PORTFOLIO

People who've worked with me use words like relentless and unstoppable. I've stopped arguing with them. I'm intense. I also love what I do, and I don't think those two things are separate.

Here's where I am. I run go-to-market engineering at BlastPoint, a customer-intelligence company for utilities, banks and credit unions. I lead a small team that builds the machinery behind our sales motion: the data products that start conversations, the agents that research every new lead, the routines that tell reps what changed this week. I studied marketing. I'm not an engineer. I still ship code every week, because I build with Claude Code and check everything it makes.

That's the part of the world I care most about right now. The distance between having an idea and shipping it has collapsed. A marketer with a clear question and a free afternoon can build something that used to take a team and a quarter. Most people haven't noticed yet. The ones who have are quietly running circles around everyone else.

So I try to work like the gap is already gone. I prototype first and ask questions second. I treat every campaign as an experiment with one variable. I check my own numbers before anyone else has to, and I'd rather report a small true number than a big flattering one. Then I write it all down, because a system that only works when I'm in the room isn't a system.

The intensity comes from caring about the people on the other side of the work: the rep who gets an hour back, the buyer who finally gets an email worth reading, the founder at a bootcamp who leaves with a plan. That's the job. I love it.

Most cold email fails for a simple reason: it's about the sender. My product, my case study, my fifteen minutes. Enterprise buyers delete it before the second line.

The fix I've found is a better reason to talk, and the best reasons are sitting in public data.

Start with the data. If your buyers are regulated, rated, ranked or reviewed, there's a public record about them, and there's more of it than you think. Then match it to the right person: not the account, the human who owns that number and has to explain it to a board. Then customize it to what they care about, not what you care about.

We call the result a Provoke. A data trigger plus the right persona equals one reaction: "Wait, how do you know that?" A line like "your energy savings per customer dropped 11% while bills went up 6%" gets a reply from the VP who has to explain it, way more often than a feature list ever has.

Every lead lands in one of three buckets. Carrot: you're winning, here's how to cash in. Stick: you're behind, here's the number that proves it. Insight: huh, that's weird, want to know why?

The best version turns the data into something worth opening on its own. At BlastPoint I built PressurePoint, a public ranking of 1,199 U.S. utilities on affordability pressure. It has opened 40+ executive-level meetings, and the same finding that earns a meeting is a story a journalist might want.

You don't need engineers for any of this. Ask Claude what public data exists about your buyers and what one artifact you could build from it. Then keep asking. Go build something weird and cool that works.

I've run more than 60 growth experiments, and the rule that matters most is the boring one: change one thing.

If you change the subject line, the list and the offer in the same week, you'll learn nothing, no matter which way the numbers move. Change one variable, measure it the same way every time, keep what wins, and write down what you learned so the next person doesn't have to learn it again.

A few other rules I hold myself to.

Short wins. A cold email should fit on a phone screen: under 75 words, one ask, and able to stand on its own. If it needs a follow-up to make sense, it wasn't finished.

AI writes from formulas, not vibes. I don't let a model invent an opening line. I give it the data point and the shape of the sentence. Back in 2023 that took personalized lines from 15 an hour by hand to 1,000 an hour, and personalized lines lifted conversion by 33%.

Opens aren't engagement. Bots open emails and click links. Count replies, meetings and revenue.

Let the data overrule you. I've thrown out my own plans when the numbers said they were wrong. That's the whole point of measuring.

Before a big outbound campaign for a client's ad team, I checked the list. It was supposed to be 353 large accounts, sized by a data vendor's company-size field.

I compared that field against real headcount. Of the 346 companies I could match, 270 were wrong. 204 had fewer than 501 employees. One account listed at 10,001+ had a single employee.

Then I made my own mistake. My first fix leaned on global headcounts, which made local branches of big companies look huge. I caught it, restored the real local employers by hand, and the final list was 134 accounts that actually fit.

So now every field is a claim to me, and claims get checked before anyone spends money on them, including mine.