Dear Junior Product Designer, Let’s Ship

A real-talk letter on designing with clarity, not chaos

This is the note I wish someone handed me on day one. No buzzwords. Just field notes with one quick story per point so it’s advice, not a portfolio tour.

1. Conference Room A. The agenda isn’t the work.

Truth: Projects are messy. Process is a toolbox, not a temple.

First Monday at LootCrate. The whiteboard had a glorious double diamond, color coded, arrows everywhere. By Wednesday our promos changed, shipping rules flipped, and that diamond looked like a map of a city that no longer existed. We scrapped ceremony for a 3-day loop: map the funnel, paper prototype, tiny live test. We shipped Friday, learned Monday, and repeated. Conversion climbed because our cadence matched reality.

Pocket lesson: Treat process like utensils in a kitchen. Reach for the spatula you need, not the whole drawer. Write the goal at the top of your file and pick the smallest move that reduces risk this week.

2. You don’t need to test everything. Research is not a buffet.

Truth: Testing is triage. Test what changes the decision.

At Zillow, our conversational assistant had choices everywhere. Voice tone, turn taking, escalation rules. We did not “test everything.” We made a Risk Board on a sticky note: decision, risk if wrong, cheapest way to learn. Tone got five hallway tests. Escalation rules earned a short Wizard-of-Oz with support. Different risk, different bite.

Try this: Triage before testing. Make a Risk Board. For each decision, note risk if wrong and the cheapest way to learn. Do the top three. Park the rest. If a study won’t change a decision, it’s a snack you don’t need.

3. Visual Design: The Quiet Boss Of The Room.

Truth: Type, spacing, and contrast direct attention and trust.

SAP had complex tools with many modules. We standardized spacing tokens and a clear type scale. No fanfare. Suddenly screens felt learnable. People understood where to look first, second, third. Support questions dropped because hierarchy did the talking.

Try this: Do the 90-second flash test. Show the screen, ask “What is this?” and “What happens next?” If answers wobble, fix type and space before features.

4. Tools: Put Down The Swiss Army Chainsaw

Truth: You need a sharp knife and good taste, not every plug-in.

I watched designers collect plug-ins like souvenirs. The folks who shipped kept a sharp knife. At Zillow, a humble Figma flow plus a 3-minute Loom moved a cross-team decision faster than a pixel-perfect playground.

Pocket lesson: Go deep on one design tool, one prototyping method, one storytelling format. Add more only when the problem truly needs it. Save the team from death by new tools.

5. Leadership Speaks Metrics, Not Vibes (Stakeholders do get UX—translate it)

Truth: People approve risk, time, and money—not adjectives.

On an SAP rollout, “we need two days of usability testing” sounded like delay. We reframed it as “avoid three sprints of rework on the grid.” The spreadsheet smiled, budget appeared, and nobody asked for a 40-slide deck.

Pocket lesson: Give every design activity a business subtitle. Lowers abandonment. Reduces support load. Avoids compliance risk. Increases throughput. Talk business.

6. AI: Power Tools, Not Your Replacement.

Truth: AI accelerates work. Judgment still decides.

At Zillow, we used AI to sift thousands of chat transcripts, cluster intents, and draft reply scaffolds for a mortgage pre-qual assistant. Helpful, but the real design work was ours: setting guardrails, escalation rules, and disclosure language, then choosing tone per step. AI sped up drafts; human judgment set the choreography and accountability.

Try this: Use AI to rough in flows, cluster insights, and generate edge cases. Keep humans on framing, ethics, accessibility and accountability. Example of prompts to use:

• 1) Idea starter, not idea replacer.
  • Use it for: rough concepts, flows, empty-state ideas, microcopy directions.
    Prompt: “Give me 5 alternative flows for a subscription checkout with promo codes and gift options. Return as steps, each step one sentence. Flag the riskiest step for drop-off.”
• 2) Research turbo: from noise to signal
  • Use it for: clustering notes, drafting insights, writing discussion guides.
    Prompt: “Cluster these interview notes into 5 themes. For each theme: user quote, root need, risk if we ignore it, metric it affects.”
• 3) Microcopy that moves metrics
  • Use it for: headlines, CTAs, error states, status messages.
    Prompt: “Write 7 CTA variants to reduce anxiety for mortgage pre-qual. Constraints: ≤22 chars, verbs first, plain language, no jargon. Tone: calm, capable.”
• 5) Edge cases and red-team
  • Use it for: breaking your happy path before users do.
    Prompt: “Act as a malicious user. Try 20 ways to break this password reset flow. Return each as: step, expected outcome, failure we create, fix we should make.”

7. The Open Office. Be visible without becoming a full-time billboard

Truth: Being visible matters. Performing visibility burns you out.

On a cross-team SAP program, we killed status theater with a weekly one-pager: outcome metric, top risks, decisions needed, next ship date. Leadership followed the work; meetings didn’t multiply. We broke the silos and fostered a psychological safety space and shared ownership.

Pocket lesson: Share small, steady signals. A one-pager or a 90-second Loom. Show before and after, tied to the metric. Decline updates that don’t move outcomes.

7. The Helpdesk Bell: The 15-Minute Rule (Ask early, learn faster)

Truth: Stuck happens. Move as a team. After 15 minutes, ask for help and learn faster.

At Zillow, I hit a wall on mortgage eligibility logic. After 15 focused minutes I wrote what I tried, the exact blocker, and two options, then pinged PM and engineering. Ten minutes later we were unblocked and I had learned a pattern I used all quarter.

Try this: When stuck, set a 15-minute timer. Document attempts, hypothesis, and next options. Ask with context. Close the loop by posting the fix so the next person moves faster

Pocket Plays for Your Next Sprint

Outcomes live in the hallways, not the rooms. You move between process and research, visuals and tools, finance and power tools, visibility and help. Your job is to lower risk, raise confidence, and make the next step obvious for users and for your team.

• Outcome first: pin one metric to your canvas.
• Risk Board: decisions, risk, cheapest learning.
• Hierarchy before features: make the first glance do the heavy lifting.
• Tool diet: one tool, one proto, one story.
• Speak KPI, not vibes.
• AI as co-pilot: draft, critique, choose.
• Visibility by signal, not stage.
• Ask early with the 15-minute rule.