Architecture Trade-off Analysis Method
also called ATAM
A structured evaluation that scores an architecture against prioritised quality-attribute scenarios and identifies the points where those attributes conflict.
Developed at the SEI, and useful mainly for its central discipline: an architecture cannot be evaluated in the abstract, only against specific scenarios for the qualities that matter.
The method elicits quality-attribute scenarios from stakeholders, prioritises them, and walks the architecture against each, classifying the design decisions it meets along the way as sensitivity points (a decision that strongly affects one attribute), trade-off points (a decision that improves one attribute and degrades another), and risks.
Trade-off points are the output that justifies the exercise. Caching is a trade-off point between latency and consistency; synchronous replication between durability and write latency; a shared database between delivery speed and independent deployability. Naming them makes the architecture's tensions explicit rather than emergent.
The full formal method is heavy and rarely run as written. The parts worth stealing for any design review: scenarios rather than adjectives, and an explicit list of the decisions where two things you want are in conflict.