PAGE 00MANIFESTOQ3 2026 · 2 SLOTS LEFT
00manifesto

Sangmin Lee, Search Visibility Engineering. JSON-LD architecture, structured-data engineering, and SEO + AEO readiness for brands losing organic revenue and AI citation share. Engineer-led delivery. Same engineer diagnoses and ships the fix. Available Q3 2026.

headline 01 of 03 · rotating

JSON-LD architecture and AI search readiness for brands losing organic revenue. Engineer-led: diagnose, ship, monitor. Same engineer. Both engines.

Book the diagnosis callSee what's broken
Q3 2026 · 2 OF 4 SLOTS BOOKED
01services

Diagnose. Implement. Sustain.

Three engagements. Same engineer. No handoffs.

type · 01

Consulting. Analysis only. No code ships.

type · 02

Execution. Production code. Real change.

type · 03

Managed. Continuous on-call engineering.

01 · diagnose02 · implement03 · sustain3 phases
consulting
01diagnose

Audit

Diagnostic only. No code ships. You walk away with revenue mapped per fix, not a deck.

$25,000one-time · 2-3 weeks
// shipsper-fix revenue projection + 90-day prioritized roadmap. Foundation diagnosed for SEO + AEO in one pass.
includes
  • Schema graph audit · entity relationships
  • AI visibility scan across 5 engines
  • Crawlability + Core Web Vitals review
  • Per-fix projected revenue impact
  • 90-day prioritized remediation roadmap
best for · scoping the problem before committing engineering
Book the audit call
// start hereexecution
02implement

Sprint

Production code ships. Schema architecture, JSON-LD per template, validator-clean structured data, with CI gates and pre/post measurement.

$70,000+fixed scope · 4-6 weeks
// ships@graph in production, validator suite in CI, citation tracking pre/post, engineer pair sessions. Both engines covered.
includes
  • Full @graph architecture across templates
  • Per-template JSON-LD implementation in production
  • Entity resolution + AI engine emulation
  • Schema validator suite integrated into CI
  • Engineer pair sessions for your team
  • Pre/post AI engine citation measurement
best for · post-audit remediation, ready to ship
Book the sprint call
managed
03sustain

Retainer

On-call engineering. Schema decays. I keep it from decaying. Both engines stay clean.

standard$10,000/mo
enterprise$25,000/mo
monthly · 3-month minimum
// shipsregression monitoring (CI), citation tracking (5 engines), monthly pair sessions, Slack 24h SLA.
foundation includes
  • Schema regression monitoring · CI-integrated
  • AI engine citation tracking · 5 engines
  • Monthly engineering pair sessions
  • Slack channel · 24-hour response SLA
  • Quarterly architecture reviews
enterprise adds
  • Dedicated weekly cadence (60 min)
  • 4-hour incident response SLA (vs 24h)
  • RankLabs dashboard with custom alerts
  • Multi-domain / international entity graph
best for · post-sprint. Schema work compounds when maintained.
Book the retainer call// protects sprint investment
alternative path
// QUICK SCAN

5 days. Single AI engine. Written report. $4,500 credits 100% to Audit within 30 days.

$4,500
02method

Most search-visibility budgets are shopping for content while the foundation rusts.

Schema is engineering. The people selling it usually don't write code.

Schema goes stale. Reviews ship but aggregateRating never wires up. CMS updates break BreadcrumbList. No helper, no validator, no owner. Rich-result CTR walks over to your competitor.

Same fix on the AI side. Solid @id and sameAs anchors give answer engines one source to ground against. Drift narrows. Citations stabilize.

the foundation · entity graphSchema entity graph for Sangmin LeePerson Sangmin Lee at the center, connected to Service entities (Audit, Sprint, Retainer), Article (Writing), and Organization (RankLabs).@type: ServiceAudit@type: ServiceSprint@type: ServiceRetainer@type: PersonSangmin Lee@type: ArticleWriting@type: OrgRankLabsentity relationships · what AI engines extract from your site

I write the JSON-LD. I write the helper. I open the PR. Your team reviews and merges. No agency, no handoffs, no deck.

03tooling

I built RankLabs because the tool I needed didn't exist.

Existing crawlers validate JSON-LD against Google's published spec, but validity isn't rich-result eligibility, and answer-engine extraction wasn't what they were built for. RankLabs runs against your site, extracts every entity, and tells you what's missing on which template. The terminal below is a real run on a demo target.

ranklabs-crawlerv1.0 · live
[14:23:01]// ranklabs-crawler v1.0 · started
[14:23:01]target: https://example-commerce.com
[14:23:02]crawling / · extracted 4 JSON-LD entities
[14:23:02]Person · Organization · WebSite · BreadcrumbList
[14:23:03]crawling /products/*
[14:23:04]247 product pages found
[14:23:04]192/247 missing Offer.priceCurrency
[14:23:05]crawling /blog/* · 38 articles
[14:23:06]14 articles missing @type: Article
[14:23:08]// crawl complete ·13,263 pages · 47 issues · 9.4s
pages per engagement250K+
schema.org types60+
issue patterns200+
AI engines tracked5
03proof

Past work that taught me what's broken.

Selected engagements. Anonymized where the contract requires.

commerce · 2026 · ACTIVE · LIVE IN BRANCH PREVIEWanonymized

Mid-market commerce · single-domain @graph consolidation across 45K+ PDPs

Migrated fragmented per-template JSON-LD into a single resolvable @graph. Target: restoring Product rich results in Google Shopping, recovering Featured Snippet wins on long-tail product queries, and resolving entity coverage in AI engines as the secondary bonus. PR live in branch preview today; measuring rich-result CTR + AI citation share through Q3.

JSON-LD@graphSchema.org/ProductRankLabs
experimentation · 2021–present · via SAE.SOanonymized

Fortune-1000 health & wellness brand · A/B testing infrastructure

300+ client-side experiments delivered (and counting). Engineered DOM-diffing and MutationObserver strategies to inject experiments into dynamic WordPress/Next.js apps without breaking core experiences. Five years of production CRO/MVT alongside PMs and analysts. The experimentation muscle stays current.

MutationObserverWordPressNext.jsOptimizely
founding · 2025–present

RankLabs · AI visibility infrastructure

Built proprietary crawler, structured-data extraction pipeline, page-level scoring, and field-level schema validation, engineered to flag the rich-result-killers and the AI-citation-blockers in the same crawl. Powers diagnostics across every client engagement. I shopped Screaming Frog, Botify, Sitebulb, and Onely's stack before I built.

Custom crawlerSchema validatorAI engine emulation
origin · 2015–2016 · publisher-scale

Complex Networks · multi-site frontend revamp + BI partnership

Led frontend revamp on 3 sister sites single-handedly before they consolidated. Partnered with BI team on cross-site technical SEO and conversion experimentation. High-traffic publisher SEO: schema, internal linking architecture, page speed, crawl-budget, ranking signals. This is where I learned what classic SEO at publisher scale actually looks like, a decade before AI search was a category.

Frontend architectureMulti-siteBI partnership
04writing

Engineering notes.

Public writing on JSON-LD, schema architecture, and what AI search actually retrieves.

  1. 2026-05-02

    How AI answer engines actually resolve your @graph

    ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews each parse JSON-LD differently. The deltas matter. Here's what I've measured running the same @graph through all five.

    read
  2. 2026-05-02

    10 JSON-LD bugs I find in every commerce audit (and what they cost)

    Field-level JSON-LD bugs that fail silently across 45K+ PDP catalogs. Each one mapped to the rich result it kills, the AI engine that drops you, and the dollar cost I model for clients.

    read
  3. 2026-05-02

    Schema decays. Here's the half-life I've measured.

    Production schema goes invisible to AI engines on its own, with no obvious failure. Here's the regression timeline I've measured across client engagements, and why CI gates aren't optional.

    read
View all writing
05contact

Stop pouring budget into a broken foundation.

If your SEO retainer hasn't compounded, your AI citations have stalled, or your last technical audit ended in a deck nobody read, that's not a content problem. It's an engineering problem. The same engineer who diagnoses ships the fix.

Book a 30-min call30-min call · no deck · engineer to engineer
or write me directly
I read every message. Reply within 24h.// legitimate interest (GDPR Art. 6(1)(f)) — you requested the contact. privacy policy