BRAND CONTEXT

Stage 9 — Advertising Platform Guidelines

prayer/brand-context-st-peters/09-advertising-compliance.md

Stage 9 — Advertising Platform Guidelines

Brand: St. Peter's Basilica Prayer Placement (v1) | Audience: US + Canada, women 40+, Catholic / devotional | Written: 2026-09-14 Upstream: Stage 8 (08-copy-guidelines-hook-bank.md) — the 20 scored hooks, the power-words and forbidden-words lists (§1.6–1.7), the Final Quality Checklist (§8). Stage 4 (04-offer-brief.md) — the mechanism and the 15 named objections/hesitations (R1–R6, S1–S4, C1–C2). COPY-CREATIVES.md — the live copy this document audits. evidence/2026-09-14/competitor-ads.md — what the category is actually running, and what Meta is actually approving, right now.

What this document is. Ad-operations guidance. What Meta's ad review actually rejects in this category, why, and the exact rewrite that clears it. A media buyer opens this file, checks copy against it, and ships.

What this document is not. Legal advice. There is no legality analysis, no "consult an attorney," no jurisdictional note, no compliance-resources link list anywhere below. If a claim in the source template read as a legal disclaimer, it was dropped, not softened. This is a platform-policy and ad-review document, full stop.

1. Category Classification

For ad-review purposes, this brand is: a paid devotional intercessory service with physical fulfillment, sold as a one-time purchase plus a recurring subscription, in a category with a documented enforcement history.

Three things stack on top of each other and every one of them changes how review treats the account:

  1. Religion-adjacent commerce. The product is explicitly and unavoidably Catholic — the site, the saint, the rite. Meta classifies this under its Personal Attributes policy the moment copy implies anything about the viewer's religion, not merely because the product itself is religious. Selling a Catholic product is not restricted. Implying the viewer is Catholic is.
  2. Recurring billing. The Monthly Prayer Membership ($9.90/mo or $79/yr) puts every ad that funnels to it under the same account-health scrutiny as any subscription commerce offer: undisclosed recurring charges are one of the highest-volume sources of user complaints, and user complaints are what move an ad account's health score, independent of whether any single ad gets rejected.
  3. A prosecuted category. competitor-ads.md places this brand in the same mechanic as the Washington AG v. Rogovy prayer-fulfillment prosecution's target category — write-a-prayer, we-place-it-at-a-holy-site, photo-confirmed. That history doesn't change what Meta's ad-review classifiers do, but it is why every operator in the space that survives long enough to be studied (competitor-ads.md, Holy Land Prayer) carries a "service fee, not a donation" disclaimer somewhere — usually buried on the landing page, never in the ad. This brand's decision to put that line in the ad's own primary text (Hook 20) is ahead of the category, not behind it.

Platforms in scope: Meta (Facebook + Instagram) is primary and the only channel this document treats in depth — it is where the funnel currently runs and where all documented competitor activity lives. Google and TikTok are covered briefly in §2 for the day either is considered.

2. Platform Policies

2.1 Meta — in depth

Targeting. Meta removed religion from detailed targeting on 2024-03-18, and removed religion-based exclusion targeting on 2025-03-31. Both routes are gone. You cannot build an interest audience of "Catholics," and you cannot even exclude non-Catholics from a broad audience. The brand's plan of broad US+CA, women 40+ is not a strategic preference — it is the only route that exists. This is the single most important operational fact in this document, and it has one direct consequence for copy: targeting used to do the qualifying work of finding a Catholic audience. It cannot do that anymore. Creative has to do all of it, by resonance, without ever stating it. Every hook in Stage 8's bank that works ("worn smooth," "ordained by now," "the bronze foot") does this correctly — it uses Catholic-coded behavior and imagery to self-select the right reader, never a Catholic-coded label applied to the reader.

Personal Attributes (the #1 rejection driver for this category). Meta's ad policy prohibits content that asserts or implies knowledge of a person's race or ethnicity, religion or philosophical belief, age, sexual orientation or gender identity, disability, medical condition (including mental health), financial status, union membership, criminal record, or name — directly, based on their being a member of the audience the ad is served to. For this brand, three of those categories are live risks, not one:

  • Religion — the obvious one. "As a Catholic, you know…" asserts it. "Many Catholics find…" or "For anyone who has ever prayed a novena…" does not, because it describes a group or a behavior rather than diagnosing the viewer.
  • Financial status — less obvious, and it is live in this brand's own ad kit today. COPY-CREATIVES.md Ad 3 opens "You've done the math at the kitchen table a dozen times. You've prayed over the bills before opening them." That is an unqualified assertion that the viewer is in financial distress. It is the same policy violation as a religion assertion, just aimed at a different protected category. See the audit in §9.
  • Medical condition — also live. Ad 1 opens "You've prayed at her bedside. You've prayed in the hospital parking lot." — a direct, unconditional assertion that the viewer has a family member with a medical condition serious enough for a hospital stay.

The mechanical fix is the same for all three, and it does not require abandoning second person. Brand voice (Stage 8 §1.3) requires second person singular, always — so the fix is not to shift to third person or a group address, which the brand's own voice rules forbid anyway. The fix is assertion → condition. "You've prayed at her bedside" (assertion, states a fact about the viewer) becomes "If you've ever prayed at a bedside" (condition, invites self-identification without diagnosing it). See the twelve worked rewrites below — this is the highest-value section in this document.

Special Data Categories and the pixel/CAPI implication. Religion is one of Meta's Special Ad Data Categories. For a religion-adjacent advertiser this means two things beyond creative: (1) never build a Custom Audience or Lookalike from a data source that encodes religion (e.g., a list scraped or purchased by parish, diocese, or "Catholic donor" tag) — Meta's system-level enforcement on this is separate from and stricter than the creative-review enforcement above, and can suspend ad-account access rather than just reject a single ad; (2) event data sent via Conversions API for this brand (purchase, add-to-cart on the candle bump, subscription-start) should never carry a custom parameter that encodes the buyer's inferred religion, denomination, or parish — a "prayer_type" or "novena_name" field is fine, a "religion" or "faith_affiliation" field is not, even server-side.

Page and account health. Two operator-visible facts from competitor-ads.md matter here: (1) six of the eight known competitor pages/domains have gone dark since the last pull — a category with unusually high advertiser churn, most plausibly from account-health spirals rather than the offers failing; (2) the two survivors (Holy Land Prayer, Lourdes: Springs of Grace) run the same script for weeks at a time rather than iterating fast, which is consistent with an account that has learned not to trigger repeat manual review with rapid creative churn. Practical takeaway: once a script clears review, leave it running rather than re-submitting variants for no reason — every submission is a fresh classifier pass, and volume of submissions is itself a signal reviewers weight.

Subscription commerce. The Monthly Prayer Membership CTA needs price, billing frequency, and the cancel path visible before the click that starts it, on whichever surface carries that CTA (thank-you page, Email 5, landing footer). This is covered in full in §6.

2.2 Google Ads — briefly

Google restricted personalized-advertising signals for sensitive categories (religion among them) years before Meta did, so the same "broad targeting only" constraint already holds by default — there is no religion affinity or in-market audience to lose. Google's live risk surface for this brand is its Misrepresentation policy (unrealistic outcomes, unsubstantiated claims) and its general prohibition on exploiting a personal hardship or crisis in ad copy for commercial gain — which maps directly onto the same financial-status and grief-lane risk covered under Meta's Personal Attributes above, just enforced under a differently-named policy. Treat any Google Search/Display expansion as needing the same personal-attributes and outcome-claim discipline as Meta, with no targeting compensation available either.

2.3 TikTok Ads — briefly

TikTok's advertising policy treats religion and places of worship as a restricted industry in a meaningful number of markets — stricter than Meta's default posture, sometimes requiring pre-approval or additional certification rather than standard automated review, and in some regions prohibiting the category outright. If TikTok is ever added as a channel, budget for a manual-review cycle before the first dollar spends, and expect the personal-attributes and outcome-claim discipline in this document to be a floor, not a ceiling — TikTok's human reviewers in this vertical run more conservative than Meta's classifier.

3. Banned Phrases — NEVER USE

Organized by the risk each one carries. Every entry maps to a real rule above or to a rule already binding in Stage 8 §1.7 — this section restates the ones with direct ad-review consequences and adds the ones specific to platform enforcement.

3.1 Outcome and miracle claims

And miracles have followed          Your prayer will be answered
He will call / she will forgive you / it will heal
Guaranteed results                  Proven to work
This changed everything for me      God answered within days

The category leader (Holy Land Prayer, per competitor-ads.md) runs exactly this in its live ad copy — "through his grace they were answered" — while its own landing page disclaims outcomes entirely. That contradiction is the leader's exposure, not a model to match. This brand's own positioning block already forbids all of the above; the point here is that Meta's Misleading Claims enforcement will catch the ones the brand's own rules miss.

3.2 Personal-attribute assertions

As a Catholic, you...                Your faith has been tested
You've been struggling financially   Your marriage is falling apart
You're going through a health scare  As a grieving mother, you...

See §2.1 and the twelve rewrites below. This is the single largest rejection category for this brand specifically.

3.3 Medical claims through a devotional lens

Heal your anxiety through prayer     Cure [condition] with a placed prayer
Prayer as treatment for depression   Spiritual healing for [medical condition]

Never appears in the current kit. Flagged here because the financial- and health-adjacent angles (Ads 1, 3) sit one careless edit away from crossing into this.

3.4 Church-affiliation implications

Blessed by the Vatican               Endorsed by St. Peter's Basilica
An official Church prayer service    Sanctioned by the diocese
Your priest recommends this          A Vatican-approved prayer placement

The brand's own Positioning Block already states "We are not the church... not St. Peter's Basilica, the Vatican, or any diocese, and we do not speak for them." That line exists precisely to make the above list unnecessary in copy — but it is a landing-page/footer line today, not present in any of the eight ad primaries. Meta reviewers checking a religion-adjacent advertiser for implied institutional endorsement will look at the ad and the landing page together; keeping this disclaimer landing-page-only rather than restating it is a gap, not a violation, but it is worth closing (see §6).

3.5 False urgency (the category's worst-operator tell)

Only [N] slots left today            This offer expires at midnight
80% OFF — today only                 Room for one more (Lourdes' own live line)
Price increases tomorrow             Last chance before it's gone

competitor-ads.md documents this exact device live: note_to_the_wall's "80% OFF," and Lourdes' "There's still room for one more" tied to a Saturday pilgrimage group. Both operators sit in the "Light-medium" saturation band for urgency — meaning it isn't even universal in this category, and this brand's own rules (Stage 8 §8, item 10) already forbid it outright: "Urgency originates only from her calendar or a person she named." A named Thursday (Hook 8, Hook 15) is allowed. A manufactured slot count is not, on brand grounds and on ad-review grounds simultaneously.

4. Risky Phrases — USE WITH CAUTION

Risky Phrase The condition that makes it safe vs. unsafe
"A candle is lit beside it." Unsafe as an unqualified statement of fact — it appears, unqualified, in all eight of the brand's current live ad primaries (§9), and Stage 8 already ranks the candle-availability claim the worst of nine unresolved findings, directly contradicted by a first-hand account ("when we went today there weren't any," reddit-voc.md). Safe only once fulfillment confirms it as standing operating practice, or rewritten candle-free per Hook 10/11's pattern (itemize what is certain: printing, carrying, placing, photograph). Until resolved, treat this as a Misleading Claims risk, not just a brand-consistency one — an unsubstantiated factual claim about physical service delivery is exactly what that policy exists to catch.
"Forever resting" / "para siempre" Unsafe as a permanence claim with nothing behind it — no surface in the brand's kit states the card is placed permanently rather than cleared on a schedule (as real basilica prayer intake generally is). Safe only if reworded to describe the photograph as permanent ("the photo is yours to keep") rather than the placement.
"There's one more step of faith you can take." Safe as written — a call to action framed as a personal choice, not a promise. Becomes unsafe only if a future variant attaches a number or timeframe to what that "step" produces.
"No one will promise you he'll change. We certainly won't." (Ad 2) Safe — this is the brand doing the Belief-5 refusal correctly, and it is a genuinely strong pattern. Stays safe only as long as it is never followed, anywhere in the same asset, by a testimonial that quietly implies the opposite.
"You've prayed so much you could have been ordained by now." (Hook 6) Safe — names a behavior (praying, a lot) and a wry self-aware joke, never a status. Would become unsafe only if rewritten to "As someone this devout, you..." — the moment behavior becomes an assigned identity.
Grief-lane conditional openers ("If you're carrying a loss...") Safe as a condition. Becomes unsafe the moment "if" is dropped for punch under time pressure — watch for this specific edit in fast-turnaround creative reviews; it is the most likely place brand voice and ad-review protection quietly diverge.
Financial-hardship imagery ("the rent," "the number that won't leave your head") Unsafe used the way Ad 3 uses it today — as a direct, unqualified assertion. Becomes usable only inside a conditional frame ("if the number on the bill has ever kept you up") or removed to a general statement about a shared human experience rather than the viewer's specific finances.
"You'll see it with your own eyes." / "you'll see exactly where it rests." Safe — this is a promise about the deliverable (a photograph), which the brand controls and can keep, not a promise about an outcome it cannot control. The distinction is the whole differentiation; keep it sharp in every rewrite.

5. Safe Phrases — APPROVED

Pulled directly from Stage 8's 20-hook bank (all 5/6 or 6/6) and the brand voice guidelines. A media buyer can paste any line below without a second look.

Headlines (from live ad kit + hook bank, ≤40 characters where noted)

Her Name, Carried Into the Basilica
Your Marriage Prayer, Resting in Rome
Carry Your Prayer Into St. Peter's
Place Your Child's Prayer in Rome
Their Name, at Rest in the Basilica
Nobody Ever Told You What Happens to Left Prayers
There Is a Difference Between Unanswered and Unwitnessed

Primary text openers (curiosity / pain, cold-safe, no personal-attribute risk)

Nobody ever told you what happens to the prayers people leave at holy places.
There is a difference between a prayer unanswered and a prayer nobody ever witnessed.
Centuries of hands wore that bronze foot smooth — your prayer has worn smooth too.
You have prayed so much you could have been ordained by now.
Grief doesn't get smaller — it gets quieter, and it still has nowhere to go.
There are prayers you only say alone.
A mother never stops praying for her children — not at 2, not at 22, not ever.

Mechanism / proof lines (candle-free, production-method-safe)

Your $34.90 covers the printing, the carrying, the placing and the photograph — nothing else.
One man carries your card in, sets it down, and photographs exactly where it rests.
Not a picture of somebody else's prayer — your words, legible, where they were set down.
Your prayer card, photographed where it rests inside St. Peter's, in three days.

Objection / boundary lines (the brand's strongest differentiation — close every cold ad with one)

This is a service fee for carrying and photography — never a donation, never a promise.
I cannot promise you how God answers a prayer — no honest man can.
A shrine's own site takes your intention free — that is the sending, not the seeing.
We promise the placing — we never promise the outcome.

CTAs (already in use, no edit needed)

Learn More
Send Message

6. Required Disclosures

Platform-protective only — nothing legal, nothing jurisdictional.

  1. The not-affiliated line, restated in the ad, not just the landing page. The brand's own Positioning Block already carries "We are not the church... not St. Peter's Basilica, the Vatican, or any diocese." That line currently lives on the landing page and FAQ only. For a religion-adjacent advertiser, having it appear nowhere in the ad itself is a gap worth closing on at least one surface per funnel — the offer-page persona letter or the FAQ block reached from the ad's "Learn More" click is sufficient; it does not need to be in the primary text.
  2. Clear pricing before checkout. Already the brand's strongest discipline (Hook 10's itemization) — keep it exactly this specific everywhere: printing, carrying, placing, photograph, $34.90, nothing else.
  3. Subscription terms stated before the subscription CTA. The Monthly Prayer Membership's price ($9.90/mo or $79/yr), billing frequency, and cancel path must be visible on whatever surface carries the "Join" or "Keep It Carried" CTA — the thank-you page and Email 5 both currently pitch the membership; confirm both state price and frequency above the button, not only in the FAQ.
  4. Honest, single delivery window. This is a live, unresolved contradiction and it is an ad-review risk, not only a copy bug: Stage 8's hook bank locks "three days" as "the single most important lever in the offer," requiring it "byte-identical everywhere it appears" (Hook 9's note). COPY-CREATIVES.md, meanwhile, states "most within 5–7 days" on the landing page's "What to Expect" block, in the FAQ, and in the delivery email template (three separate instances). Ad review checks the ad's promise against the landing page's promise; today, no ad currently states a delivery window on its face, but the moment one does — and Hook 9 is written specifically to be used as an ad headline — it will contradict the page it lands on. This must be resolved to one number before any ad using a delivery-window claim goes live.

7. Pre-Flight Checklist

Run before any ad in this account goes live. Every box, every time — this is not a sampling exercise.

PERSONAL ATTRIBUTES
□ No sentence states a fact about "you" regarding religion, medical condition,
  financial status, or relationship status without an if/when/many/anyone-who
  qualifier ahead of it.
□ Behavior is named ("you've prayed," "you've done the math"); status is not
  ("you're Catholic," "you're broke," "you're sick").
□ Third-party pain (her, him, the child) is used instead of first-party medical
  or financial assertion wherever the angle allows it.

OUTCOME CLAIMS
□ No verb about the prayer points forward — no "will be answered," "will heal,"
  "will call," "will change."
□ Every cold ad closes with a categorical boundary line (Hook 20 or the
  Positioning Block), not a soft sign-off.
□ No testimonial implies an outcome, even indirectly ("and then he called").

FACTUAL / CANDLE / PRODUCTION-METHOD
□ The candle line, if present, is either confirmed as standing practice or
  rewritten candle-free per Hook 10/11.
□ No claim about how the photograph was produced, in either direction — no
  "real photograph," no "not a rendering," no "AI-generated."
□ No permanence claim ("forever," "para siempre") attached to the placement
  itself, only (if at all) to the photograph as a keepsake.

LANDING-PAGE / CHECKOUT CONSISTENCY
□ Price, delivery window, and what's promised on the ad match the destination
  page exactly — check the live 3-day vs. 5–7-day conflict specifically until
  it is resolved brand-wide.
□ Subscription price, frequency, and cancel path appear before the CTA that
  starts the Membership.

HOOK BANK / VOICE
□ Every headline and every primary-text claim traces to a hook scored 5/6 or
  6/6 in Stage 8 §2, or to the Positioning Block. If it's not in the bank,
  it gets a second look before it ships.
□ Second person singular throughout; the first noun in third-party creative
  is a person, not a product.
□ Nothing from Stage 8 §1.7's forbidden list appears anywhere (miracle,
  unlock, guaranteed, healing journey, loved one, prayer warrior, etc.).

URGENCY / COMPETITORS
□ No countdown, slot count, or price-increase warning. Any date used is hers
  (an anniversary) or a real operational one (a named weekday), never invented.
□ No competitor named, anywhere, in any customer-facing asset.

8. Common Rejection Reasons & Fixes

The rejections this account will realistically hit, given the category and the current copy.

Rejection reason (Meta's language) Why this brand hits it The fix
"Personal attributes" Ad 3's opener directly asserts the viewer's financial distress; Ad 1's opener directly asserts a family member's medical condition. Both are unconditional "you've..." statements with no qualifier. Add the conditional opener: "If you've ever done the math at the kitchen table more times than you wanted to admit —" (Ad 3) and "If you've ever prayed at a bedside, or in a hospital parking lot because it was the only private place left —" (Ad 1). See §9 for the full rewrite.
"Misleading claims" / unsubstantiated statements The unqualified candle line, present in all eight current ads, and the "forever resting" / "para siempre" permanence claim in Ad 5 and its Spanish variant. Either confirm the candle as standing operating practice (then it is a fact, not a claim) or drop to the candle-free itemization pattern (Hook 10/11). Reword "forever resting" to describe the photograph, not the placement.
"Unacceptable business practices" (exploiting a personal crisis) Any future edit of the grief or financial angles that drops the "if that's you" conditional for a punchier, fully assertive opener — the exact edit a fast creative-testing cycle tends to make. Keep the conditional. If a variant needs to be shorter, cut elsewhere in the sentence, never the qualifier.
Landing-page mismatch (ad promises don't match destination) The three-day promise (locked in Stage 8 as the single most important lever) versus the 5–7-day window stated on the landing page, FAQ, and delivery email. Resolve to one number before any ad states a delivery window on its face. This is upstream of ad copy — it is a fulfillment-ops decision that ad ops cannot fix alone.
Subscription / recurring-billing complaints If the Membership CTA on the thank-you page or Email 5 doesn't show price and frequency before the click, buyers who forget they subscribed generate chargebacks and "I didn't authorize this" reports — these hurt account health even when no single ad gets rejected. State price, frequency, and cancel path directly above the button, not only in a linked FAQ.
Extended "under review" hold on first submissions New advertisers in a religion-adjacent, recurring-billing category are more likely to draw an initial manual-review hold beyond the typical automated turnaround, independent of whether the copy is compliant. competitor-ads.md shows the two long-running survivors keep the same creative live for weeks rather than iterating fast — plausibly a learned response to this. Don't resubmit or duplicate the ad to "try again" during a hold — repeated submissions of the same or near-identical creative read as pattern behavior to the review system and slow things down further. Wait out the stated review window. If genuinely rejected and the copy is clean against this document, use the single in-platform Appeal action once, citing the specific policy line misapplied — never argue the theology of the offer in an appeal, argue the specific claim.
Account-health spiral after a real violation Once an ad is actually pulled for cause (not a hold), the instinct to relaunch the same script under a new ad ID to dodge the flag is what turns one rejection into an account-health problem — Meta's system tracks this pattern across ad IDs within an account. Fix the specific line, keep the same ad ID/edit history where the interface allows it, and resubmit once. Treat every rejection as informative about the account's review history, not just that one ad's fate.

9. Live Copy Audit — the 8 Meta Primaries in COPY-CREATIVES.md

Every one of the five English and three Spanish primaries, audited against §§3–7 above.

AD 1 — Parent Healing

Verdict: PASS WITH EDIT.

  • Risk line: "You've prayed at her bedside. You've prayed in the hospital parking lot." — unqualified assertion of a family member's medical condition. Personal Attributes risk.
  • Risk line: "A candle is lit beside it." — unqualified candle claim, present without qualification.
  • Rewrite (opener): "If you've ever prayed at a bedside, or in a hospital parking lot because it was the only private place left, you know the words start to wear smooth."
  • The close — "This is a service fee for printing, carrying, placing, and photography — not a donation, and no outcome is promised or sold. What we can promise: your prayer will be there." — is exactly right and needs no change.

AD 2 — Marriage Restoration

Verdict: PASS WITH EDIT.

  • Risk line: "For the marriage you still believe in even when it's hard to remember why." — a softer, partially-qualified version of the same pattern (the "if that's you" qualifier arrives one sentence later rather than before). Lower risk than Ad 1/3 but worth tightening.
  • Risk line: "A candle is lit beside it." — same candle issue as all eight.
  • Rewrite: move the conditional earlier — "There's a marriage some of us still believe in, even when it's hard to remember why. If that's you —"
  • "No one will promise you he'll change. We certainly won't." is the strongest anti-outcome line in the entire kit. Keep it verbatim.

AD 3 — Financial Provision

Verdict: REJECT RISK — highest risk of the eight.

  • Risk line: "You've done the math at the kitchen table a dozen times. You've prayed over the bills before opening them." — unqualified, direct assertion of the viewer's financial distress. This is the clearest Personal Attributes violation in the set; financial status is an explicitly named protected category.
  • Reinforcing risk: "the rent, the business, the job, the number that won't leave your head" — repeats and sharpens the same assertion.
  • Risk line: "A candle is lit beside it."
  • Rewrite: "If you've ever done the math at the kitchen table more times than you wanted to admit, or prayed over a bill before you opened it — there's one more step of faith you can take."
  • This ad should not go live until the opener is rewritten. Everything after the second sentence is sound.

AD 4 — A Child's Safety

Verdict: PASS WITH EDIT — the cleanest opener of the five English ads.

  • "A mother never stops praying for her children" is a general statement about a role, not an assertion about the viewer, and "If you've been praying for your child's safety and waiting" is correctly conditional. No rewrite needed on the opener.
  • Risk line: "A candle is lit beside it." — the only fix this ad needs.
  • "We promise the placing — we never promise the outcome" is a strong, clean line. Keep it.

AD 5 — Grief & Loss

Verdict: PASS WITH EDIT.

  • Opener is sound: "Grief doesn't end. It just gets quieter" is general, and "If you're carrying a loss" arrives early and correctly conditioned.
  • Risk line: "A candle is lit beside it."
  • Risk line, on-screen text: "THEIR NAME. AMONG THE CANDLES. FOREVER RESTING." — "forever resting" is an unsubstantiated permanence claim about the placement itself.
  • Rewrite (on-screen text): "THEIR NAME. AMONG THE CANDLES. YOURS TO KEEP." — shifts the permanence claim to the photograph, which the brand actually controls.
  • "We won't promise your heart will heal. We promise your prayer will be there" is the best line in the set. Keep it.

AD 1-ES — Sanación de un Padre

Verdict: PASS WITH EDIT. Same structure and same fix as Ad 1: "Has orado junto a su cama. Has orado en el estacionamiento del hospital" needs the conditional opener ("Si alguna vez has orado junto a una cama..."), and "Se enciende una vela a su lado" carries the same candle risk.

AD 4-ES — Seguridad de un Hijo

Verdict: PASS WITH EDIT. Opener already conditional and clean, matching the English version. Only fix needed: the candle line, "Se enciende una vela a su lado."

AD 5-ES — Duelo y Pérdida

Verdict: PASS WITH EDIT. Opener is clean. Two fixes: the candle line, and the on-screen text "PARA SIEMPRE" — same permanence-claim risk as the English version, same fix (reframe to the photograph as keepsake).

Summary across all eight

  • Pass as written: 0 of 8. Every ad carries the unqualified candle line.
  • The single most common failure: the unqualified "A candle is lit beside it" / "Se enciende una vela a su lado" claim — present in all eight ads, unresolved per Stage 8's own flag as the worst of nine open findings, and directly contradicted by a first-hand account in the evidence pack. This is the one fix that, made once and applied everywhere, clears every ad in the kit of its most common problem.
  • The sharper, narrower risk: Personal Attributes exposure in Ads 1 and 3 (and 1-ES) — both open with unconditional assertions about the viewer's medical or financial circumstances. Ad 3 is the single highest-risk asset in the kit and should not run until rewritten.
  • The best-constructed ad in the set is Ad 4 / Ad 4-ES — conditional framing from the first sentence, no personal-attribute assertion anywhere, and the cleanest anti-outcome close ("We promise the placing — we never promise the outcome"). It is the template the other four should be edited toward.

End of Stage 9. This document takes the copy-side voice rules and forbidden list from Stage 8 as its floor and adds the platform-enforcement layer on top: what Meta's classifiers and reviewers specifically look for in a religion-adjacent, subscription-commerce advertiser, and the exact rewrites that keep this brand's already-ahead-of-category honesty framing from being undercut by an unqualified opener or an unconfirmed factual claim.