Starting Again

A senior in the industry once told me he closed his business because of the stress from overhead, not the work. I nodded at the time. I am only starting to understand what he meant.

It’s been three years (unofficially) since I started Underscore, one year with full-time hires. People ask me why I started it. The reason was simple at the time: I wanted to design landing pages the way I believed they should be done. Nice things. Beyond ‘good enough’.

The gold rush

The second year is where the pressure showed up. Not because of the work, but because of the constant need to keep revenue ahead of cost. Once you carry overhead, your calendar changes. Your attention changes. You start making decisions for the business model, not the work.

When the AEO and AI search wave hit, I chased it. I set aside what we’d built and pivoted the entire agency into marketing. Everyone was in it. As the phrase ‘gold rush’ suggests, that’s exactly what it was. I gave it my best shot. The landscape turned out to be different from the one I expected.

I have seen how project-by-project quoting turns into a price comparison, where decisions stop being about what’s right and start being about what’s affordable. It is such a waste. What became clear is that I’m happiest inside a product team. I like long arcs. I like shipping, learning, and iterating. I like building conviction with engineers and PMs.

The return

Underscore will be winding down and I am detaching my identity from it. Underscore the business and me the operator are two different things. I am learning to hold them separately. It’s a strange place to sit.

The future of design will be shaped by designers with real foundations and real AI fluency. That’s the intersection where I think I can contribute most.

So in May 2026 I’m looking forward to joining a team again. Bringing design, better business sense, strategic thinking, and AI search experience with me.

I want to ship and build a better system for designers to grow whether through ops or as a coach. The work is better when the system is better.

If you’re building something where design matters beyond ‘good enough’, I’d like to hear about it.