What Comes Next Is a Choice

I’ve spent my career learning how to improve experiences. Now I’m applying the same feedback loop to what I want to do next.

For most of my career, I’ve worked on one question.

What makes someone convert?

I don’t mean what makes someone click a button. I mean what changes an experience enough for someone to take the next step that matters to them and to the business.

Eventually, another question became just as important. Once you learn something, how do you make sure the next decision starts there instead of starting over?

That’s experience optimization to me. It’s a feedback loop. Every action creates evidence, and that evidence should change the next action.

I start with where we need to end up, work backwards, then test something small and learn from it. Crawl, walk, run. Then do it again.

Where This Started

• 2000: Experimentation became the job. I was the first web designer at eDiets. By 2001 we were A/B testing with a tool we built ourselves, and by 2002 a profile quiz was changing what people saw. I ran experiments there for 11 years.
• 2011: I learned the whole system by doing it myself. AlexDesigns became my full-time consulting work: research, design, build, testing and analysis, much of it for ecommerce businesses. I learned as much from losing tests as winning ones.
• 2015: I learned how to scale it. I moved into a large digital consultancy and helped companies build optimization programs across bigger teams and more complicated organizations, where experimentation, personalization and data all had to work together.

The scope kept changing. The feedback loop didn’t.

Working inside other organizations taught me two things. It isn’t your business. You’re there to help it succeed on its terms, even when that means supporting a direction you wouldn’t have picked. And you don’t know what you don’t know until someone with different expertise shows you.

Some of the work I wouldn’t have chosen, in marketing operations, enterprise technology and connected data, taught me the most.

What the Last Six Months Changed

Across all of that, one problem kept coming back. Research, tests, data and results lived in different places. When a project ended, much of what it taught us was hard to find again.

I never thought of that as a people problem. It was a connection problem.

I’ve spent a lot of that time building with AI. I expected it to make me faster, and it has. What I didn’t expect was how much it would change the shape of the work.

With connected tools like MCP, I can start with a question and pull analytics, research, past tests and customer data around it instead of opening each system one at a time. I still check the sources.

By the twentieth experiment, a team shouldn’t have to rediscover what the first nineteen already taught it.

I can also investigate an unfamiliar platform, prototype an approach and find the likely constraints before I bring in a developer. Work that once took weeks can sometimes be explored in days. Small teams still need specialists, but they can get much farther before they need twenty people.

Easier to Build, Harder to Get Right

I recently built a personalized demo on Optimizely that generated a one-to-one landing page for each visitor.

The first working version came together fast. That was the easy part.

The hard part was making it personal. If ten different people go through it and get roughly the same answer, it isn’t personalization. It’s generic, no matter how polished it looks. Most of the work went into the questions underneath. Which signals actually matter? What should change for each person, and what should stay the same? How would we know it worked?

That’s the shift. Building gets cheaper. Judgment gets more valuable.

AI makes it easier to do good work faster. It also makes it easier to do bad work faster.

The work I find most interesting now is deciding where AI belongs, what it should trust, where a person has to decide, and how to keep one weak output from quietly becoming the input for the next ten decisions.

Asking the Same Questions About Myself

Somewhere in the middle of all this, I realized I couldn’t keep helping companies question their assumptions while leaving my own alone.

I’d already sensed change was coming, even if I didn’t know what form it would take. I’d spent more than ten years in one chapter, and it taught me more than I expected. Continuing to grow meant putting myself somewhere that challenged me differently.

The team was a big reason I stayed invested as long as I did. More than once we took on something nobody was sure would come together, and it did because we worked through it together. I wanted to give that team everything I could, and they made me better. I wouldn’t trade it.

So I’m running the same loop on myself, aimed at the next ten years.

Five Things I’m Reminding Myself to Do

Not Starting Over

I’m at my best when I can bring people, data, technology and experimentation together around a real problem and help a team learn its way toward a better answer. I don’t have to know everything to do that. I have to get the right expertise into the room and keep everyone pointed at the customer and the result.

Some of the work I didn’t expect to matter is what prepared me for what comes next. Every losing test made the next hypothesis sharper. More than twenty-five years of that doesn’t reset because the tools changed.

This isn’t starting over.

It’s compounding what I’ve learned.

---