The three layers of running cost

The three layers of running cost split what a software system costs after go-live into infrastructure (databases, storage, media), operations and availability (keeping the system working and current — monitoring, security patches, dependency updates, adaptation to provider changes, with the response-time commitment as the price variable), and capability development (new features, further automation). The first two are the cost of having the system at all; only capability development is discretionary.

Why it matters

A single "maintenance and support" line hides the one decision a buyer can actually make: what can be dialed down without breaking anything. Split into layers, the running cost becomes negotiable — storage is arithmetic, the response-time commitment is a choice, and the development budget can be declined for a quarter without consequence.

Common confusion

The three layers are not a maintenance fee expressed three ways — and the split is not a claim that all engineering after go-live is optional. Software does not wear out, but its surroundings change: security fixes, dependency migrations, and provider-API changes are part of keeping the system current, and they live in the operations layer. What is optional is new capability. A vendor who folds capability development into "maintenance" has merged a choice with an obligation.

Where we use it

The framework carries What a quoting system costs to run, and the contract checklist asks for the retainer itemized by layer.