I Kept Bouncing Between Design and Code. Here's Where I Landed.

For years, I’ve had two tabs open in my head.

In one, I’m a developer. I write code, I ship things, I debug at midnight and feel great about it. In the other, I’m a designer, or maybe a product person: the one who asks who this is for, what problem we’re solving, and whether we’re building the right thing at all.

I kept switching between them. Backend developer one month, UX designer the next, then cybersecurity, then back to front-end. Every time I picked one, the other one pulled me back.

This week, I finally said it out loud: I’m a product person who can build, not a builder who can design. That one sentence untangled a lot.

Why I kept bouncing

Here’s the honest part. I never doubted that I could build. I have the repos, the degree, the late nights. Code gives you proof: it runs or it doesn’t.

Design doesn’t work like that. There’s no compiler for “is this good?” So every time design got hard, the voice in my head said maybe you’re not good enough for this, and I went back to the thing I could prove.

What I’m learning is that “good enough to design” isn’t a talent you’re born with. Most of design is problem-solving: who is this for, what are they trying to do, where do they get stuck? The polish comes from practice and feedback. And without feedback, the only critic you hear is the one in your head, which is usually harsher and less accurate than any real person would be.

The funny part? When I was bored, I’d open Figma and design something just for fun, the same way I’d open my editor and code something just for fun. I was doing both all along. I just only gave myself credit for one.

I’ve been doing discovery the whole time

Then something clicked that I’d never connected before.

Before my current job, I worked in business retention at a telecom company. My job was to talk to business customers who wanted to cancel and figure out how to keep them. And the way I did that was by asking questions. Who, what, when, where, why. How many users do you have? How many devices? What’s not working? Why are you thinking about leaving?

I never thought of that as product work. But that’s discovery: understanding the user and their context before you recommend anything. It worked, too. I averaged an 80%+ save rate, and one day I saved 98% of the customers I talked to.

Looking back, almost every job I’ve had put me right next to the user: answering member questions in insurance, talking businesses out of canceling, and now running operations where I see exactly where workflows and tools break down. I spent years thinking that was “not tech experience.” It turns out it’s the part a lot of product people would love to have.

Putting it into practice

So I decided to stop talking about design and actually test some.

Helix & Hue Studio is a concept site I designed and built for a curly and coily hair stylist in Charlotte. Its whole job is to answer the first question a new client asks: does this stylist do hair like mine? My case study said I’d written a usability test plan. Which was true. I just hadn’t run it.

Getting it ready taught me more than I expected:

• The site wasn’t actually testable yet. Prices said [PRICE]. The FAQ said [Prep instructions]. Half my tasks couldn’t be finished. I filled in realistic sample content so people could complete every task.
• I found broken buttons I’d never noticed. “See the work” jumped to the top of the page instead of the gallery. The bottom “Book a visit” did the same thing, which would have looked broken in the middle of a test. Small things, but they’re exactly what trips up a real person.
• Empty boxes change how people judge everything. The gallery was full of gray placeholder text. That’s fine in a wireframe, but in a test it tells people the site isn’t finished, and they get more critical of everything else. So I added photos.
• I did a practice run first. I ran the whole test with an AI agent playing a participant, so I’d know what to expect. It was useful for learning the flow, and it surfaced one good question: if I like a style in the gallery, which service do I book to get it? But it also showed me what a simulated user can’t do. It never actually tapped anything. Real people will, and that’s the point.

I also made myself a simpler note sheet: one row per task, a check, an H for “needed help,” or an X, plus a few words. Because the detailed version was never going to happen mid-session. (Hi, ADHD.)

“Never in my life has my hair looked like that”

The studio is fictional, so there’s no real portfolio of work to show. I used AI-generated placeholder photos, and the site says so in the footer. Then I checked every photo against the curl type it was filed under.

And I’m glad I did. The generator kept pushing everything toward the same glossy, defined ringlets. Looser 3A curls came out too tight. 4B and 4C came out too defined and too shiny.

Then I looked at the 4A photo. I have 4A hair. My first reaction was: never in my life has my hair looked like that.

That reaction is exactly what this site is supposed to prevent. Someone with 4A hair taps the 4A filter, sees hair that looks nothing like theirs, and gets the same message Instagram already gives them: this stylist’s work isn’t for hair like mine. So I regenerated it until it looked like a real day-two twist-out on hair like mine.

The lesson I’m taking with me: photo accuracy isn’t decoration on this site. It’s the core of the design. And it has to be checked by people who actually have that hair type, which is why I’m testing with them.

What’s next

Over the next couple of weeks, I’m running the Helix & Hue usability test with five real people with curly and coily hair. I’ll write up what they actually did, where they got stuck, and what I changed because of it. That’s the part that turns a case study from “here’s what I assumed” into “here’s what I learned.”

I also updated my resume and portfolio to tell one story instead of three: someone who starts with discovery and ends with shipped code.

If you’ve been bouncing between two paths like I was, here’s what helped me: stop asking which one you’re better at, and ask which one you keep coming back to. Then look at your past jobs again. You might have been doing the work longer than you think.

If you have curly or coily hair (anywhere from 3A to 4C) and want to help with the test, send me a message. Twenty minutes, your phone, and my thanks.

Bria