Design Is the New Edge for Solo Founders (and I'm Bad at It)
Design Is the New Edge for Solo Founders (and I'm Bad at It)
The thing I have to admit, as a solo dev who has been shipping products for fifteen years, is that I'm bad at design. I've always been bad at design. I picked engineering specifically because design felt like the part of the work where I had no instincts and the feedback loop was too slow to develop them. For most of those fifteen years, this was fine. The product won on what it did, not on what it looked like.
That stopped being fine sometime in the last twelve months. AI tools have erased the speed gap that used to be a solo dev's main advantage. Anyone with Cursor and a weekend can ship a working SaaS now. The products that stand out in 2026 are the ones that don't look like everyone else's weekend project. Which means the people who skip the design step — me, and probably you — are going to lose to the people who don't.
This is the post I've been writing in my head while reluctantly opening Figma every weekend. What changed, what I've tried, what works, and what to do if you're also bad at design.
The thesis: when shipping is free, looking like a template is a death sentence
The pitch deck version of this argument is that design has always been important. The honest version is that design used to be a tiebreaker between two products that did similar things, and now it's the entire decision because both products almost always do the same thing.
Here's the math. In 2022, building a working SaaS from scratch took a small team three months. In 2026, the same SaaS takes a solo dev a week. The supply of "products that exist and roughly work" has gone up by roughly an order of magnitude. The supply of attention to discover them has not. So the filter has shifted from "does it work" to "is it worth my time to find out if it works."
The signal a potential customer uses for that filter, in the first three seconds, is design. Not the product, not the marketing copy, not the testimonials — the visual quality of the page they land on. If your landing page looks like a default Tailwind starter, you're in the bucket of "another weekend project" and you don't get the click. If it looks like someone made deliberate choices, you get the chance to make the case.
That's the bar now. Not "good design." Just "deliberate design that doesn't read as default."
What I tried this month
I tried three things, with mixed results.
Claude Design (the model-side design tooling Anthropic shipped) for the early exploration phase. This worked well as a thinking partner — describe what you want, get a few directions, iterate on the language. Where it fell down for me was the actual visual output; the generated mockups were fine but generic, and the moment I tried to push them toward something specific I was back to driving the visual decisions myself. So the model is useful as a brainstorming surface, less useful as the thing that produces the final design.
Figma for the parts that actually shipped. I'm slow in Figma. I've used it on and off for years and I'm still slow. But it's the right tool when I know what I want and need to get there cleanly. The honest read: I should accept that Figma will take me three times as long as it would take a designer, budget that time, and stop trying to skip it.
Lovable for the prototype phase. This was the most surprising data point. Lovable produces visually competent prototypes faster than I can in Figma, and the resulting code is reasonable to take into a real codebase. It is not a replacement for design taste — the output is still a "Lovable-shaped product" if you don't push it — but it is a useful productivity multiplier for solo devs who are bad at design and need a starting point that doesn't look terrible.
The honest combined verdict: Claude Design for thinking, Lovable for prototyping, Figma for the parts that ship. None of these are a substitute for taste, but they collectively close some of the gap between a non-designer and an actually-shipped product.
The three things that kill solo-built products on sight
You can take this as a checklist. If your product has any of these on the landing page, fix them this week.
Default Tailwind colors with no customization. The slate, gray, blue palette that ships with Tailwind is everywhere now, and people pattern-match it as "didn't bother." A custom palette is a fifteen-minute fix and an enormous trust signal. You don't need to be a colorist; pick a primary, pick a neutral, vary the saturation, and you're already ahead of 80% of solo-built sites.
The generic gradient hero. You know the one. Diagonal pink-to-purple, big bold heading, three feature cards underneath, a CTA button. This was a fresh look in 2021. In 2026 it reads as "Stripe-influenced template I didn't customize." Replace it with literally anything specific to your product — a real screenshot, a real animation, a real demo, even just a real photograph. The generic gradient is the visual equivalent of the LLM-tell.
"Built with [framework]" badges. They were a community signal in the early framework days. Today they're a signal that you're more interested in the stack than the product. Take them off your landing page. Put them in your README if you want, but the customer doesn't care what you built it with.
A practical, non-designer's checklist
This is what I've assembled into a pre-launch checklist for myself, with no designer on the team.
Pick one custom color. Just one. Use it deliberately. Build the rest of the palette around neutrals.
Pick one custom font. The default sans on every framework is one of three fonts. Pick something else. There are thousands of free options that look great. This costs nothing and signals deliberateness.
Use real content in screenshots. Every screenshot on your landing page should show real data, real text, real interactions. Lorem ipsum and "John Doe" example users tell the customer you don't have a real product yet.
One CTA per page. Solo-built products tend to throw three CTAs onto a hero "in case." Pick the one that matters most and remove the others.
Sit on the design for a day before shipping. Most of the bad design choices I've shipped were ones I would have caught if I'd waited 24 hours and looked again with fresh eyes. The cost of the day is nothing; the cost of the bad design is real.
The honest take
Design is now the differentiator that engineering used to be, and the people who treat it as optional are going to keep losing to the people who treat it as table stakes. I've been one of those people. I'm trying to stop.
The good news is that the bar isn't actually "good design" yet — it's just "design that doesn't read as default." That's a bar a non-designer can clear with deliberate effort and twenty hours of practice. The bad news is that the bar is going to keep rising as more solo devs figure this out, and the people who start now will have a year of compound improvement on the people who start later.
Start now. Pick a custom color this week. The rest follows from there.