What a CPO Sees in 90 Seconds on junochen.com

Current-State Verification

Three issues flagged July 10. Here's where they stand.

1. "Product Designer" title. As of July 19, the direct HTML source has been updated. Browser title and meta description now carry design-leadership language, including Head of Design availability. Partially resolved. The rendered fetch path still returns "Juno Chen — Product Designer" as the visible page title. Cached snapshots show older headline copy beginning with "The stakes change." This is source/render drift, and it matters more than the source fix. When a recruiter pastes your URL into Slack, when Google's index hasn't refreshed, the preview card says Product Designer. Force a cache purge. Resubmit to Google Search Console. Verify in incognito on a device that has never visited your site. Until the rendered output matches the source, the fix is cosmetic.

2. what-do-you-count.html. As of July 19, still linked from the homepage. Still sitting in the Field Studies section, tagged Pricing / Strategy / Apr 2026. The external-use boundary classifies all TinyFish pricing and packaging details as internal-only. The homepage contradicts that boundary every day this card stays live. P0 unchanged. Remove the card today.

3. Full case study pages returning HTML before client-side redirect. As of July 19, unverified at the individual page level. On July 10 the pattern was clear: each case study page (CS 01 through CS 06) delivered its full HTML to the browser before a client-side JavaScript check against localStorage could fire a redirect. Anyone who views source, disables JavaScript, or uses a fetch tool sees the complete case study without authentication. I could not re-verify each page in this pass. Check it yourself in under five minutes: open each case study URL in a fresh incognito window with JavaScript disabled (Chrome DevTools → Settings → Debugger → Disable JavaScript). Full page loads? Gate is still cosmetic. Nothing loads or redirect fires? Fix landed. Until you've confirmed, assume the July 10 pattern persists.

CS 04 (Equinox+) and CS 06 status. Equinox+ teaser title still reads "Equinox+ · Teaser · Juno Chen" with no meta description tag. The hero proof rows are strong: Product Design Director, 600K+ premium members, 200+ clubs, 0→1 in 90 days, 4.8-star App Store rating. But the title and missing meta description mean this page previews generically in any share context. Add a mandate-forward title and a one-sentence meta description that leads with the role and scope. CS 06 was not available for verification.

The Core Question

The guideline for this audit asks whether junochen.com claims "human-agent workflow coherence" as Juno's design territory.

The homepage doesn't use that phrase. What it does, in the agentic section, is ask: "What does a user need to see when an agent acts on their behalf?" That question is the buyer-facing version of workflow coherence. It names the same territory in language a CPO would actually use.

Sharpen it further. Workflow coherence as an abstraction doesn't give a buyer anything to hold onto. The decision moment does. The moment a human decides to trust, override, or act on what the system recommends. That's the territory: concrete, singular, ownable. Every element on the page should point here.

The Diagnosis

The homepage is oscillating between two signals. Some elements read: design leader who has operated at scale, built from zero, and is now working on the trust layer of agentic AI. Other elements read: strong senior practitioner who does impressive AI-adjacent work. Both are present. Both are credible. The inconsistency is what kills you.

A CPO, VP Design, or CDO scanning this page has 90 seconds and ten tabs open. They are pattern-matching against a mental template with two bins: leader and practitioner. There is no middle bin. When the sort is inconsistent, the buyer doesn't average the signals. They default to the lower read. Practitioner. Next tab.

The raw material to sort cleanly into the leader bin is already on the page. One more pass makes every element answer the buyer's question the same way.

Break 1 — The Headline Claims a Domain, Not a Decision

Current headline positions Juno as a design leader for trust-centered platforms. The territory is correct. "Trust-centered platforms" is a category label. It tells the buyer what you work on. A CPO scanning ten portfolios needs to know what you do that changes outcomes, mapped to the problem sitting on their desk this week.

That problem: they are building products where an AI agent acts on behalf of a human, and nobody on their team knows how to design the moment where the human decides whether to trust the agent's output, override it, or let it proceed.

Revised headline:

Tagline (beneath the headline):

Headline names the decision. Tagline names the role. Without the tagline, a buyer reads the headline and thinks "great framing for an elevated IC." The tagline carries the leadership signal the headline deliberately leaves out. One claims the territory. The other claims the title.

This pairing works because the headline applies to every case in the portfolio. Alibaba: a buyer deciding whether to trust a payment flow across regulatory regimes. Thermo Fisher: a supply chain operator deciding whether to adopt a new digital workflow. Red Cross: a responder trusting a consolidated platform during a crisis. Carrier IQ, Brand Pulse, Retail Velocity: a user deciding whether to trust what an agent surfaced. The headline holds across every case on the page.

Break 2 — The Hero Body Introduces Noise

The hero eyebrow is working. Availability, design leadership, AI-native platforms, San Francisco Bay Area. Four signals in one line. Keep it.

Hero body proof points are strong. $50B+ GMV is a scale signal. Zero-to-live in twelve months is a velocity signal. Agentic AI concept-to-production is a currency signal. Three proof points, three buyer anxieties addressed: Can she operate at scale? Can she build from nothing? Is she current?

Then there's the TinyFish line. In a hero section that otherwise reads as design leadership at enterprise scale, a reference to current work at a company the buyer hasn't heard of raises a question the hero should not be raising: Is she a founder? Is this a side project? Is she actually available? The eyebrow already says available. The TinyFish line complicates that signal in the first five seconds. Remove it from the hero. If TinyFish context matters, it belongs in a conversation, not in the opening frame.

2-sentence professional statement for Head of Design targeting:

Names the leadership function. Proves scale and velocity with published numbers. Connects enterprise track record to AI-native work without making the agentic labs sound like a hobby.

Break 3 — Case Cards Lead with Projects, Not Mandates

The homepage frames its case studies under a line about design leadership where wrong calls carry real cost. Good. That framing line tells the buyer these aren't redesign projects.

The cards undercut it. Four of five lead with project descriptions rather than leadership mandates. The difference matters at this level: "Redesigned procurement for 40M buyers" is a project. "Led design for a $50B+ GMV procurement platform" is a mandate. A CPO scanning these cards is asking one question: Did she lead, or did she execute?

CS 01 — Alibaba.com, revised:

Leads with "led design." Scale numbers follow the mandate verb immediately. The tag "agentic payment" does quiet work connecting enterprise case work to AI currency. Keep it.

CS 02 — Thermo Fisher, revised:

Already the strongest card on the page. The revision sharpens it by naming the 12-month timeline and reframing the work as a change-leadership problem. A CPO hires a Head of Design to solve adoption problems, not screen problems.

CS 03 — American Red Cross, revised:

Leads with "led." Six-to-one complexity signal up front. Stakes language is specific and earned. "Shipped to national deployment" is the outcome the current card doesn't deliver.

The pattern across all three: mandate verb first, scale number second, stakes third, outcome fourth. This is the order a buyer processes information. They want to know you led before they want to know what you led.

Break 4 — Agentic Labs Signals Maker, Not Leader

Section framing is excellent. "What does a user need to see when an agent acts on their behalf?" is the exact question every company building agentic products is trying to answer. This is your territory. The question alone earns a second look from any CPO building in this space.

Then the section says "built solo."

To a CPO hiring a Head of Design, "built solo" signals one thing: this person is a maker who doesn't need a team. That is the opposite of what they need to hear. They need: I understand how to build these systems, and I understand how to lead a team that builds them at scale. "Solo" positions the labs as side projects rather than applied research from a design leader investigating the trust layer of agent-driven workflows.

Reframing recommendation:

Drop "solo." Drop the count. Frame this as a design leader who builds to learn, not a practitioner who builds because building is what they do.

The count discrepancy. Direct HTML contains four agent systems in the active carousel: Carrier IQ, UAT Sentinel, Retail Velocity, and Brand Pulse. The rendered snapshot surfaces five systems in an older card-style layout, adding Creator Scout. The external-use boundary defines three as portfolio-ready: Brand Pulse, Retail Velocity, and Carrier IQ. UAT Sentinel and Creator Scout fall outside that boundary.

Reduce the carousel to three. Brand Pulse, Retail Velocity, Carrier IQ. These three cover the range: brand intelligence, field operations, logistics verification. UAT Sentinel is a QA automation tool. Creator Scout's function isn't documented in the published portfolio. Neither has been cleared for external positioning. Displaying work you haven't cleared for external use on the same page where you're positioning yourself as a leader who makes careful decisions about trust and transparency is a contradiction a sharp buyer will catch.

Break 5 — Field Studies, What Earns Its Slot

Three essay cards. Each one evaluated on a single criterion: does it reinforce the decision-moment territory or dilute it?

"Trust Is the New Interface" (AI Design / Agentic UX / Jun 2026). The excerpt opens with an agent failure framed as a design problem. Right move. A CPO clicking this link finds a published position on the exact problem they're hiring for: what happens when an agent fails, and whose job it is to design for that failure. This reads as a position paper from someone with a point of view, not a blog post from someone with a content calendar. Lead with it. It directly supports the decision-moment territory the rest of the site is claiming.

"What Following Taught Me About Leading" (Leadership / Reflection / Sep 2025). A tango metaphor for leadership. Apply the 90-second lens: a CPO in their last 20 seconds on the page reads this title and thinks personal brand content, not leadership signal. It doesn't hurt. It doesn't help. It's the oldest piece on the page by seven months. If you keep it, it's the second card, never the first. But if the choice is between two cards that both reinforce the agentic-trust territory and three cards where one is a personal reflection from last year, two is the stronger number.

"What Do You Count When the Playbook Doesn't Exist?" (Pricing / Strategy / Apr 2026). Remove this card. P0. Links to content that violates your external-use boundary. Addressed above.

The section headline is about what stays with Juno. This is soft. A buyer scanning this section should understand immediately that these are published perspectives on design leadership in AI-native environments. Consider a framing line that connects the essays to the decision-moment territory. These are positions, not musings. Signal that.

This Week

Prioritized by urgency:

1. Remove the What Do You Count card from the homepage. P0. Today.
2. Force-purge the rendered cache. "Product Designer" is still appearing in rendered output. Resubmit to Google Search Console. Test in incognito on a clean device. Until the rendered version matches the source, the title fix doesn't exist for the people who matter.
3. Remove the TinyFish line from the hero body. It competes with the leadership signal everything else is building.
4. Reduce the agentic carousel to three systems. Brand Pulse, Retail Velocity, Carrier IQ. Pull UAT Sentinel and Creator Scout until they're deliberately cleared for external use.
5. Swap the headline and add the tagline. Current headline names the domain. Revised headline names the decision moment. Tagline names the role. Together they answer both questions a buyer asks in the first five seconds: What does she do? and At what level?
6. Rewrite CS 01–03 card copy. Lead with the mandate verb. "Led." "Built." "Shipped." Project description follows the leadership claim.
7. Add a meta description to the Equinox+ teaser page. Proof rows are strong. Metadata is invisible.
8. Verify case study page access control. Five minutes, incognito, JavaScript disabled. Full HTML loads? Gate is still cosmetic.

The homepage is significantly better than it was on July 10. Hero proof points are real, case cards have substance, and the agentic section is asking the right question. What remains is resolving the oscillation. Every element needs to sort into the same bin. Right now some answer "leader" and others answer "impressive practitioner." At the Director+ level, the buyer doesn't split the difference. They sort you into the lower bin and move to the next tab.

You have the evidence. You have the territory. The homepage just needs to stop hedging about whether you're the person who answers the question it's already asking.