Lessons From Going Viral

Ok I didn’t go that viral

But something I posted on my other account this time last week suddenly blew up over the weekend - 21k views, 463 likes, 300 new followers, 150 saves

Here are some interesting things that I think you can take from that:

1. Please stop chasing virality

I didn’t even go viral - 21k views is tiny compared to what a lot of people get on the platform.

Yet I was still dealing with probably a dozen ill-informed trolls from the anti-UPF cult all weekend, when I could have spent that time, energy and headspace on coffee, learning, beer.

I’m not trying to get clients with my content right now - I’m optimising for speaking gigs, podcasts etc, so I need to show I have an opinion worth sharing. If I wanted coaching? Going viral would be bottom of my list.

2. The negative comments have to be part of it

Like Bear Grylls drinking his own piss - it’s not nice, but it’s necessary.

I can’t pretend I enjoyed having people tell me how stupid, attention-seeking and money-grabbing I was. But I also can’t pretend it wasn’t nice to have so many people agree, show support and thank me for posting what I did.

You can’t have it one way, rough will come with the smooth

3. The post was ugly, but it said something

People are obsessed with shiny algorithm-optimised content. I did a very poorly designed post, with a very important message. It still did well.

Optimise for having something to say - it’s better for your business.

4. AI content is a fantastic thing.

I wrote the whole post myself - I always do.

But the post was meticulously researched using AI tools, to ensure that my argument was water tight.

So that when the negative comments came, I knew I had a decent reply.

Plus - I needed it to be well researched to have impact.

Using AI to be more interesting and have better stories is incredible - using it to post about why consistency is great, not so much.

5. Know why you’re posting I was calling out people in my space - but because they were directly going against my worldview and therefore I have an opinion that I should share that helps communicate things.

I wouldn’t just post about anyone I didn’t agree with, because half of what I don’t agree with is not remotely relevant to my audience or purpose for posting.

Connections

• [[Boring or Interesting]] — “The post was ugly, but it said something” is the Boring or Interesting thesis in action. The post polarised — trolls arrived and supporters arrived — which is exactly the filtering mechanism that spark describes. Ugly design repelled the aesthetics-first crowd; a strong opinion magnetised the depth-seeking crowd. The 150 saves matter more than the 21k views because saves are a stated-preference signal: people bookmarking what they genuinely value, not what they passively scrolled past.
• [[Stated Preference]] — The gap between virality metrics (revealed preference — what the algorithm rewarded) and actual business value (stated preference — what you need for speaking gigs) is the core tension of this spark. The 21k views optimised for the algorithm’s revealed-preference engine, but the real value was in the opinion itself landing with the right 300 followers. Building for stated preference means accepting that ugly posts with strong arguments outperform polished posts with weak ones — because the audience you want is filtering on substance, not design.
• [[The Lionel Messi of Content]] — “Using AI to be more interesting and have better stories” is Messi’s precision applied to research rather than output. Most coaches use AI to produce more content (running harder on the Red Queen treadmill). This spark describes using AI to make fewer, better-aimed moves — meticulously researching a single argument so it lands with force. Three explosive movements, not ninety minutes of running. The AI-as-research-weapon framing is the underexplored Level 3 application most coaches miss entirely.
• [[Tell people who you’re not for]] — “Know why you’re posting” and “calling out people in my space” is the active version of the Stewart Lee strategy. Lee passively displays bad reviews; this spark describes actively generating the polarising content that produces them. The anti-UPF trolls are not a cost of virality — they are the filter working correctly. Each negative comment from the wrong audience is a signal to the right audience that this is someone willing to take a position, which is the prerequisite for trust.
• [[Science curiosity not science knowledge protects against polarisation]] — The anti-UPF trolls embody Kahan’s finding perfectly: scientifically literate people who lack science curiosity become more dogmatic as they learn more. The trolls were not uninformed — they had read extensively about UPF — but their reading had weaponised confirmation bias rather than generating genuine inquiry. The “water tight argument” prepared with AI was necessary precisely because the opposition was not ignorant but selectively informed, which is the more dangerous failure mode.
• [[The Red Queen]] — “Stop chasing virality” is a Red Queen exit strategy. Every coach who optimises for viral reach is running harder on the same treadmill — more hooks, more controversy, more algorithm-friendly formats. This spark argues for competing on a different dimension entirely: having something worth saying. The 21k views came not from algorithmic optimisation but from genuine opinion, which is the creative escape the Red Queen framework predicts will outperform incremental improvements.

Tensions

• The post went semi-viral precisely because it was polarising — which means the “stop chasing virality” advice is complicated by the fact that strong opinions are themselves viral fuel. You cannot fully separate “optimise for having something to say” from “optimise for engagement” when the algorithm rewards controversy. The mechanism that attracted the right 300 followers is the same mechanism that attracted the trolls. The advice works only if the coach can tolerate the troll cost, which is a personality filter as much as a strategy.
• AI-as-research-weapon requires existing expertise to wield. The argument was “water tight” because AI helped research it, but the direction of the argument — the worldview, the position, the specific claim worth defending — came from years of domain knowledge. A newer coach using the same AI tools to research an argument they do not deeply understand would produce something that looks well-researched but collapses under scrutiny. The AI amplifies the quality of the thinking it is given, not the thinking itself.