Lesson 4 · Operations
Build vs buy & vendors
Decide on total cost of ownership, and read the contract before you sign it.
Build-versus-buy is decided on total cost of ownership (TCO), not purchase price. Building costs the initial development plus maintenance, on-call, upgrades and the opportunity cost of the engineers — and over a system's life, maintenance usually outweighs the first build. Buying costs licences plus integration, migration, training and the vendor's future price increases.
Buy when the capability is a commodity with a mature market; build when it differentiates your product or no vendor fits. Weigh lock-in, the cost of leaving later: data export, proprietary APIs, retraining. Lock-in is not always bad. It is a price, so know what you are paying.
When you buy, you manage a vendor. Read the service-level agreement: what is promised, how it is measured and what the remedy is — usually service credits, not your losses. Check security and compliance, data ownership and exit terms before signing. Negotiate; multi-year commitments and good timing buy discounts, and procurement is an ally. Reassess at renewal, because the right answer three years ago may not be right now.
Picture it
A team wants to build its own feature-flag service: “it is just a table and an API — two weeks”. The manager prices five years: the build, then about a fifth of an engineer for maintenance and on-call, plus the audit log, percentage rollouts and SDKs that will follow. Roughly $250,000. A vendor costs $30,000 a year, or $150,000 over five. They buy, and spend the two weeks on the product.
Key formulas & frameworks
- TCO of building
- Development + Maintenance + Operations + Opportunity cost
- TCO of buying
- Licences + Integration + Migration + Training
- Lock-in
- The cost of leaving later
- Reading an SLA
- What is promised · How it is measured · What the remedy is
Practice
Score: 0 / 5
Up nextScaling the organization
Next lesson