📉 I Had a LinkedIn Blind Spot I Didn't Know About

Hi friends 👋

My LinkedIn reach has been lower than usual lately.

And before I started blaming the algorithm (which, to be honest, is our first instinct), I thought it's time to do a proper audit.

Well, an algorithm wants people to stay on the platform, so it pushes the posts that make people spend time on the platform.

We've all heard that.

But I've had posts with 10–15% engagement getting only 200 impressions, while other posts with 4% engagement got thousands. And nothing's more frustrating than that kind of inconsistency when you don't know why it's happening.

Spoiler alert! This analysis surfaced the blind spots. Specifically, it identified the types of posts, and the patterns that were hurting not only those posts, but also the ones I posted after them.

Here's some context.

For almost a year now, I've been tracking every single LinkedIn post I publish in a spreadsheet.

The date, the type of post, impressions, reach, likes, comments, saves, reposts — everything. And I have a formula that calculates the engagement rate for each one automatically.

It's something I recommend to every founder or creator I work with, and I do that for my own posts as well.

So I entered over 70 posts in the spreadsheet.

But here's the thing: collecting data is one thing. Actually sitting down to analyze it properly and spot the real patterns is another thing entirely.

I’d been meaning to do a deep dive for a while and this week, I finally did. And I didn't do it alone.

I added my latest posts to the file, shared it with Claude Code, and asked it to analyze everything.

Oh my. Let me tell you, it did an excellent job.

It pulled out what had been consistently working, what wasn't, and (the most valuable part) the blind spots. The things I hadn't been seeing myself, even though the data had been sitting right there.

That last part is the real value of having an outside eye on your own content, even if that eye is AI.

It helped me readjust my strategy going forward: double down on what's worked, and drop the habits that have been apparently working against me.

If you want to do this for your own content, here's the exact process I went through:

Step 1. Start tracking (if you aren't already)

Copy my spreadsheet template and start logging every post you publish.

👉 Copy the template

At minimum, track: day of the week, post type, hook (first line), impressions, members reached, reactions, comments, and saves.

The magic isn't in any single post's numbers, it's in the patterns that emerge across 20, 30, 50 posts.

Step 2. Once you have data, use this prompt with Claude Code

When your spreadsheet has a decent amount of posts (I'd say 20+ to start seeing real patterns), share it with Claude Code and use this prompt:

Step 3. A note on context

If Claude doesn't know much about you yet, give it some background before you run the analysis. Tell it who you are, who your audience is, what services you offer, and what you're trying to achieve with your content.

The more context it has, the more specific and useful the feedback will be. Generic input gets generic output. That's true for AI just as much as it's true for content.

You may ask, but Kate, why is it necessary to use Claude Code? Would the regular chat feature not give the same results?

You could use it, but the results aren’t the same.

Actually, first I tried the regular chat, and it’s not great at reading the files, especially if the file is a bit complex. Claude Code, on the other hand, is great at doing that. And it doesn’t weave in its own research or opinions like the chat feature does. It’s more analytical and bases its outputs strictly on what's in your data.

So yeah, you can try it, but just know that the results differ wildly. At least, it did for me.

That being said, the whole process took me less than an hour and gave me more clarity than getting a manual overview of my last two week’s content.

If you've been posting consistently but feel like something's off, or that an algorithm doesn’t love you, this is worth doing.

Hit reply if you try this, I'd genuinely love to hear what you find.

If you’re interested in starting a newsletter like this, try out beehiiv (that’s what I use).

See you next week,
Kate 🌟