Juno, nothing on this page has changed since the July 19 audit. The page was redeployed August 7, but every gap flagged then — the H1, the row order, the Open Graph metadata, the missing first-screen artifact — remains open.
What was already right before that audit is still right, and naming it calibrates what "done" looks like when you edit your other teasers. The mandate paragraph claims agency ("I named it, built the research case, secured the mandate, and redesigned it"), anchors scale at $50B+ GMV, surfaces your exact title, and delivers four quantified outcomes. The two-column layout lets a reader scan either the paragraph or the project brief and extract the claim from either path. The browser title carries the mandate frame. The metadata leads with the role. These decisions landed by the July 9 cycle and they're correct. Stop touching them.
Now the gaps.
1. Title language
The H1 reads: "The signals before anyone touched a screen."
Flagged in the July 9 audit. Flagged again July 19. Still unchanged.
The H1 is the largest text on the page. A hiring manager's eye lands there first, and at eight seconds it may be the only text they process at full attention. Their working theory about your candidacy forms from that line. Right now it communicates a research-first sensibility — real, and relevant — but it doesn't communicate what you did, what you owned, or what moved. The mandate paragraph does all of that. The H1 doesn't.
Your mandate paragraph already contains the replacement. Pull the claim up:
Named the trust gap. Secured the mandate. Redesigned the platform at $50B+ GMV.
Every clause is a verb you performed. The scale anchors the final clause. The word "trust" tells the reader what kind of problem this was. If you want to preserve the research-first signal, the current H1 works as a subhead beneath this line. It cannot be the first thing the eye reads.
2. First-screen information hierarchy
This is the strongest part of the page. At a 1440×800 viewport, the first screen delivers all three elements without scrolling:
- (a) Mandate: The paragraph names the gap, claims the agency, specifies the redesign.
- (b) Scale: $50B+ GMV, 200K+ suppliers, 25M+ desktop sessions, plus the "25% of traffic, 80% of transaction value" compression that tells the reader why desktop mattered.
- (c) Outcomes with numbers: +7% DAU, +20% transactions, +2.2pt NPS, −47% security concerns.
One structural adjustment. The Outcome row sits last in the project brief. A hiring manager scanning the right column top-to-bottom hits Company and Project before reaching the numbers. The company name is already in the eyebrow ("Alibaba.com · B2B Platform Redesign · 2024"); repeating it doesn't add information. Lead with what the reader can't get from the eyebrow.
Proposed row order:
- Role — Head of Design & Research, North America
- Outcome — +7% DAU · +20% daily transactions · +2.2pt NPS · −47% security concerns
- Scale — $50B+ GMV · 200K+ suppliers · 25M+ desktop sessions
- Team — Design · Research · PM · Engineering
- Project — B2B Site Refresh · 2024
Drop the Company row entirely.
A metrics detail: the rendered Outcome row says "+20% transactions" while your own metadata description says "+20% daily transactions." The qualifier matters. "Daily transactions" is more specific and more credible to a hiring manager who reads enterprise metrics closely. Make them match. Keep "daily."
The Open Graph description also omits the −47% security-concern result. When someone shares this link on Slack or LinkedIn — which is how most hiring managers first encounter a portfolio page — the preview text drops your most concrete outcome. Add it.
3. Craft signal
The first screen contains no visual design artifact — no screenshot, annotated UI, or interaction detail. What's visible is typography, prose, the project brief, and a decorative "01."
The sprint sections below the fold have strong before/after screen captures — three homepage states, three search states, three product-detail states. None of that evidence enters the viewport before the reader scrolls. A hiring manager evaluating for a design leadership role sees a strategy document where they expected a designer's portfolio. The work exists. It's below the fold.
You don't need to redesign the hero. Promote one artifact into it.
The strongest candidate is a before/after pair from Sprint 01 — the sign-in gap. A single side-by-side (legacy homepage on the left, authenticated state on the right) placed below the mandate paragraph and above the CTA would prove this is a design case and give the reader a reason to scroll. The "after" state raises the question of what changed and why, which is exactly the question you want them carrying into the full case.
If a full before/after pair crowds the hero at this viewport, a single annotated screen from the authenticated state with two or three callout labels — "trust signal," "procurement context," "authenticated navigation" — would also work. At Director level, the hiring manager is deciding whether you can diagnose a complex system and show what you changed. An annotated screen proves judgment. A polished mockup proves execution. Judgment is the scarcer signal at this seniority, and that applies to every teaser you build.
4. Call to action
The CTA reads: "REQUEST THE FULL CASE STUDY" in uppercase, dark red, with a right arrow.
"Request" implies a gate. The reader has to ask for access rather than view the work. Even if the full case is password-protected, the word creates friction before the reader decides whether the content is worth the friction. Use "View" or "Open."
The second problem is that "the full case study" is generic. The reader doesn't know whether they'll get a deeper strategy narrative, more screens, the research methodology, or all three.
Proposed replacement:
View the full case — research framework, three sprint redesigns, and component system →
This tells the hiring manager what depth awaits and gives them three reasons to click. If the full case requires a password, add a single line beneath in smaller type: Password-protected. Request access via the link. The primary CTA reads as an invitation; the access note handles logistics separately.
Verdict
Fix the H1 to lead with the mandate. Promote one before/after artifact into the hero above the fold. Those two changes are the difference between a page that rewards careful reading and one that earns the click at scan speed.
- OpenAI Payments maps here: The Product Designer, Payments posting asks candidates to make complex financial logic trustworthy at scale — Alibaba's transactional trust evidence is the most direct public proof you have for that claim.
- Figma's recruiter data confirms: Their portfolio guidance says reviewers may spend only minutes or seconds on an initial scan and recommends showing evidence most pertinent to the target role rather than comprehensive history.
- The access-control gap persists: OWASP's client-side authorization guidance classifies browser-only gating as presentation rather than access control, which means the full case behind this teaser's CTA is effectively public whether you intend it to be or not.
- Gusto wants this vocabulary: Their Head of Design posting asks candidates to define trust, uncertainty, and graceful failure in AI-generated outputs — the same diagnostic language your Alibaba mandate paragraph already uses about a different system.