Why your biggest posts usually aren’t you best ones

I was showing a colleague my Notion system the other day and a few minutes in he made a face and said:

===

"Is it really worth it? For what, 10 subscribers?"

===

That’s fair because he was looking at a column of pretty sad little numbers. One subscriber from this post. Five from another. And a whole lot of zeros.

So I flipped the page and showed him something more interesting.

A post with 2,500 likes that brought in zero email subscribers. And a post with about 250 likes that brought in 180 new followers and 72 email subscribers.

===

"Okay. That's a lot better. I can just imagine, you show people stuff like this, and they’ll be like, ‘where do I sign up?’"

===

But honestly, none of that surprised me.

I’ve been writing on Threads since July 2023, back before you could schedule a post or see much of anything useful about how it performed. I had posts do a million views. Two million. Ten million. And I was tracking it all by hand in Notion.

It was crude too. I’d look which days people subscribed, and assume they came from the post that had the most likes that day.

But even though it was crude, it told me what I needed to know.

The post with the most likes were almost never the ones that helped me grow my email list.

Which is why I stopped worrying about likes and views a couple years ago. I’m more interested when content brings people onto my email list or into my DMs because that means I’m finally creating something valuable enough for people to respond to.

Now that same crude tracking method can be automated with the power of AI.

It’s no longer “probably that post”, it’s “definitely, this post brought in exactly 72 new email subscribers.”

But tracking might be the smallest part of the system at this point.

Every post I've ever written now lives in Notion and in Claude Code, with metrics sitting right next to it. Which means any agent you point at it can read the whole thing. Analyze it. Find the patterns. And surface what's actually working, so you can make more of the stuff that sounds like you, instead of what someone else thinks you should sound like to get more likes.

You can ask what your best performing content was last quarter and get a real answer instead of a guess.

You can dig up posts from two years ago that people used to love and turn one of them into three new ones. You can build writing skills that draft in your own voice, from your own ideas, using your own back catalog as the source material.

And then you schedule it all right from Notion. Ask AI to plan your finished content out for you over the next week, and it will be published automatically at the scheduled date and time and your metrics will be synced back for you while you go do other things.

No copying and pasting between apps. And no writing a page-long prompt so an MCP server can interpret it differently every single time.

Two tools. Claude Code on your desktop, and a Notion account. That's all you need.

Which is what I'd like to teach in a live, three day workshop.

By the end of it you’ll have something that looks like this:

https://www.skool.com/kevin/classroom/c6cd1e91?md=8b4687eb5fef4d589365eb75cbb8c2a2

If you want in, just reply back with "I'm interested" and I'll send you the details.

Talk soon,

Kevin Hood