Insights

Working notes. Published occasionally.

One or two pieces a month, sometimes fewer. I write when I have something worth saying — not on a content calendar. If a topic needs 2,000 words, it gets 2,000. If it needs 400, it gets 400.

More recent notes.

Why your AI content doesn't sound like you.

Most AI content fails not because the model is bad, but because nobody built a voice guide before hitting generate. Here's the fix, and why it matters more than anyone's telling you.

Read

The three things that separate an agent from a prompt.

Autonomy, a defined output, and integration into your stack. If a tool only has one or two of those, it's not an agent — and the confusion is costing marketing teams real money.

Read

Small and senior beats large and junior.

A five-person marketing team with senior people usually outperforms a twenty-person team with juniors and managers. Here's why, and the team structures that actually work at different company sizes.

Read

What the platform-side years taught me about SEO.

Working inside Surfer and MarketMuse showed me the patterns that only become visible across thousands of companies. Three of those patterns, and why your agency probably can't see them.

Read

Kill the agents that aren't pulling their weight.

The quickest way to ruin an AI agent rollout is to never retire the ones that stop earning their keep. A framework for quarterly agent reviews, and why "would anyone notice if we switched this off?" is the only question that matters.

Read

Content that decays drags everything down with it.

Unrefreshed content doesn't just fail to rank — it actively lowers the authority of the content around it. An argument for quarterly content maintenance, and what to do with pieces that aren't worth saving.

Read
Browse by topic

The things I write about.

Four subjects mostly. Occasional drift into adjacent territory when something genuinely interesting shows up, but the core rotation is here.

Next step

Reading the notes is one thing. Talking is another.

If something here lines up with what you're trying to solve, the next step is a conversation. One call, thirty minutes, no pitch deck.