The clearest winner in this period is a post I almost didn't write: "I became a Belgian citizen in April." 5,593 impressions. 138 likes. 17 comments. 4 follows from a single post. By every metric I track, it is the strongest post in six months of data (leaving aside a cringe out-of-character post that I will talk about in a bit.)

The post with the best comment-to-impression ratio (roughly 1%, which is the highest conversion rate in my dataset) is "Signs a LinkedIn post was written by AI." An open list format.

The best performing contrarion post: "AI made me dumb." 3,324 impressions, 20 likes, 18 comments. High reach and conversation, without relying on a life event to carry it.

And then there's the post I need to name directly because ignoring it would be dishonest: a low-effort, slightly ragey post about Sam Altman's take on AI energy consumption. It got the highest impressions of any post I have ever done. I don't think that post deserved it. It was ragebait. It attracted AI naysayers (the opposite of my target audience.) And the LinkedIn algorithm rewarded it anyway. That's the number that most directly challenges my current strategy assumptions.

This edition of the newsletter is interrupted to tell you about a tool I looove on LinkedIn. It is called LinkedHelper and has been a lifesaver when it comes to expanding my network and also outreach on LinkedIn. Aaand you can get a 10% discount (after your 2-week free trial) if you use my promolink - Click here.

The three performance drivers I identified in January (personal story + image, reflection tone, and a photo of my face) have all held. But they've gotten more complicated.

Let me explain.

Personal story + image is still the best performing post type. The Belgian citizen post confirms it. But the pattern is splitting into two sub-patterns I didn't anticipate: personal image + genuine tension question produces comments; personal image + warm update produces likes, but not conversation. The "Hot minute — quick re-introduction" post got 2,713 impressions and 63 likes. It got 1 comment. The photo did its job. The close didn't.

Reflection tone is holding on impressions, but it has a comment problem. I post in Reflection tone almost every week, which means it's a new baseline. The posts that generated the most comments in months 4–6 are tagged Contrarian, not Reflection.

And the personal photo question, which I described in January as unresolved: I now have a clearer read. Personal photos are a reach and warmth lever, not a conversation lever. They bring people in and make them feel something. Whether they comment depends on whether the post ends with a question that requires a take. Given that comments are my stated primary metric, the photo is a supporting variable — not the lead one.

Current ranking, based on what the data actually shows for comments: Contrarian tone + open list format is first. Personal image + genuine question is second. Reflection tone alone is third.

The Spryngbase Diaries (my build-in-public series), which is supposed to function as my comments engine, has produced exactly zero comments in the two most recent episodes tracked. EP08, "The One Where We Got Roasted," got 12 likes and 0 comments. That's the post I was most invested in strategically. It's also the worst performer in the dataset on the metric I care most about.

Meanwhile, a shitpost about Sam Altman got the highest impressions I've ever recorded.

Ughh I hate you, LinkedIn.

This incident made me realize that virality and relevance are not always congruent. It is quite difficult to be viral with your target audience on a platform like LinkedIn when you are expressing your opinion about an external incident. The algo doesn't know the difference between the right 10,000 people and the wrong 10,000 people. It just knows that someone is engaged.

I started experimenting with content that was a bit more ragebait. More topical. It felt inauthentic and performative. I knew that as I was writing it.

Buutt, I did it anyway. Because I was watching the numbers. And that is exactly the drift I warned against in January.

Memes. Ragebait. Shitposts. Your own photos. Basically whatever you posted on Facebook in 2011.

On a more serious note, LinkedIn is still one of the best channels for lead-gen in B2B. If you read the January piece and this one back to back, the thing I want you to understand that the data alone wouldn't tell you is this: experimentation is good. Relying solely on vanity metrics as a measure of success is problematic.

My posts are getting more impressions. But the number of leads engaging with my content is still small. I fear the chase for metrics distracts me from the actual purpose i.e. using LinkedIn as a channel to garner leads for my business. The two things are not the same, and I have been conflating them.

I also wrote in January that I was "unsure how long the trend of putting your own picture up to grab attention is going to last." I am now noticing that I see less of the content I most respect (Marie Martens, Jess Cook) and more general memes and quick-consume content filling my feed. 

The Facebookification of LinkedIn is well underway. Whether the personal photo trend survives that shift, I genuinely don't know.

What I do know is that I still believe stories are the only way to be compelling and interesting enough to your audience. I just think LinkedIn is testing my patience with making content that is relevant to my audience versus trying to gain reach via impressions and engagement. Those are different games. Winning the wrong one is still losing.

I am not drawing a clean conclusion here because I don't have one. The Spryngbase Diaries strategy needs a rethink (either the format, the platform, or the close.) 

The open list format deserves more experiments. And I need to be more honest with myself about what metric I am actually optimizing for, before I decide which content to prioritize.

Six months of data is enough to see patterns. It is not enough to know which patterns to trust.

I'll report back.

Grow-th Architect is where I (Shreya Vaidya) think out loud about brand-building, positioning, and surviving on LinkedIn. If this resonated, subscribe, or share it with someone who needs to hear it.