Think Like a Product Manager
Before you write a single prompt, answer two questions that separate a focused build from a scattered one.
🔁 Recap: The Loop, One Day Later
Yesterday you learned the vibe coding loop — Describe, Build, Test, Refine, Repeat — and used it to build your boarding-pass page. Today that loop doesn’t change. What changes is what happens before you type your first prompt.
Warm-Up: Prompt Challenge
Quick round — in 60 seconds, rewrite this vague prompt using yesterday’s template so it’s specific enough to actually build:
“Make an app about airplanes.”
Compare with your neighbor. Whoever’s version has the most specific “who it’s for” and “styled like” wins bragging rights.
❓ The Two Questions That Matter Most
Every strong build starts before Gemini ever sees a prompt. Ask yourself two questions:
Who is this for? Not “everyone” — a real person or type of person. A new classmate. A nervous first-time flyer. Yourself, six months from now.
What’s the ONE thing it has to do well? Not five things. One. If your quiz app does one thing brilliantly — matches someone to an aircraft — and nothing else, that’s a stronger build than an app that half-does five different things.
Everything past those two answers — colors, animations, extra features — is optional polish. Nice to have, not required.
🔬 Why These Two Questions Actually Work
These aren’t just study habits — they’re backed by how both builders and brains actually work.
“Who is this for?” Once you’ve stared at your own build for twenty minutes, you can’t see it fresh anymore — every button makes sense to you because you put it there. Researchers call this the curse of knowledge, and it’s the single biggest reason builders misjudge their own apps. Naming one specific person — even a made-up one — forces you to imagine someone who doesn’t already know what you know. It’s also the same idea behind “Jobs to Be Done,” a framework real product teams use: people don’t just want an app, they’re “hiring” it to do a specific job for them.
“What’s the ONE thing it must do well?” Partly, this is about you. Your working memory holds roughly seven things at once — that’s measured, not a discipline problem. Every extra feature you try to build at the same time competes for that same slim budget, which is exactly why yesterday’s kitchen-sink pitfall produces messier code. Partly, it’s about your user: one of the most copied rules in software history, the Unix philosophy, is one sentence — “do one thing and do it well.” Software that tries to do five things adequately loses to software that does one thing brilliantly, every time.
💡 Why This Changes Your Prompt
Compare these two prompts for the same rough idea:
| Without the PM lens | With the PM lens |
|---|---|
| ”Build a quiz about airplanes" | "Build a 5-question quiz that matches a first-time flyer to one of four aircraft personality types, based on how they’d want to travel” |
| Vague target, vague outcome | Specific audience, specific single job |
The second prompt is longer, but it’s not more complicated — it’s more focused. Gemini has a clear target to build toward, and you have a clear target to test against.
“Everyone” is not an audience. If you try to build something for everyone, you’ll end up compromising on all of it — too simple for the people who wanted depth, too complex for the people who wanted speed. Pick one real person in your head (even a made-up one) and build for them specifically. It’s easier to widen a focused build later than to sharpen a vague one.
You want to build a 'flight planning helper.' Using today's lens, what should you decide before you write your first prompt?
Reveal Answer & Explanation
Who specifically it’s for, and the one job it must do well — before anything else.
“Flight planning helper” could mean a dozen different apps: a checklist for student pilots, a route-distance calculator for hobbyists, a packing list for frequent flyers. None of those are wrong — but trying to build all of them at once produces something that does none of them well. Pick one person and one job, and the rest of your build gets dramatically easier.
⏭️ Up Next
Next: three simple UX habits — clarity, feedback, and hierarchy — that make whatever you build today easier to actually use.