Defending decisions with evidence

Defending decisions with evidence means justifying choices using measurements, tests, user impact, incidents, compatibility data, and stated constraints—not authority or preference. When challenged, clarify the criterion and show how your option scores.

| Evidence type | Example use | |---------------|-------------| | Metrics | p95 dropped 40% after cache | | Tests | Contract test caught breaking change | | Incidents | Outage implicated sync path | | User research | 30% abandon on slow step | | Constraints | Team lacks K8s ops for v1 |

					Interviewer: "Why not GraphQL?"
You: "Given 3 mobile clients on stable REST contracts and 2-week deadline,
OpenAPI + versioning meets need; GraphQL adds gateway ops we can't staff.
I'd revisit if clients need flexible field selection."
				

On interviews: welcome pushback—treat it as a chance to refine criteria, not as personal attack.

Common pitfalls: doubling down without new data; hand-waving "industry standard"; ignoring legitimate new constraints mid-interview.

The trade-off is confidence versus intellectual honesty when evidence is incomplete.

Checklist:

• Tie recommendation to explicit criteria.
• Cite concrete evidence types.
• Acknowledge weaknesses of your choice.
• State what would trigger a different decision.