@lemma_ shot theirlaunch video inside the world of Obsession - 2.3M views
@Heypocket put Louis Litt from Suits in theirlaunch video - 1.3M views
@intelligenceco literally put@ndrewpignanelli's grandmother in thevideo - 4.6M views
And you remember all three.
In April we ran our own launch.
https://x.com/filippocarnvl/status/2048112389216981381
It didn't get millions of views, more or less 370K across the launch.
But the ROI was unbelievable:
In three posts, this video contributed to six figures in deals to our media company and a major chunk of our portfolio came from it alone ( we did videos for companies like Merge, Parallel, Extropic etc).
Tech founders loved it.
But, we didn't put a dollar of ads behind it.
The only thing it had going for it was that nobody had seen a launch video that looked (and sounded) like that.
The video got watched because it was new. And in 2026 getting watched is the whole game, because the people watching are the people buying. Attention is money now, as you've probably heard a million times.
Most launches don't get watched.
It's the same founder at the same desk saying "we're excited to announce” with the same motion graphics and everyone looks the same.
I'd take 370K views and a lot of inbound over millions of views and a bunch of congrats comments any day.
So here are four things those three launches (and our launch too) had in common, that brought them real customers.
1) Every format will eventually die
X shows your post to a few hundred people first. If they keep scrolling, that's it, the post is done.
Early this year the talking head launch style worked really well. A founder, a good hook and narrative, showing the product on screen.
It worked because nobody had seen it before. Then every company and every agency started doing it, and by summer it was getting annoying to most founders.
That's how every format goes. It works until enough people have seen it, and then it stops.
You remember the three launches mentioned earlier? Now try to name one talking-head launch you scrolled past last week. You can't, and you've probably seen forty of them.
Pocket went viral with their first video before the product even existed and got thousands of pre-orders off it.
https://x.com/AkshayNarisetti/status/2077439335130124544
The one thing they did was that nobody had seen a launch look like that before.
2) The parking lot method
The way I explain it to founders is a parking lot.
Five cars, all the same. You walk past and you don't look at any of them. Make one of them a Ferrari and everyone looks at it, not because it's expensive but because it's different from the other four.
Now park five Ferraris. Nobody looks again.
That's the feed. Your launch is one of five hundred cars, and most of them are the same car.
There are a few ways to be the Ferrari, and we've done a couple of them in our own launch.
3) Steal from outside your category
Don't copy other launch videos 1:1. Take something from them, but mix it with things from other worlds.
Take the reference from somewhere your buyer has never seen next to software.
@lemma_ took a horror film. They recreated the diner scene from Obsession and showcased the product inside it.
Pocket took a legal drama and hired Rick Hoffman, the guy who plays Louis Litt.
A true-crime doc, an old infomercial, a reality show, whatever.
The rule is that the product should be the only thing in the frame that belongs to software.
It works because the viewer has no reference for what they're looking at, so they have to keep watching to figure it out.
4) Do the opposite of everyone else
Write down what every launch in your category did this month. Fast cuts, big music, a founder talking at 1.5x speed, "introducing."
Then take the loudest thing on that list and flip it.
Every launch that month was fast and loud. intelligence.co made theirs slow and calm: a soft voiceover, a grandmother, a laptop running agents on the grass - 4.6M views!
When every video is shouting, the quiet one is the one people actually watch.
Ours did both of these. Plus one more thing that's harder to copy.
What was actually in the 18 seconds of our launch
Most launch videos that month were a founder at a desk for 90 seconds. Ours was 18 seconds of text, flashing transitions, an ASCII hacker look nobody had done before, and a Kettama track that nobody had ever used in a launch video lol.
That's how we mixed different things and made something new.
And that's the opposite move: 18 seconds instead of 90, text instead of a talking head.
The video also didn't describe or pitch what we do, it did it.
Instead of saying "launch videos are the best way to launch" in the video, we proved it. We sell launches that get watched, so our own launch had to get watched or nothing we said afterwards would matter. It got watched 370K times.
What to do with this
Every format you're watching go viral right now is already dying. Ours will too. At some point enough people will have seen the ASCII look and the flashing text, and we'll have to come up with something new again. That's the job.
So before you greenlight your launch video, ask one question: has my buyer seen this exact shape in the last 90 days? If yes, you're parking the fifth grey car.
Take a reference from outside tech, or do the opposite of what everyone in your category is doing. Ideally both.
The launch that gets watched is the one that gets paid. Ours got us 60+ demos for our service and 6 figures in revenue from 18 seconds, and it didn't need millions of views to do so.
We're looking to create more of these unique launches with ambitious founders. If your launch is on the calendar, lets talk!
EPTA.
