From Dreamweaver in 2003 to AI Experimentation Today

It was around 2003 when I built my first real website, for an insurance agency. At the time I was all in on Adobe Dreamweaver. If you could build a site in Dreamweaver, you felt like you were ahead of the curve. Looking back, that's kind of funny. Now I can build AI-powered websites in hours, not weeks, definitely not months. But that first site? In my mind, it was legit. It was also graphic-heavy and probably slow as hell. Didn't matter. **The first signal that hooked me** My goal was simple from day one. Get free traffic and turn it into leads. I managed to rank top 5 for "midwest insurance" and it started producing leads. A handful a day. That was it. That was the moment I realized this wasn't just building websites. This was building demand. **Topical authority before it had a name** What I didn't realize at the time was that I was already thinking in topical themes. I wasn't just building one page and hoping it ranked. I started expanding: - Different insurance types - Regional variations - Supporting content around the main keyword I didn't call it topical authority back then. I just knew the more relevant content I built around a theme, the more Google rewarded the site. No frameworks. No courses. Just build, rank, expand, repeat. That same loop still runs today, and it's why most of my [content frameworks for AI search visibility](/blog/seo-visibility-ai-powered-search) start with topic clusters before they ever touch a single page. The mechanics changed. The logic didn't. **Falling in love with user experience** Around that same time I started paying attention to something most people overlooked. What happens after the click? Ranking wasn't enough. If users landed on the site and couldn't figure out what to do next, the traffic didn't matter. That's when I came across [Don't Make Me Think](https://sensible.com/dont-make-me-think/) by Steve Krug. One idea stuck immediately. "Don't make me think." Simple. Obvious. Easy to ignore. But it changed how I built everything. - Navigation had to be clear - Pages had to be scannable - Calls to action had to be obvious No guessing. No friction. That mindset was just as important as ranking. Traffic gets you in the door. UX is what turns it into results. I still apply Krug's first chapter to every funnel I build. It's also baked into how I think about the [funnel breakdown calculator](/tools) on this site, every step has to be obvious or the whole thing leaks. **Testing before testing platforms existed** Same thing with optimization. We were A/B testing, we just didn't call it that. There was no [Optimizely](https://www.optimizely.com/), no plug-and-play tools, no dashboards. If I wanted to test something I built two versions of a page, swapped them manually, and watched what converted better. Headlines, layouts, forms, button placement. Everything was a test. And honestly? It forced you to actually understand what was driving behavior, not just what some dashboard told you. I'm still skeptical of teams who run tests without being able to explain why something won. The tool gives you the data. Understanding still has to come from you. **The shift that accelerated everything** About five years in, I moved into a role where I helped build a consumer lending website and affiliate program for what was, in its prime, a small $450 million lending company. This is where things leveled up. And yes, I designed the front end in Dreamweaver. Hand-rolled HTML, CSS that I was very proud of, and the occasional table-based layout I will deny under oath. Float-clearing divs were my love language. The site shipped. The leads came in. Nobody complained that it wasn't React. We weren't just ranking. We were: - Dominating consumer loan keywords - Supporting affiliates with banners and tracking code (before UTMs were standard) - Building reporting in [Google Analytics](https://marketingplatform.google.com/about/analytics/), which felt like science fiction. Just two years earlier I was hooked on [AWStats](https://www.awstats.org/), basically dressed-up server logs with bar charts. Then GA showed up. I've been hooked ever since. I started working directly with affiliates, eventually managing a small group and even charging for support. That's where I really learned the next lesson: traffic is one thing, scaling it across systems is another. It's also where I first ran into the analytics tradeoffs I later wrote about in [why I chose Google over Adobe Enterprise Analytics](/blog/why-i-chose-google-over-adobe-enterprise-analytics). Different tools, same question, can the people on my team actually use this thing. **Full circle, same mindset, new tools** Fast forward to today and the mindset hasn't really changed. I still launch my own sites. I still test everything. I still chase organic traffic. The difference is speed and scale. Now I'm leaning heavily into AI. Not just to build content faster, but to recreate something I used to do manually, my own version of an experimentation engine. Back then Dreamweaver felt like the edge. You could drag, drop, hand-code when you needed to, and ship a real website. That was the bar. Today the bar is automation. Instead of designing one page at a time, I'm designing content automation tools that span the whole stack: website content, social media, a podcast system, even video. That's the thinking behind [ContentWeaver](/ai-projects/marketing), which I designed and developed end to end, and honestly it's the biggest shift I've made in years. The easiest way to describe it? Think of it as the luxury version of [Semrush](https://www.semrush.com/), [Ahrefs](https://ahrefs.com/), [Sprout Social](https://sproutsocial.com/), a video and podcast studio, and a link outreach tool, all stitched into one workflow. Not five tabs, five logins, five invoices. One place. Same itch I had in 2003. Bigger surface area. Instead of swapping pages by hand, I'm building systems that: - Generate variations - Test messaging across channels - Learn from performance data - Iterate automatically A lot of this overlaps with where I think things are heading next, which I laid out in [digital marketing priorities for 2026](/blog/digital-marketing-priorities-2026). Less manual labor. More closed-loop systems. More humans doing the strategy and judgment, less doing the busywork. **The funny part** Back then I thought mastering Dreamweaver was the future. Turns out the real skill wasn't the tool. It was understanding intent, building around topics, designing frictionless experiences, testing what works, and adapting faster than everyone else. That part hasn't changed at all. If you're working on something similar, building organic demand, layering in AI, trying to figure out what actually converts, [let's talk](/contact). I'd rather trade notes than write another generic playbook.