I want to walk you through something that happened this week that I think is worth documenting.
We ran a full storefront overhaul on Logoclothz.com using Claude as the primary build agent. Not for one task. For everything. Competitive research, theme code, product catalog updates, schema audits, content cleanup, navigation rebuild. Eight phases over three days.
This is the build log.
Why We Did This
The store had a pricing communication problem. The all-in delivered price at Logoclothz is genuinely competitive when you compare it spec-for-spec against the big players. But the buy box was running a triple price ladder (MSRP, Was, Sale), showing “Shipping: Calculated at Checkout” right next to a free-shipping promise, and never using the words “delivered” or “included.”
We were winning on price and losing on perception.
The strategy that came out of the research phase was simple: fuse the all-in message to the price. Sell speed and honesty, not “cheapest.”
Phase 0: Research First, Always
Before touching a single line of code, Claude ran a ten-competitor teardown. Live pages, pricing structures, real carts. The goal was to understand how the competition communicates price and what signals customers actually respond to.
This is where Fable 5 earns its keep. When you drop a full competitive brief into a long-context model and tell it to find the gaps rather than confirm your assumptions, you get different answers than you would from a standard prompt. Claude found the exact friction point: we were hiding our competitive advantage behind a pricing structure that looked like every other store that was actually more expensive.
That research phase produced two strategy documents before a single code change was made.
The 8 Phases
Phase 1: All-In Pricing. Changed the shipping language site-wide, removed the MSRP ladder from all product displays, and added a clear “All-In Delivered Price” box to every buy box. The message: no setup fees, no art fees, no shipping added at checkout. Ever.
Phase 2: Product Page Trust Cleanup. Replaced about 165 lines of inline-styled HTML with clean reusable components. Consolidated the trust signals, fixed a Quick View bug that was injecting product-page scripts into category popups, and added a four-step “How It Works” strip.
Phase 3: Learn Hub Rebuild. Tagged all 19 blog posts with topic categories via API. Added a topic pill bar and a sizing-series navigator. Fixed 61 broken relative links and removed 13 hand-pasted “related guides” blocks that had been duplicated across posts.
Phase 4: Navigation Overhaul. 19 raw top-level nav items collapsed into 6 curated mega-panel items. Zero URL changes. The mobile menu was left untouched. A “Free 90-Min Proof” CTA and phone number were added to the desktop bar.
Phase 5: Per-Family Product Claims. This one required careful thinking. The store sells products with different fabric types, and each fabric type has different certifications. Spandex is not fire rated. Poplin is. Recycled fabric has its own certifications. We applied a Badge Profile custom field to all 81 products via API, and the theme now renders the correct claims per profile. If a product has no profile set, it shows no fabric claims at all. Fail-safe against over-claiming.
Phase 6: Internal Linking and ADA. Added a Size Family field to all 81 products. Product pages now automatically link to the matching sizing guide, size category, and related products. Nav contrast ratios were pushed to WCAG AAA.
Phase 7: FAQ Page Rebuild. The FAQ comparison page had an empty body. Rebuilt it via the Pages API with real content: eight FAQ accordions, eight comparison cards, four resource cards, and FAQPage JSON-LD schema for AEO.
Phase 8: Schema Audit. Found duplicate unlinked Product entities on every product page - one from the app layer, one from the theme. Fixed by giving both a matching @id so they merge into a single entity in Google’s knowledge graph. Added OfferShippingDetails with the free UPS Ground information, which enables the “Free delivery” annotation in Google Shopping surfaces.
What I Learned About Using Claude for This
A few things stood out.
First, context is everything. Claude did not just get a task list. It got the competitive analysis, the brand rules, the product taxonomy, and the existing code structure before any phase started. The quality of output is directly proportional to the quality of context you provide.
Second, Fable 5’s long context window changes what is possible. You can drop in a full codebase, a competitor teardown, and a strategy brief and ask it to find the gaps. That is a different kind of work than a short-context model can do.
Third, the schema audit phase was the one I would have skipped if I was moving fast. It is not visible. It does not look impressive. But duplicate unlinked Product entities are exactly the kind of thing that quietly prevents your products from getting rich results in Google Shopping. The invisible work matters.
The Open Items
A few things are still on the list: scoping the NFPA claim on a Page Builder widget that is rendering on the wrong product type, checking hero text size on real phones, and resubmitting the sitemap to Search Console.
The branch is ready to merge. The theme is live. The matrix is running.
If you want to understand how to use Claude for real e-commerce work, not just content generation but actual system-level changes, come learn with me at Ready, Plan, Grow!.
Find me across the web
Stay curious, my AI friend. It's the secret sauce - think like you are seven. - Ryan