Content architecture
Designed the content model in Contentstack — field structure, entry types, modular components — so editors could publish consistently and engineers could render predictably.
One engagement, in depth
A national DTC brand. Headless Shopify Plus. Contentstack. The work that built the architecture fluency behind every Morelight engagement.
The engagement
A national sleep and mattress brand running a headless Shopify Plus storefront on Contentstack. I joined as content lead and owned content strategy, structured content architecture, and the full template set across a complex catalog — from product detail pages to editorial articles to category landing pages.
Stack
Shopify PlusContentstackHeadless frontendProfoundCrownPeak
What I owned
Template set
Defined heading hierarchy, schema fields, alt-text requirements, and FAQ patterns across three core template types: PDP, editorial article, and category landing page.
Structured data
Implemented Product, FAQPage, BreadcrumbList, and Article schema across the catalog. Ensured entity signals were consistent and machine-readable.
AI visibility monitoring
Built and maintained a prompt set in Profound to track brand citation rate across ChatGPT, Claude, Perplexity, and Google AI Overviews. Reported monthly on citation gaps and what changed.
Accessibility review
Ran CrownPeak scans on the template set. Triaged critical and serious issues, wrote remediation tickets for the engineering team, and tracked closure rate over time.
Governance
Maintained the baseline after each release. When new templates or content pushed issue volume back up, the next cycle brought it back down. The flatline from mid-2026 — new pages not returning the site to peak issue counts — is the product.
What the engineering team shipped
My role was architecture, specs, and tickets. The headless frontend engineers merged the changes. That is the operating model — senior strategy and specification on my side, implementation on theirs.
- Heading model refactored across three template types — from marketing-driven heading soup to question-H2 and answer-paragraph structure
- Product and FAQPage schema added to PDP template, surfacing structured answers in AI and rich results
- Alt text and name-role-value fixes on interactive components — closed the majority of critical CrownPeak findings on the template set
- Answer blocks added to key PDP sections — short, extractable answers to common product questions, structured for AI citation
CrownPeak — 24 months on a 378-page catalog
Total website issue volume — accessibility, SEO, usability, and content quality — cut from roughly 4,000 to under 500 over two years on a live headless Plus catalog. Spikes are releases. Drops are template and HTML work.
CrownPeak total website issues, 24 months
All issue types: accessibility, SEO, usability, content quality. 22 Sep 2024 – 22 Sep 2026. Spikes are releases; drops are template and HTML remediation. A one-off auditor cannot show this arc.
WCAG issues by level, 12 months
Level A
Level AA
Level AAA
Level A (dark), Level AA (blue dashed), Level AAA (pink dashed). 22 Sep 2025 – 22 Sep 2026. Level AA was already well-managed; Level A dropped from ~1,000 to a plateau of 83 open.
~4,000 → ~500
Total website issues, 24 months
Accessibility, SEO, usability, content quality
9.6 / 10
CrownPeak accessibility score
Platform average 6.9
0
Priority issues
Across 378 pages
1.14
Issues per page
378 pages scanned
vs CrownPeak platform averages
Accessibility9.6vs 6.9 avg
SEO9.5vs 6.7 avg
Usability8.6vs 6.8 avg
83 Level A and 42 Level AA issues remain open — a controlled backlog on a live catalog, not a finished audit. AA conformance is a program, not a badge.
The same heading, label, and name-or-role fixes that cleared CrownPeak priority issues are what answer engines extract from PDPs and guides.
Organic and performance results
Metrics from the published Contentstack case study.
30x
organic article traffic growth
<1s
page load on the headless stack
50%+
fewer developer resources required post-migration
This is the architecture I bring to every engagement
If your brand runs a similar stack and needs AI visibility and accessibility handled at this level, the next step is a Snapshot.