12.11 — The SaaS industry, ARR fraud, the con#

Ocarina rejects SaaS. The SaaS industry is a circular ecosystem where every tool sells to every other tool, where declared revenue is inflated by convention, and where every layer of the capital chain (LP, VC, founder) deliberately maintains a plausible deniability so as to sue the others when the time comes.

1. Contemporary SaaS#

2. Cross-subscriptions#

   ┌──────────┐ pays     ┌──────────┐
   │ Vercel   │─────────>│ Stripe   │
   │          │<─────────│          │
   └──────────┘     pays └──────────┘
        │                     │
   pays │                     │ pays
        v                     v
   ┌──────────┐          ┌──────────┐
   │ Linear   │<────────>│ Notion   │
   └──────────┘   pay    └──────────┘
        │                     │
        v                     v
   ┌───────────────────────────────┐
   │ Slack, Figma, Loom, Sentry,   │
   │ Datadog, Clerk, Cursor, …     │
   │ all subscribed to all         │
   └───────────────────────────────┘

• Vercel pays a subscription to Stripe (billing), Datadog (obs), Linear (planning), Figma (design), Loom (internal presentations), Slack (comms), Notion (docs).
• Stripe pays a subscription to Vercel (hosting for its dashboards), Linear, Figma, Slack…
• Linear pays a subscription to Vercel, Stripe, Notion, Figma, Slack.
• Notion pays a subscription to Linear, Stripe, Slack, Figma.

And so on. Everyone pays everyone.

3. MRR → ARR and fraud by convention#

Definitions#

Conventional fraud#

ARR = MRR × 12 is a mental model:

1. MRR contains prepaid annual plans divided by 12. If a client pays €12,000 upfront for the year, some count +€1,000 of MRR. If they churn at month 3, the “MRR” was never recurring.
2. MRR contains ignored promos / discounts. A client at 50% off for 6 months is counted at full price in MRR by some.
3. Projected ARR ignores churn. The ARR projection assumes 0 churn, while realistic SaaS B2B/B2C churn is unpredictable.
4. MRR contains optimistic users after trying the free-tier. A trial that converts to paid is counted immediately as perpetual MRR.
5. Future upsells are sometimes included in some ARR presentations ("net new ARR", “expansion ARR”).

Result: an MRR of €100K can produce a declared ARR of €1.2M while the actual annualized revenue will, perhaps, be €600-800K.

Why nobody says it#

Because no one has an interest in saying it — incentives align everyone on inflation:

4. The con, plausible deniability#

The fraudulent mechanism works as long as it grows. The moment it breaks (recession, rates climb, failed IPO, audit), people need to be able to turn around. Every layer thus built, from the start, a plausible deniability:

┌─────────────────────────────────────────────────────────────┐
│ LP (Limited Partner)                                        │
│ "The VC presented me a mark-to-market. I believed it. I am  │
│ a victim."                                                  │
└────────────────┬────────────────────────────────────────────┘
                 │ sues ↓
                 v
┌─────────────────────────────────────────────────────────────┐
│ VC (General Partner)                                        │
│ "The founder presented me his ARR numbers. I took them at   │
│ face value. He's the 'representations & warranties' party." │
└────────────────┬────────────────────────────────────────────┘
                 │ sues ↓
                 v
┌─────────────────────────────────────────────────────────────┐
│ Founder                                                     │
│ "My CFO computed. My auditors validated. I signed what my   │
│ teams produced."                                            │
└────────────────┬────────────────────────────────────────────┘
                 │ sues ↓
                 v
┌─────────────────────────────────────────────────────────────┐
│ CFO / Auditor (Big 4)                                       │
│ "The founder provided the data. I applied accounting        │
│ standards. If the MRR definition was fuzzy, the US-GAAP /   │
│ IFRS framework is fuzzy."                                   │
└─────────────────────────────────────────────────────────────┘

Each one leaves themselves a legal exit: “I didn’t know, those numbers were presented to me, it’s <layer below> who lies”.

Note: this diagram is a deliberate caricature of the plausible-deniability mechanism. Legal reality is messier: suits don’t systematically descend layer by layer; the SEC or DoJ often intervene directly without following this order. The structure remains nonetheless representative of each layer’s self-protection incentives.

Publicly documented case studies:

5. The role of SaaS for testing tools#

That’s where the argument touches Ocarina directly — three concrete examples:

Cypress.io#

• 2017 — release, open source.
• 2019 — Cypress Dashboard SaaS (parallelization, analytics, recording).
• December 2020 — Series B of $40M led by OpenView.
• Model: you can’t truly parallelize Cypress without paying for the Dashboard. Lock-in.

BrowserStack#

• Model: you pay per parallel session. An e2e suite of 200 tests, 4 browsers, 30 min: the annual bill gets salty.
• Sold as “infrastructure” when it’s a rental of standard Selenium VMs.

Sauce Labs / LambdaTest#

• Same as BrowserStack. Commodity sold as platform.

Ocarina’s observation#

You can do exactly the same with:

• Python 3.14 (free).
• Selenium (free).
• Ocarina (MIT, 1 dep).
• GitHub Actions (free up to reasonable volume).

Annual cost: €0. Subscription cost: €0. Audit: trivial (a single repo to read). Lock-in: zero (MIT, code readable in an afternoon).

6. Why the manifesto is a war cry on this#

1. The technique needed to test e2e is largely free (Selenium, browsers, Python).
2. The industry invented an artificial paid layer between this technique and the end user (Juicero press).
3. Engineers are the ultimate dupes: they pay through their company for tools that the open-source community produces more cleanly.

In cases of huge companies like Vercel or Stripe, the model is understandable.
In the case of small startups, it’s nothing but cargo cult or completely naive mimicry by kids who want to play grown-ups.

Ocarina is the alternative — not by saying “our SaaS is cheaper”, but by saying “there is no SaaS, the code is there, read it, MIT”.

7. Connection with the other primer files#

• 10-infopreneurs-tugan-bara-ai.md — Same mechanic on the info-product side: selling air to people who resell it themselves.
• 08-ocarina-in-testing-industry.md — The shift Ocarina proposes against this industry.
• ../11-independence/03-explicit-refusals.md — The refusals of SaaS, cloud, platform.
• ../11-independence/02-auditability.md — Why 1 runtime dep is an accounting argument, not just an aesthetic one.

8. And then?#

Ocarina’s implicit thesis:

1. The 2022+ rate hike will deflate ARRs mechanically.
2. LPs will start asking for much more serious audits (already visible in 2024-2025).
3. Founders who can no longer raise will trim their SaaS stack.
4. The moment when a team will say “we cancel the 15 €5K/month subscriptions and write our stack in-house” is imminent.