the ownership economics of business software

Every company now rents most of its operational software. This paper argues that for a specific, identifiable class of tools, renting has become the expensive option — and provides the model to identify that class in your own stack.

1. How metered pricing compounds

Per-seat, per-event, and per-evaluation pricing share one property: the bill is indexed to your growth, not to the vendor's cost. A support desk priced per agent doubles with your support team; an analytics tool priced per event doubles with your traffic. Over a three-year horizon, a tool priced this way typically costs 3–5× its flat-priced equivalent — before the discounted renewal negotiation that never quite lands.

2. What changed about self-hosting

• Containers collapsed deployment to a pull and a restart — the ops tax that justified SaaS margins is mostly gone
• Postgres, Redis, and ClickHouse now cover nearly every storage pattern a business tool needs
• Open licensing matured from ideology to procurement requirement — MIT/Apache code is an auditable exit door
• Cloud marketplaces made 'runs in your VPC' a checkbox, not a project

3. The five-question ownership test

• Is the tool a system of record? (support history, identity, analytics — own it)
• Is pricing indexed to your growth? (per-seat/per-event — own it)
• Is the data too sensitive to live in someone else's database? (AI traces, prompts — own it)
• Is the vendor's network effect the product? (payments, deliverability — keep renting)
• Is it genuinely undifferentiated and cheap? (keep renting, revisit yearly)

4. The operations objection, priced honestly

Someone must upgrade, back up, and monitor owned software. That labor is real — and purchasable as a flat monthly line item, which is precisely what managed hosting is. The honest comparison is never 'SaaS vs. free'; it's 'rent software + data + exit rights forever' versus 'own software + data, rent the labor while you want it'. Run at your numbers, the second bargain wins for systems of record almost every time.

5. Sequencing the transition

Start where metering hurts most and migration risk is lowest — usually feature flags, internal dashboards, or the support desk. Each migration funds the next: teams we work with typically redirect one tool's annual subscription into a fixed-price sprint that eliminates that subscription permanently. Ownership compounds the same way rent does — the direction is the only choice.