Most brands running Meta ads run out of angles by month 3. The founder's original hook is exhausted, the "hero benefit" angle is fatigued, and creative testing stalls. Claude — briefed properly — can generate 10-20 net-new landing page angles a week, grounded in real customer language, not hallucinated benefit lists.
Below are the 6 prompt patterns we use to feed the Meta ad testing pipeline. Each is a copy-paste template with a filled-in example.
Table of Contents
- Why angle testing beats creative testing
- Prompt 1 — Mine customer reviews
- Prompt 2 — Comment excavation
- Prompt 3 — Competitor gap angles
- Prompt 4 — Situational trigger angles
- Prompt 5 — Objection-reversal angles
- Prompt 6 — Category redefinition
- How to sequence them in a week
Key Takeaways
| Question | Answer |
|---|---|
| What's an "angle"? | The specific pain, situation, or transformation the ad + landing page speak to. Not the creative format. |
| How many angles should I test at once? | 3-5 running concurrently. Kill losers weekly. |
| How does Claude help? | Turns raw customer/comment/review data into angle candidates faster than a human brainstorm. |
| What data do I feed it? | Reviews, ad comments, competitor pages, customer surveys, CS ticket themes. |
| How long is the angle test? | 2-3 weeks per angle to get a clear read. |
| Where do winning angles go next? | Dedicated landing page + 3-5 creative variants at that angle. |
Why angle testing beats creative testing
Two brands running the same budget with the same creative team: Brand A tests 8 new creatives per week, all on the same angle. Brand B tests 2 new creatives per week, but 3 different angles. Brand B wins.
Reason: creative variance within an angle produces small lifts. Angle variance produces order-of-magnitude lifts. A brand that finds one new working angle per quarter grows faster than a brand that iterates on creative execution of one exhausted angle.
Claude's role: generate the angle candidates faster than a human brainstorm can. You still test them; Claude just gets you to the test.
Prompt 1 — Mine customer reviews
The best angles come from customer language, not marketer language. Prompt Claude with the last 100 reviews:
BRAND: [product name]
CATEGORY: [category]
I'm going to paste 50-100 recent customer reviews below.
Read them and extract:
1. The 5 most common SPECIFIC use cases customers mention (not
generic "I like it" — actual situations they were in).
2. The 5 most common transformations customers describe
("finally stopped X", "now I can Y", "no more Z").
3. The 5 most common surprises — things customers didn't expect
but noticed after buying ("didn't realize it would...").
4. The 5 most common comparisons — competitors or old solutions
they were using before.
For each, quote 1-2 short customer phrases verbatim.
Then propose 5 landing page angles I could test on Meta cold
traffic, each grounded in one of the extracted patterns.
For each angle, give me: (a) 1-line angle summary, (b) headline
draft in the customer's language, (c) which extracted pattern
it's based on.
REVIEWS:
[paste 50-100 reviews]
Output: 5 angle candidates, each grounded in verbatim customer language. These convert.
Prompt 2 — Comment excavation
The comments on your existing Meta ads are a goldmine most brands ignore. Both positive and negative:
BRAND: [product name]
I'm pasting comments from our 3 top-spending Meta ads over
the last 60 days. Some positive, some negative, some questions.
Extract:
1. The 3 most common OBJECTIONS in the comments.
2. The 3 most common QUESTIONS about the product.
3. The 3 most common SITUATIONS commenters describe themselves
in ("I always...", "when I...", "my problem is...").
4. The 3 most common COMPARISONS to other brands or solutions.
Then propose 3 landing page angles that:
- Explicitly address one of the top objections in the hero,
- Answer one of the top questions in the "how it works" section,
- Speak to one of the described situations in the problem block.
For each angle, draft a hero H1 (max 12 words) + subhead
(max 20 words).
COMMENTS:
[paste 200-500 comments]
Objection-first angles convert well on brands that have been running for 3+ months and have real skepticism to overcome.
Prompt 3 — Competitor gap angles
Find angles the competition isn't running:
BRAND: [product name] — sells [1-line description]
KEY COMPETITORS: [3-5 competitor names]
I'm going to describe each competitor's Meta ad angles based on
what's in their ad library (I'll paste 3-5 ad copies per
competitor).
Extract:
1. What angle is each competitor leaning on hardest?
2. What angles are 2+ competitors both running (saturated)?
3. What customer situations / pains are NOT being addressed by
any of the 3-5 competitors?
For the gaps, propose 5 landing page angles our brand could test.
For each: (a) the gap it fills, (b) why competitors are missing it,
(c) draft hero H1 + subhead.
COMPETITOR AD COPY:
[paste]
Gap angles are how emerging brands break into saturated categories. They convert on the audience the incumbents have already warmed up but aren't retaining.
Prompt 4 — Situational trigger angles
People buy when something happens in their life. Prompt for the triggers:
BRAND: [product name]
CATEGORY: [category]
Someone would buy [product] because a specific situation just
happened in their life. List 15 situational triggers — real,
specific, dated events — that would make a person shop for
this product today.
Examples of what I want (not from our category):
- Just moved into a new apartment (furniture)
- Kid started daycare (skincare, laundry)
- Just started a new job with commute (bags, coffee, apparel)
- Doctor mentioned cholesterol (food, supplements)
- Went on first hike of the year (outdoor gear)
For each of the 15 triggers, propose:
- A hero H1 that speaks to that trigger
- What time of year / life stage it's most likely to be relevant
Situational triggers give you 15 different targeting + angle pairs. Test the top 3-5 per quarter.
Prompt 5 — Objection-reversal angles
Turn the top objection into the hook:
BRAND: [product name]
TOP 3 OBJECTIONS: [list — from your CS tickets, comments, reviews]
For each objection, write a landing page hero that DIRECTLY
LEADS with the objection and reverses it as the hook.
Example:
Objection: "too expensive"
Hero: "Yes, it costs 2x more. It also lasts 6x longer. Here's the math."
Objection: "one more subscription"
Hero: "Not a subscription. Buy once, refills only when you need them."
Give me a hero + subhead for each of my 3 objections.
Objection-reversal is high-conversion because it disqualifies the wrong customer fast and pre-answers the doubt for the right one.
Prompt 6 — Category redefinition
Redefine the category the customer thinks they're shopping in:
BRAND: [product name] — sells [description]
Most customers think they're shopping for [obvious category].
Propose 5 alternative category framings — reframes that would
change what the customer is comparing us to.
Example (not from our category):
- "Insulated commuter backpack" → "Wearable laptop insurance"
- "Meal delivery kit" → "The cheapest way to eat well without cooking"
- "Sleep supplement" → "The cheapest thing that fixes the 2am wake-up"
For each, give me:
- The reframe
- The customer this reframe attracts
- A hero H1 for the landing page at this reframe
Category redefinition angles are the rarest and highest-value — a working one can carry a brand's paid acquisition for 6-12 months.
How to sequence them in a week
The full cycle we run for a client:
- Monday — Prompt 1 (reviews) + Prompt 2 (comments). Extract 10 angle candidates.
- Tuesday — Prompt 3 (competitors). Add 3-5 gap angles to the pool.
- Wednesday — Prompt 4 (triggers) OR Prompt 6 (redefinition). Add 5 more.
- Thursday — Score all 20 candidates. Pick 3 to build landing pages for.
- Friday — Ship 3 landing pages via the Claude landing page workflow.
- Next week — Meta ads launched, 2-3 weeks of testing.
- Every quarter — Prompt 5 (objection reversal) as a refresh pass on losing angles.
This is one operator's week. It replaces the "we can't figure out what to test next" plateau every Meta account hits.
Talk to Branva
Angle-generation + landing page + ad-copy pipeline is exactly what our team runs as a service. If your Meta account is stuck on one angle, book a working session and we'll show you 5 fresh angles pulled from your own reviews and comments in ~30 minutes.
Related: the essential types of ecommerce ads — the format-by-format breakdown of where each ad type fits.
Frequently Asked Questions
How do I score which angles to test first?
Weight by: (a) how much verbatim customer language backs it, (b) how differentiated from what you currently run, (c) how easy the landing page is to build. Start with the ones scoring high on all three.
Should I test one angle or one creative variant?
Angle > creative for growth. If you're stuck, angle first, always.
What if I don't have customer reviews yet?
Use Prompt 3 (competitors) and Prompt 4 (triggers) first. Once you have 100+ orders, come back to Prompts 1 and 2.
How often do I refresh angles?
Test cadence: 3 new angles per month, kill 1-2 losers per month. Working angles stay in rotation for 3-6 months before fatigue.
Related reading
- How to build Meta ads landing pages using Claude — where angles become pages.
- How to build Shopify landing pages using Claude — the general landing page workflow.
- How to connect Meta ads to Shopify using Claude — the tracking layer.
- Brand Brain — the canonical source of truth — where angle library lives.
- The Paid Growth pillar — the systems behind scaling paid.