Startup marketing, billing, pricing, AI monetization by Arnon Shimoni
-
Building a billing system is still hard, even with AI
8 years building billing systems and I’m still wrong about just how hard it actually is. Not “getting better at it hard”, but like really genuinely discovering new layers of complexity I never saw coming. — Back when we started Paid at the end of 2024, I remember people telling me “billing is not hard,…
-
The “stranded asset” in AI pricing
With hybrid “AI pricing”, users face siloed entitlements, creating “stranded assets” where purchased credits remain unused. This misalignment leads to significant churn risks as companies reassess value versus spend amid rising AI costs and limited access to resources.
-
Building a startup marketing engine when you have nothing
I think lots of founders misunderstand what early-stage marketing is. You don’t start when you get funding, or when you’re ready, or even when you hire your first marketing person – you start when you decide who you want to be in the world but more importantly: who you need to beat. When I cofounded… -
Operational debt is worse than tech debt
I’ve touched billing systems inside several companies now, and the same pattern shows up every time: the hard part of migrating off seats isn’t the software, it’s the operational debt baked into the business. This is why billing system migrations take at least 8-10 months. You start (well, engineers start) by thinking a pricing change…
-
The trust collapse: Infinite AI content is awful
Content is infinite and free now. Trust isn’t. We’re abandoning digital channels entirely because we can’t tell what’s real anymore.
-
Notes on where seat-based pricing is going
Everyone (mostly me) keeps saying seat pricing is dead, and some people are arguing with me online. I think that’s cool, but it’s because they don’t understand why it’s dying. For twenty years, SaaS revenue scaled with headcount.More humans = more seats.That logic held so long that we stopped questioning it. Then the humans disappeared.…
-
MIT’s 95% AI failure rate is wrong
The MIT NANDA report claiming 95% of AI projects fail is making rounds on LinkedIn. Everyone’s sharing it like it’s gospel. Has anyone actually read the methodology? They interviewed 52 organizations. That’s their entire sample. We’ve worked with over 250 companies at Paid, and the reality looks completely different. The flaw in their math MIT…
-
The free tier death cult
One user consumed tens of thousands in model usage on a $200 plan. That’s from Anthropic’s own admission when they killed Claude Code’s unlimited tier. Not $200 worth of usage. Tens of thousands of dollars on a $200/month plan. TechCrunch said Cursor hit $500 million ARR. Fastest growing application ever, but still losing money on…
-
The SaaS pricing apocalypse
You know that feeling when you’re looking at your competitor’s pricing page and the math doesn’t add up? $9/month. Unlimited usage. Same features as your $900 enterprise plan. Your first thought: They’re burning VC money. Classic Silicon Valley playbook – grow at all costs, figure out margins later. You’ve seen this movie before. They’ll either…
-
The 5 stages of SaaS death
Last week, I met a founder who’d just lost a pretty big $2M deal to a 3-person startup with a half-finished product (I call these “3 Stanford grads in a trenchcoat”). He was annoyed they didn’t even have a basic dashboard or settings page. I think he wasn’t angry about losing the deal, but mostly…