|
Hi Reader - There's a robot in The Hitchhiker's Guide to the Galaxy called Marvin. Full name: Marvin the Paranoid Android. He's the smartest thing in the universe, and he's miserable about it. In one scene, he tells a mattress he's thirty billion times smarter than it is. To prove it, he runs a little test. "Think of a number, any number." "Er, five." "Wrong. You see?" The mattress comes away very impressed. I think about that scene a lot when we use AI to analyze ads. Because the analysis part is brilliant. We often look at apps that launch 100s (sometimes 1000s) of different ads in a single week. No human wants to watch all of those. Our AI, BruteForce, went through every one: the hook, the format, the emotion, the mechanic. Then it sorted them into families. Three families made up about 80%. So I did what anyone would do. I asked it what we should test next. And that's when it turned into Marvin. An enormous brain with total confidence. And some genuinely good ideas. What it doesn't give you is an ad that's ready to run. It DOES gets you about 80% of the way there. However, when you try to ship 80% of an ad, you realize it sucks. The technical term is ‘slop’ - even if it no longer is a mutant with six fingers. It may be very photorealistic slop made with Seedance 2.5, but it is still slop. Say you ask it for a sleep app script. You get a tired mom who can't sleep. She downloads the app. She wakes up refreshed and happy. It's good enough that everyone in the meeting would nod. Pre-AI, a lot of companies would even ship it. That's the problem. You've seen that ad a thousand times, and in a world of ‘good enough’ ads that are proliferating in volume, it is going to die a silent death. To sharpen a script like that and take it from 80% to 100%, we built a storytelling skill with 12 frameworks, 5 story spines and 8 guides in it.
(Yes, we have a file on Kishotenketsu. I'm a lot of fun at parties.) Marvin has read every damn framework in the world. He can recite them to a mattress. What he hasn't done is lie perfectly still at 3am so someone he loves won't feel guilty. So someone has to pick the framework. Does this story need another obstacle, or a turn that changes what we just saw? Then you workshop it. Does she come back for her phone, or does he say something first? Does he turn off the alarm, or does she sleep through it? That's the last 20%, and that's the part that takes a human. Not just a rando human off the street, but one that has breathed and studied direct response copywriting principles that have stood the test of time for DECADES. (Here is the kind of narrative and story you get - and yes: we make THOUSANDS of these every month). Which is why, even though I spend 12-14 hours on Terminal a day, I (and our team) work hand in hand with the AIs to make sure every ad we put out portrays the real raw lived human experience of the people we serve. Anywayyy, if you're spending $50k+/month and want AI doing the analysis with humans finishing the work, reply "MARVIN" and we'll talk. Or don't. I’m sure the next ChatGPT prompt pack will let you crush it and post ‘RIP creative strategists.’ Later, P.S. I put 314 podcast episodes and 15 books into the Mobile Growth MCP, a library your AI can search to answer YOUR hardest questions about growth. New episode of Intelligent Artifice is live: How a two-person app studio launched 16 apps with AI, with Gabe Kwakyi. Here are some goodies from our socials: A 5 year old mobile game with 14M+ installs had hit a growth ceiling Sometimes we tell clients to spend less Hard-paywall apps convert 5x more downloads to paid than freemium apps After analyzing 126 Wispr Flow ads and 1,025 reviews, I found 4 major creative gaps |
Intelligent Artifice features hard won insights from the front-lines of performance marketing in a Generative AI world - delivered straight into your inbox. For free.
Hi Reader - Years ago, I was working on a word game, and we were killing it in the US. So we ran an experiment in the UK. After all, they speak English too, right? I think it was George Bernard Shaw who said: England and America are two countries divided by a common language. (Whoever it was that said it, I was about to prove it.) Anywayyy for our word game, the UK did even better than the US. CPI was strong. Cost per first puzzle played was strong. D1 ROAS was extremely strong, and D7 held...
Hi Reader - Before I scaled an app from $30K a month in ad spend to $1.5M(profitably), first I cut it to zero. I've told the zero part before. What I didn't tell you is what happened in the two months after. Because turning the spend off turned out to be the easy half. Zero spend fixes exactly one thing: it stops you paying to pour users into a leaky bucket. It does not fix the bucket. And in a five-person company, zero spend isn't a strategy. It's a countdown. Or a ticking time bomb. Pick...
Hi Reader - Most companies run a hackathon once a year. You know the yearly kind. Forty-eight hours, matching hoodies, a demo that gets a standing ovation, and a LinkedIn post about "our culture of innovation." The non-engineers come for the beer. And by March, nobody remembers where the repo lives. That made sense when the tools changed once a year. AI changes them every few weeks. A yearly hackathon is out of date before the pizza gets cold. A couple of weeks ago, in my typewriter email, I...