes, in logical order:
If these happen out of order → script fails.
Credibility must exist in the message — but it does NOT have to appear before the first claim. Some scripts start with a claim, an insight, or recognition. All are valid only if credibility is eventually established.
Acceptable forms of credibilityMay appear before the claim, during the explanation, or after (closing the loop).
Credibility closure test (mandatory): after the full script, ask “Do I now know why this person or brand knows what they’re saying?” If NO → ❌ KILL, no matter how elegant the copy is.
Pathos is not drama. It is accuracy.
Recognition beats intensity. Always.
A “core desire” = something people will pay to get or pay to stop feeling.
A. Core Desire Selector (pick ONE primary)Lines up with widely used “hardwired desire” frameworks (e.g. Whitman’s Life-Force 8) and psych research on core motives (e.g. Reiss’s 16 basic desires).
B. The “Unsayable at Lunch” Test (pass/fail)Write the desire as a raw confession:
If it’s lunch-safe, it’s usually not deep enough.
Logos is not a feature explanation. It is causal clarity.
Every script must pass this sequence test: Recognition → Reframe → Resolution → Action. Not necessarily labelled, but felt.
The script must sound like the persona’s internal monologue.
The script mustIf it wouldn’t be said out loud naturally → reject.
Platform voice proof: “This hook could be a top comment under a viral video: YES/NO.” If either is weak → rewrite before scripting.
“Can I clearly picture my life AFTER this works?”
People act on pictures, not prose.
Every word must justify its existence.
CutShorter isn’t better. Necessary is better.
Urgency must come from cost of delay, consequence of inaction, time-sensitive relevance.
Action should feel like the logical release of tension.
No exceptions.
“Did this script make me want to ACT — or just agree?”
Agreement is intellectual. Action is emotional + logical alignment. If it doesn’t create felt momentum → reject.
“Persuasion is not about what you say. It’s about what the listener is ready to believe next.”
How the message is felt, not just heard. To translate a high-conviction script and prior research into visual storytelling that reinforces belief, accelerates comprehension, increases retention, and never distracts from the SMAB.
“Do the visuals make the message easier to believe, faster to understand, and harder to forget?”
If visuals exist only to “look good” → kill.
First principle — “Visuals do not decorate persuasion. They either strengthen it or sabotage it. A great script with weak visuals underperforms. A weak script with great visuals still fails.”
Shot selection, visual sequencing, on-screen elements, editing rhythm, text overlays, sound + silence.
It does NOT decide: angle, hook, script, CTA — those are already locked.
Every visual must serve ONE primary role:
❌ If a visual serves more than one role → usually weak. ❌ If it serves none → kill immediately.
If the viewer remembers the shot more than the message → reject.
Visuals must follow the same sequence as the script:
If visuals jump ahead or lag behind the script → kill.
On-screen text is not subtitles. It is belief reinforcement.
If text competes with audio → wrong.
Editing must follow cognitive load, not aesthetics.
Sound must support meaning, not hijack attention.
If the ad still works muted → good. If it only works with sound → risky.
“Does this visual sequence help me imagine my life after this works?”
“If I removed this visual, would belief weaken?”
If the answer is NO → remove it.
“Great visuals don’t impress. They disappear while persuasion happens.”
The last line of defense before money is spent. To decide — objectively and ruthlessly — whether a creative deserves to go live.
“If this ad fails, will it fail because the market rejected it — or because we shipped something unfinished?”
If the answer is the latter → kill or rework.
First principle — “Shipping is a privilege, not a default. Every creative must earn the right to spend money. This section is not for improvement. It is for approval or death.”
It evaluates the entire creative system end-to-end: angle (SMAB), pain & transformation strength, format choice, hook + script delivery, visual execution. Individually strong parts can still fail collectively.
The pre-ship check is not creative. It is risk management.
A creative cannot enter this section unless ALL are locked:
If any are missing → STOP.
Before scoring, ask one human question:
“Would I confidently put ₹X / $X of my own money behind this creative?”
If instinctive answer is NO → kill or downgrade to rewrite. This question exists to catch what numbers sometimes miss.
No “soft launch” below 45.
“If this ad fails, what will be the most likely reason?”
If failure reason is execution-related → do not ship.
A viewer should understand what’s being sold and why within seconds.
“If someone saw this once, what would they remember?”
If recall ≠ promise → kill.
“Is there a real reason to act now — or does this feel optional?”
Urgency must come from cost of delay, relevance, momentum, consequence. If urgency is manufactured → kill or rewrite.
Every creative must end with one of these labels. No creative enters ads manager without a label.
Scale-ready
Structural
Do not test
“Testing is not an excuse to ship weak thinking. Testing is for validating strong decisions.”
I run every creative through this exact SOP before it ever touches your ad account.