Case Study: How I Built a Mobile Experience From Scratch in Less Than 4 Days (Part-Time)

I’m not a super strong front end engineer.

Even so, I built, tested, and shipped a pretty good mobile onboarding flow and home screen for a "desktop-first" app over a few calendar days, part-time.

No figma. Just my experience, taste, Claude code, and some timely code review and internal feedback.

In previous roles, work like this would have taken weeks at best. Maybe a quarter plus at worst.

The Context

My client Rootnote is a user-friendly analytics product that helps content creators track and sell the value of their online presence. From day one, it was designed for larger screens. For many key use cases, that still makes sense.

But this year, market feedback pushed us to pivot. Instead of selling primarily to universities and talent agencies, we're now going direct to athletes and creators, supported by channel partnerships.

The implication was clear: our mobile experience couldn't just exist, it had to hook a mobile-native audience and activate them at high rates. These users won't show up for demos or webinars. They need to get value immediately, on their phones.

Every day without that experience meant shorter runway and less room to test the bet behind the pivot.

Re-engaging contract design help was not really in the budget and would add weeks to the timeline.

The desktop view of the rootnote content page

The Build

So, the back and forth with Claude began.

Armed with decent responsive design knowledge and deep exposure to end users, I could make informed priority decisions. Within minutes, I was iterating through design variations that used to require dedicated design time and handoffs.

The onboarding flow went through 5-10 iterations based on my taste and internal feedback, from first prompt to production.

Most of the calendar time went to code review and testing. Moving from browser developer tools to real devices exposed new issues, but Claude handled the feedback like a champ. It delivered solid frontend code iterations in minutes.

Of course, I then had to refine the desktop styling and test everything again.

A refined selector in the onboarding flow for the key platforms that sync content level analytics

What Made This Possible

The magic wasn't just in the AI generated code. It was in the combination of that and:

• Deep user understanding and judgement built from years of experience
• Clear prioritization driven by a common sense of urgency
• A bias for rapid iteration without heavy processes and hand offs
• Exploratory testing with actual devices and user mindset

Yes, expert design help would have made it better—probably even faster too. But it wasn't in the cards, and perfect wasn't the goal. Good enough to measure, learn, and iterate was.

New guided sync experience with intro video

Zooming out

There's still lots of room for improvement. More screens need attention. The end-to-end testing suite needs a revamp. But we now have a foundation that's already gathering mostly positive, unsolicited user feedback.

Start to production in less than 4 days. Part-time.

I'm sure many could have done this better or faster. But not in most environments I've experienced over my 20-year career.

In 2025, I could write a case study like this almost every other week. For over a year now, I've felt like I have superpowers. And they're only getting stronger. With great power comes great responsibility, and judgment matters more than ever.

But man, being able to make a meaningful difference for a client and their users this quickly? Producing results in days instead of quarters?

This is my favorite stretch yet of my career as a builder.

What are you shipping in less than a week that would have been impossible a year or two ago?

New mobile content home screen