How to build Shopify landing pages using Claude

How to build Shopify landing pages with Claude

Landing pages are where most Shopify brands leak revenue. Meta and Google send warm traffic; a generic product page fumbles it. Building dedicated landing pages used to require a designer + copywriter + developer. Claude collapses most of that — if you prompt it correctly.

Below is the exact workflow: how to structure a landing page prompt, the 6 section templates we reuse, and how to iterate to a page that converts and reads on-brand.

Table of Contents

Key Takeaways

Question Answer
Why not just send ads to the product page? Product pages sell inventory; landing pages sell an offer. Different job, different structure.
What does Claude actually produce? Fully-written HTML sections you paste into GemPages, PageFly, Shopify sections, or a custom template.
How long is a landing page? 6-9 sections. Anything shorter is a product page; anything longer overwhelms.
What's the biggest miss most brands make? No offer, no proof, no urgency. Claude can't invent these — you have to brief it.
How long does one page take? 60-90 min from brief to live. First one takes 2-3 hours.

Why product pages ≠ landing pages

Your Shopify product page has to serve everyone — organic search visitors, email clickers, returning customers, and paid traffic. It's a compromise. A landing page has one job: convert a specific piece of traffic into a specific outcome.

Structural differences:

Product page Landing page
Full navigation Minimal / no navigation
All variants shown Pre-selected variant
Universal copy Angle-specific copy (tied to the ad or campaign)
Ships-to-anyone shipping info Offer-specific urgency
Optimized for SEO Optimized for the incoming click's intent
One layout Different layouts per angle / audience

You can't make one page do both jobs well. That's why landing pages exist — and why Claude + a proper prompt is such a leverage move: you can spin them up quickly instead of never getting to it.

The offer brief you feed Claude

Every landing page prompt starts with a 1-page offer brief. Claude cannot invent this — you must brief it or the output is generic. The brief has 8 fields:

1. Traffic source & angle. "Meta ad, cold audience, women 35-55, angle = 'stops toddler tantrums in the car.'"

2. The offer. "20% off first order + free shipping over $50 + free returns" or "buy 2, get 1 free through Sunday."

3. The product. Full spec sheet. Name, price, materials, hero features (3-5), what's in the box.

4. Hero benefit. One sentence — the transformation, not the feature. "Your toddler falls asleep in the car within 5 minutes."

5. Objections. The 3 top reasons this traffic won't buy. "It's expensive; I already own something similar; will it fit my car seat."

6. Proof. Reviews (paste 5-10), press mentions, before/after examples, guarantees, credentials.

7. Urgency. Real deadline, low stock, launch window, promo end date. If you don't have real urgency, say so — Claude will not fabricate.

8. Voice. The voice doc — same one you use for product descriptions.

Missing any of these and Claude fills in with generic filler. Every field.

The 6 landing page sections

Claude produces one section at a time (better output than trying to write the whole page in one prompt):

1. Hero. Headline + subhead + primary CTA + hero image or video. 15-25 words total. Answers "why should I keep scrolling."

2. Problem block. 2-3 short paragraphs on the pain the customer is in RIGHT NOW. Anchored to their language, not yours. Ends with an implied "there's a better way."

3. Solution block. How this product solves it. 3-4 short paragraphs. Include the hero benefit and 2-3 supporting benefits. Product image next to copy.

4. Proof. Reviews (3-5), before/after, press logos, guarantee. This is where trust is built or lost. The best proof section overweights specifics — "Sarah, 34, Ohio, using 3 weeks" beats "★★★★★ great product."

5. Objection handler. FAQ block that addresses the 3 objections from your brief. Direct, honest, not defensive. If sizing is an issue, address sizing. If price is the issue, justify price with unit economics.

6. Final CTA. Recap the offer + urgency + button. Not a hard sell — a reminder of what's on offer, then out.

Optional 7th block: comparison table (vs. alternatives) or ingredient/spec deep-dive. Only if it moves objections.

The complete prompt template

Here's the section-by-section prompt we use. Fill in the brief, then run each section prompt sequentially:

[Section 1 — Hero]

OFFER BRIEF:
[paste 8-field brief]

BRAND VOICE:
[paste voice doc]

Write the HERO section for the landing page.
- 1 headline (max 12 words, benefit-forward)
- 1 subhead (max 20 words, specifies WHO it's for and WHAT changes)
- 1 primary CTA button copy (2-4 words, verb-first)
- 1 line under the CTA (guarantee or offer summary, 6-10 words)

Output as HTML with h1, p, and button. No wrapper divs.

Repeat with variants for sections 2-6. Each section gets its own prompt with the same brief + voice, plus a section-specific structure spec.

Full copy-paste template lives in the Branva prompt library we share with clients.

Where the output goes

Three deployment paths:

1. GemPages / PageFly. Paste the HTML into a Text/HTML block. Style with the page builder's inspector. This is what most brands use.

2. Native Shopify section. For brands not on a page builder, drop the HTML into a custom Liquid section. More work but faster page load + no monthly page builder cost.

3. A dedicated landing page template. For high-traffic evergreen offers, create page.landing.json in your theme with the sections. Reuse for multiple offers by changing the metadata block.

We usually deploy on GemPages for iteration speed, then convert to native sections once the page is proven. GemPages adds ~200-500ms to LCP; for hot-selling pages that matters.

Iteration — how we tune before shipping

Two rounds of iteration before the page goes live:

Round 1 — voice + specificity. Read the whole page cold. Does any sentence sound like it could be on 100 other brands' pages? Rewrite for specificity. Add one detail per section that's uniquely yours (customer name, product weight, ingredient origin, delivery time).

Round 2 — trust + friction. Where does someone doubt this page? Is the guarantee clear? Are the reviews specific? Is the button copy conveying value or just "buy now"? Read as an ad-cold visitor, not as the founder.

Then ship, drive 100-200 clicks, and check the heatmap. Iterate on the section where they drop off. Most landing pages we ship take 2-3 revisions before hitting their conversion ceiling.

Talk to Branva

Landing pages are one of our highest-leverage workflows for brands running paid. If you're spending $5K+/month on Meta or Google and sending it to the product page, book a working session and we'll audit your funnel + show what a Claude-built landing page looks like on your top offer.

Frequently Asked Questions

How is this different from just using GemPages templates?

GemPages templates give you a layout; you still need copy. Claude produces the copy structured to the layout. Best used together — GemPages for layout + reusable design, Claude for offer-specific copy.

Do I need separate landing pages per ad?

For high-spend ads: yes. For everything else: one landing page per hero offer, refreshed monthly, is enough. Diminishing returns past 3-4 landing pages per offer.

What's the biggest CRO lift I'll see?

Depends on baseline. Brands sending paid to product pages typically see 20-60% conversion lift moving to a properly-structured landing page. Brands already on a landing page but with generic copy see 10-25% lifts from Claude-tuned copy specifically.

Should I use a page builder or a native section?

Start with a page builder (GemPages, PageFly) for iteration speed. Convert your top-performing evergreen page to a native Shopify section once it's proven. Native sections are 200-500ms faster and don't cost a monthly subscription.

Related reading