This is an ongoing series where I explore the intersection of design with the behavioral economics concept of signaling.

——————————————-

Thesis: In product design, speed isn’t just a convenience; it’s status. Fast systems feel competent, premium, and safe. Slow ones feel cheap and confused.

Why it signals:

– Humans map response time to care and capability. Sub-300ms says “we prepared.” Multi-second says “we’re guessing.”

– In AI experiences, speed acts as a truth mask. If you answer confidently and quickly, the burden of proof shifts to the user.

How to install it:

– Set latency budgets (p50/p95) per critical step; design the UI for each budget.

– Use optimistic UI and predictive prefetching when correctness is high.

– Replace spinners with structured skeletons that prove the system knows what’s coming.

– Do hard work on-device when possible; reserve server calls for uncertain branches.

– Make errors as fast as successes; slow failure is where trust dies.

Metrics that matter:

– p50/p95 latency per step, completion rate under budget, task abandonment before/after.

Where to use it now:

– Clinician tools: chart pulls, order sets, handoff summaries—budget to p95 < 1.5s.

– Mobile triage: first helpful step in < 2s; fix suggestion in < 5s.