22 Years Building. One Lesson That Actually Matters.
I helped build a company that sold for $102.5M. I pushed a contractor to quit freelancing and go build his idea — it became Zapier. After 171 projects, only one lesson held up.
I've shipped 171 projects over 22 years. Products, platforms, apps, APIs, agencies, startups. Some became category leaders. One sold for $102.5 million. Some quietly died. Most fell somewhere in between.
After all of that, if I had to distill everything into one lesson, it's this:
Speed of learning beats quality of execution.
Every time.
The myth of the perfect launch
Early in my career I believed the goal was to get it right before shipping. Perfect the design. Nail the copy. Stress-test the architecture. Launch when it's ready.
At The Penny Hoarder I learned how wrong that was. I joined as employee number three. We didn't have the luxury of perfection — we had a founder with a vision, a small team, and a need to move. So we moved. We built the technology team from scratch, unified marketing and engineering through automations and SOPs, and shipped constantly.
The company hit 100 million monthly page views. Then it sold for $102.5 million — one of the largest exits in digital media history. Not because we launched perfectly. Because we launched, measured, and iterated faster than anyone else in the space.
What 1.18 billion page views taught me
At The Mind Unleashed, I watched monthly revenue go from $9,000 to $435,000 and a social following grow from 300,000 to 7 million. At Discover Hawaii Tours, I took monthly revenue from $700K to $1.2M. At my own agency, The Pixel Flow, we became the number one WP Engine agency partner two years running.
The products that drove that scale weren't the ones we planned the hardest. They were the ones we got in front of users fastest and let real feedback shape.
You can't simulate a real user. No amount of research, testing, or internal review replicates what happens when someone with no context and real expectations tries to use what you built.
The Zapier story
Here's my favorite illustration of this lesson. A developer contracting for The Pixel Flow kept talking about an automation tool he wanted to build. He was good at his job, comfortable, and hesitant to take the leap.
I told him to stop contracting and go build the thing.
He did. The tool became Zapier — now valued at over $5 billion. I didn't found Zapier. But I pushed the person who did. And the reason it worked wasn't that the idea was unique — it's that he shipped it, put it in front of people, and iterated relentlessly while everyone else was still planning.
The agentic parallel
This is one reason I'm building so aggressively in the agentic space right now. AI agents embody this lesson at the infrastructure level.
At CmdCenter, I built an orchestration engine where agents don't plan endlessly — they execute a stage, pass output to the next agent, get checked by a QA agent, and fix what fails. The system has an agent memory layer that records what works and what doesn't, so every run is informed by every previous run. It's the feedback loop made literal.
At SiteClaw, the agent doesn't ask you twenty questions before building. It reads the brief, makes decisions, builds the site, renders the preview, and asks for feedback on something real — not a wireframe, not a mockup, a working site.
Build fast. Learn faster. Ship again.
That's the lesson. It was true in 2002 when I built my first production site. It's true now with agents doing the building. The only thing that changed is the speed at which you can run the loop.
