My (WIP) buy box
A guide to my investing, 2035's status symbols, and fake revenue/ARR
Every year we do some kind of offsite or multi-day hang to get our macro view together, work out operational kinks, make big calls, etc.
One of the deliverables coming out of this year’s hang was writing a “buy box” - a guide to share within the firm. The ostensible purpose was/is to help your second push your thinking by explicating your process and highlighting your weaknesses. Note: instead of an investment committee we just have a sponsor/lead and a second and no one has veto power as long as you find a second “yes.”
But beyond that it’s a great exercise to define what you think matters and how you want to make decisions, at least for right now.
So here is my buy box: Weird takes on important stories with a shot at real earning power led by a spikey team who I can back at an aligned price.
What I like
Themes and stories
Weird takes on/second order and third order effects on/downstream bets on really big important stories. Sometimes it just means looking at the biggest stories and asking “so what?”
AI
Old people/demography
GLP-1s
Climate change
Trump/instability
Attention economy
End of companies/talent wins
Networks, platforms, software replacing traditional orgs
Enabling talent markets
Age of inbound/trusted networks/owned audience
What happens when AI kills outbound sales and inbound trust.
Everything is scalable and fake, what next?
New consumer interfaces (beyond chat)
Founder attributes
Barbell: call your shot (experienced, deep, make a big bet) or call your customer (ship your way to right) with the ability to back it up.
Evidence of agency and velocity
Commercial but not games-y
Ineffable charisma. Lucid vision, sales, storytelling such that people root for them to win
Gritty, chip on shoulder but not insecure. Mature, reflective, self knowing
Asking for feedback but not permission
Obsessive/hyper fluent
Deals attributes
Cheap! (high CoC at the onset)
But more important to have ownership
Explainable access in a world with infinite capital
Capital aligned with ability to spend well
Enough money to get them somewhere
Business models
Needs durable pricing power at some point. Hard but not impossible to find in pure software.
Services differentiating software
Software differentiating services (vertical co’s, GBOs)
Hybrid hardware or IRL with software
Network effects/multiplayer mode
Real depth
Biz attributes
No double back flips
Clear path to value/value prop to customers
Company/team attributes
Building in person unless there’s a good reason not to
De-risked relationships (co founders know how to work together)
Technical (co)founder
Clear CEO/leader
2035’s hottest status symbols
As cognition and attention get fried and physical improvement is increasingly hackable, status symbols will be all about patience and self direction (often/especially toward unnecessary ends).
Between AI, short form video, waifus, and ozempic, the status symbols of today and tomorrow the historical status symbols will deteriorate in value.
Here’s specifically how that will manifest itself over the next decade:
Cooking your own food. Uber Eats is washed and wage-y coded. There’s no charm in the protein slop. Cooking is both a marker of unnecessary skill and also of health, agency, and control - especially once the appliances can fully do it for you.
Running marathons. Being skinny is/will be easy and even being muscular will be only marginally harder (you do still have to work out even if you’re juicing). But you can’t fake your way through the misery of joyless, boring running.
Language learning. We’re going to have widely accessible babelfish within a few years. There will be zero practical benefits to speaking another language (already people are outsourcing CX email jobs and using AI for translation). Language learning will be a delightful hobby and is already a great flex but in the future it’ll be entirely performative. People will start learning dead languages as a status performance.
Reading novels. Non-fiction feels more and more striver-coded to me, especially as most people just pretend and get summaries on podcasts and from AI. But reading a novel and really engaging with it beyond mere language comprehension is an exercise in interest, not output. (Watching movies will probably fit into this as well).
Having friends. This is probably going to be the biggest one. Any hobby or activity that requires you to be with other people, and ideally the same people, repeatedly. Actually socializing with people (vs playing videogames or hanging out with convincing simulacra) is already becoming very high status.
Basically there will be a revival of gentlemanly hobbies and anything that involves comprehension, tactile experience, and patience will be highly prized. Function and outcomes can be faked.
Elsewhere
This week there two separate articles in Fortune and The Information, respectively, about the “$100 Million ARR” club for startups. They each observed two things: 1) companies are growing faster than ever and 2) startups are fudging/fabricating their numbers more than ever - or at the very least there’s more debate over what the numbers mean.
Those two points are both true and somewhat mutually reenforcing but they are different things.
This comes back to what I’ve written about before in “the end of pricing power” and “revenue comes second.” TLDR:
#1 Companies are growing faster than ever.
This is demonstrably happening and it’s for a mix of good and bad reasons.
IMO one of the biggest reasons is that startups are increasingly “selling work” (attacking opex line items and replacing people, especially outside vendors). That means the contracts are bigger and easier for customers to underwrite. You can make much easier 1:1 cost comparisons between a low margin, venture-backed AI business and your contractor or support worker.
Secondly, customers are under tons of pressure to deploy AI/AI-ify themselves. So they will experiment rapidly and adopt quickly, often as temporary measures without much of an eye toward costs, which brings us to:
#2 startups are fudging/fabricating their numbers more than ever.
The biggest this happens is by treating “CARR” or some kind of contracted revenue as signed revenue (alternatively reporting on GMV as revenue). Much of this is the result of “experimental budgets” where a customer buys 10 seats for a 100 person company, which the vendor then reports back as a 10 person pilot that **will** grow to 100 seats total. It trickles down into things like LOIs, expansions, product cross-sell, etc. getting called revenue. Whatever.
This is all obviously kind of dumb but it does matter insofar as it leads to discounting by investors which then requires everyone else to similarly inflate their numbers to keep up. If the leaders are effectively inflating their numbers by >2x and you’re not, you look like a loser.
So what?
To be clear: there are genuinely some really insane growth stories happening out there for good reasons. But also not all revenue is ARR and not all ARR is the same… so people are starting to focus (or at least say they’re focusing) on margins, retention, etc. We will see some historic revenue declines and busts in the coming years given how easy it for vendors inflate numbers and for customers to swap vendors.
For my money, I really like what Bret Taylor said here:
“The faster you grow is sometimes correlated with a lack of a moat… I’d rather bet on durable, high-quality revenue and happy customers than just raw growth.”
Loved the take on "treating “CARR” or some kind of contracted revenue as signed revenue" - asking founders and going deep on what the numbers really mean is key during the process