👋 Hey, I’m Jaryd. I share one stealable idea with you a week on product, growth, AI, marketing, or strategy. My goal is to help with the hard part—growing the thing you’re trying to build. For more: Explore the top ideas | Watch me yap
Today's 5-Bit, brought to you by:
 |
Good Times — answer one question a day with friends Create my group and answer |
 |
quicker.chat — one inbox for all your DMs I want all my DMs in one place |
Good morning guys, and happy Friday!
I’m tinkering a bit with the layout/design of the newsletter at the moment. If you know me, you know this is my favorite thing to do. My goal is a quicker and more scannable read for you with less scrolling, and with the same helpful information. Lmk what you think, I love any and all feedback.
Who is actually paying for AI? Two charts
What's going on: The top 1% of US AI payers made up 19.5% of all consumer AI spending in August. That’s more than the bottom 50% combined. Top spenders average $903 a month—up 80% in 18 months. The median payer spends $25. And Claude has the biggest spenders with 7X the share of payers on the top plan than ChatGPT.
As models become better more of us are spending more
Claude converts to higher plans more aggressively
Why it matters: The AI spending charts are looking a lot like mobile gaming. A few people spend a lot and everyone else spends little or nothing. Also, it’s clear heavy work is happening on Claude, where it has less mainstream consumers like ChatGPT, and more prosumers like us buying software to build, create, and get more complex work done. AKA vibe coding apps with no users 😂
The idea behind this: Find your heaviest users and build a tier for them: usage top-ups, a power plan, faster responses, better models. Keep the cheap plan for getting people in the door and let them expand and upsell themselves.
How is Muse growing so fast?
What's going on: Muse cracked over 5M downloads 22 days after launch. ChatGPT took 56 days, Grok 103 and Claude 492. Only 23 apps have ever reached that milestone so fast and ~85% were games. How? Because Meta pointed up to half of its daily house-ad impressions at Muse, across Facebook, Instagram and WhatsApp.
Why it matters: Meta used unparalleled reach it already owned. Billions of people see their own ad slots every day, and pointing them at Muse cost almost nothing extra. Every single one of us saw Muse ads more than once a day.
The idea behind this: Use every surface you have possible access too as distribution. Settings pages, dashboards, emails, newsletters, whatever. Social handles and bios. Whatever you can touch, use it as reach.
Why did this move bring Nike to a 12-year low?
What's going on: Nike beat earnings, with $0.48 a share against roughly the $0.43 they expected. Yet their stock is down 9%—its lowest since early 2014. Revenue fell 4% to $11.2B, and direct sales fell 8%. They expected it to fall again.
Why it matters: Nike’s strategy for years has been getting away from wholesale and betting big on direct sales through its own app and shops. They called it “consumer direct offense”. The good side was more margin per sale, but the downside—less shelf space where people are shopping. Nike has been trying to win those partners back ever since.
The idea behind this: Owning the customer relationship only helps if the customer is still there. Always place your bets and go where buyers already shop, even if you pay for it by giving up some margin to get there.
Would you pay $4 to snoop on Instagram?
What's going on: Instagram wants to let you spy on profiles without leaving a paper trail. They’re internally testing a “read-only mode” which really, just makes stalking easier. It will be part of Instagram Plus for ~$4/mo.
Why it matters: Instagram found a tiny, very specific anxiety that a lot of people share, and it’s testing whether they’ll pay to make it go away. Everyone in some format has felt the fear of accidentally double-tapping a 2019 photo of an ex.
The idea behind this: Besides Instagram trying to productize dread with a small insurance to protect against embarrassing accidents, I like the meta idea of internally testing things with your team, or “experimental products”. This isn’t QAing something expected for production, it’s doing internal user testing to see if something might work/what the concerns are.
What happens when you delete all your code?
What's going on: This startup, Melius, started by ex-Ramp engineers, was building an AI tool for managing ad spend. 6 months in they realized it was the wrong idea and pivoted to an AI tool for ad creative. They deleted the entire codebase and started again. 2 months after coming out of stealth and they supposedly have $1M in ARR.
Why it matters: Delete is bold. Six months of code obviously feels like an asset, BUT, with AI writing most of it code is now cheaper than ever to throw away. The expensive part was the time spent on the wrong idea.
The idea behind this: Ask yourself this…if I was starting today, would I build this thing? If the answer is no, starting over is probably faster than fixing it. And deleting is a big move but it clears your mental plate and reduces the carrying cost of old code somewhere. I could probably delete a lot of this 😂
not even a fraction of my half baked ideas
That’s it for this week. Thoughts on the briefer format?
—Jaryd (if you’re wondering who am I…👇)
|
