What Changed
Nothing, because we haven't audited this page before. The June 30 portfolio-state coverage flagged the Red Cross and Thermo Fisher teasers as carrying "stronger mandate and metric framing" relative to the rest of the portfolio, then moved on. This is the first full pass against the six-dimension framework, grounded in live HTML fetched July 3, 2026.
The page is mostly working. Browser title is mandate-forward. The hero metadata column is clean and scope-dense. Eighteen animated demo cards deliver more craft evidence than most candidates' entire sites. Five of six dimensions pass.
The one that fails is the one that matters most at scan speed.
Your H1 leads with humanitarian mission framing. It is the single largest text element on the page, and it tells a CPO scanning at speed that this is an empathy story. Everything beneath it, the metadata, the metrics, the platform architecture, then reads as supporting detail for that empathy story. The GM-scope leadership mandate is buried.
Independence Day means low traffic through Monday. H2 headcount unlocks at most companies mid-July. The Tuesday outreach window after the holiday is the cleanest one you'll get this month. Make sure the page is ready for it.
1. Page Title
Current: "National Disaster Relief Platform · 0→1 Build — Juno Chen"
Verdict: Pass. (Confidence: high)
National scale, 0→1 build, your name. No humanitarian framing competes. The Open Graph and Twitter card titles match, so the title holds across LinkedIn shares and Slack previews. A hiring manager seeing this in a browser tab reads mandate before anything else.
Minor cleanup: the OG canonical URL still points to the .html extension while the server 308-redirects to the extensionless path. Not a positioning issue. Worth fixing if you're in the HTML anyway so link previews don't split.
2. First Screen
Current: The hero uses a two-column layout filling the viewport. Left column carries an eyebrow (American Red Cross · Disaster Relief Platform · 2021), the H1, a paragraph, and the first CTA. Right column carries six metadata rows: client, platform, role, team, scale, outcome.
The right column is doing exactly what it should. Role: Product Design Director / GM, 0→1 build. Scale: 6 months, national deployment, 6 legacy systems replaced. Outcome: $847K disbursed, 1,689 cases in 2 weeks. If a buyer's eye lands there first, the page converts.
But the eye doesn't land there first. It lands on the H1.
Current H1:
"Designing for the worst day of someone's life"
Verdict: Fail. Highest-priority fix on the page. (Confidence: high)
The H1 anchors every piece of information beneath it. Scanning at the Director+ evaluation layer works through the largest text element on the page: it sets the interpretive frame, and everything that follows is processed through that frame. Right now the frame is empathy. The paragraph beneath the H1 opens with displaced families before transitioning to six disconnected tools, federal timing pressure, and your leadership scope. By the time the mandate appears, the buyer is already inside a mission story with operational details attached. The leadership mandate has lost its framing power.
Replacement H1:
"Six systems. Three roles. One platform. Deployed nationally in six months."
Four claims: consolidation scope, multi-role architecture, output, timeline. Each one is a mandate signal. Together they tell a buyer what kind of problem you solved before the paragraph adds any context.
Revised hero paragraph (first two sentences):
"Led design and delivery of a national disaster case-management platform for the American Red Cross, replacing six disconnected legacy tools with a single system serving field volunteers, caseworkers, and finance officers. Shipped under federal oversight in a single disaster-response cycle, with 1,689 cases opened in the first two weeks."
Then bring in the human stakes. Displaced families, untrained volunteers, surge conditions. Those details are powerful and they belong in this paragraph. They add weight to the mandate. They just can't be the first thing a scanning eye processes, because a CPO reading the biggest words on the page will anchor on whatever those words establish.
3. Outcome Proof
Current: Three metrics in the hero metadata ($847K disbursed, 6→1 system reduction, 1,689 cases in two weeks). The closing section restates the architecture and adds the 10× surge-volume claim.
Verdict: Pass, with one reframe. (Confidence: moderate)
The metrics are present and correctly located. The concern is how "$847K disbursed" reads in isolation. A CPO seeing that figure reads program reporting. The question in their head: did the platform hold under real operational load on day one? The dollar figure can answer that question, but only if the framing makes it answer that question.
Reframe the outcome row to:
"$847K disbursed in first deployment · 6 systems → 1 · 1,689 cases / 2 weeks · 10× surge-ready"
"In first deployment" turns the dollar figure from program reporting into launch-readiness proof. Pulling "10× surge-ready" out of the closing section and into the hero puts the most impressive scalability claim where it does the most work. The closing section can still elaborate on what surge-readiness required architecturally.
4. Craft Signal
Current: Six workflow sections, each with three animated gallery demo cards. Eighteen UI demos total, driven by demos-engine.js. The demos show case IDs, family records, payment amounts, fraud alerts, FEMA compliance percentages, approval queues, SLA risk indicators, audit logging.
Verdict: Pass. Strongest dimension on the page.
The animated demos prove platform-level systems thinking without requiring the viewer to open the full case study. Intake forms, case dashboards, verification flows, payment disbursement, fraud detection, approval queues, command-level event dashboards. A GM-scope platform with role-specific views across the entire operational chain.
The six workflow sections map to distinct user roles (field volunteer, caseworker, finance officer, supervisor, program director), which reinforces the multi-role architecture claim through the demos themselves. The UI speaks for itself.
One structural note: the six section titles are implemented as styled div elements rather than semantic headings. No positioning impact on a human reader, but search engines and accessibility tools don't register them as structural content. If you're touching the HTML for the H1 fix, wrapping them in h2 tags costs nothing.
5. Density
Current: Hero + six workflow sections + closing proof section + two CTAs + footer navigation. For a page labeled "teaser," this is substantial.
Verdict: Pass.
The density works because the page is gated. The two CTAs ("Request the full case study") make clear this is a preview. The eighteen demo cards give a hiring manager enough to form a strong impression of scope and craft quality, then create a natural reason to request access. That request is a conversion event. A thinner teaser would make the access request feel like a tax. This one makes it feel like the obvious next step.
6. CTA
Current: Two CTAs, both reading "Request the full case study," both linking to cs-03-red-cross.html. One in the hero, one in the closing section.
Verdict: Pass.
Placement is correct. A buyer convinced at the hero level can act immediately. A buyer who scrolls through all six workflow sections hits the second CTA at the natural decision point.
The one thing I can't verify from the HTML alone: whether cs-03-red-cross.html actually gates access or loads the full case study without friction. (Confidence: speculative.) The full case-study pages rely on client-side localStorage checks rather than server-side access control, which means the HTML is delivered before any gate activates. If the gate isn't functional, the "Request" language overpromises exclusivity. Verify before Tuesday.
Summary
| Dimension | Verdict | Action |
|---|---|---|
| Page title | Pass | Fix OG canonical URL if touching HTML |
| First screen | Fail | Rewrite H1 and hero paragraph this weekend |
| Outcome proof | Pass | Reframe outcome row; add "in first deployment" and "10× surge-ready" |
| Craft signal | Pass | Wrap six section-title divs in semantic h2 tags |
| Density | Pass | — |
| CTA | Pass | Verify gate on cs-03-red-cross.html |
One fix changes how this page reads at scan speed. The metadata, the craft evidence, the gating, the density calibration are all holding. The H1 is the single element that determines whether a buyer processes everything beneath it as a leadership mandate or a mission story.
The constraint-dissolution pattern running through this work, FEMA compliance designed into the transaction flow so it never becomes a review gate, surge capacity built into the architecture so it never requires retraining, is exactly the capability that companies entering their complexity phase are hiring for. The demos already demonstrate this pattern. The H1 just needs to tell the buyer what they're about to see before they see it.
Do it before Tuesday.
- Multi-role architecture maps broadly: The one-record, three-role-view pattern visible in the Red Cross demos is the same structural pattern that appears in Headway's Provider Experience posting, where scheduling, clinical documentation, billing, and insurance verification serve distinct workflow needs from shared data.
- Ramp's agentic operations language: Ramp's Director of Product Design posting describes designers building tools, agents, and memory, and adjacent postings describe human-operator-plus-AI-agent workflows that map directly to the Red Cross platform's role-specific views over shared operational data.
- Capital One tests org transformation: The Director, Product Design role in AI in Experience Design names AI tooling, process evolution, and talent upskilling across the entire design org, which means the Red Cross case should lead with operating-model judgment rather than platform craft.
- Client-side gating remains a trust-surface risk: The full case-study pages still deliver HTML before a client-side
localStoragecheck activates any access control, which means the "Request the full case study" CTA promises a gate that doesn't exist at the server level.