Screens are symptoms. Experience problems are the root cause.
This playbook runs the same optimization logic at three depths. Strategic mode targets core journeys and platform-level UX. Standard mode addresses feature usability uplift. Lite mode handles UX papercuts, friction fixes, and growth tuning. All three modes require decisions to be evidence-based — not taste-based.
Principles
Four commitments to keep optimization honest.
Experience first, features second: we optimize user effort, cognitive load, and task success confidence — not screens. Evidence over opinion: decisions are grounded in research signals, usage data, and behavioral patterns; design intuition is valuable but must be explainable. System thinking beats patchwork fixes: prefer pattern improvements and rule alignment over one-off UI changes that don't scale. Async and explicit: UX problems are documented, not debated — artifacts over meetings, decisions leave a trace.
Stage 1 · Diagnosis
Understand before optimizing.
Key journeys are walked through to identify where experience breaks down and why. Friction points, drop-offs, task success rates, and qualitative and quantitative signals are aggregated. Competitive UX pattern comparison provides external calibration. The stage closes when experience issues are specific — not vague — and root causes are identified, not just symptoms.
Stage 2 · Framing
Decide what to fix first.
Problems are translated into hypotheses. Each is scored by impact, frequency, and effort. Success criteria are defined in UX terms — "better" is measurable, not subjective. Scope and non-goals are aligned before any exploration begins. The optimization brief is the contract between design intent and execution.
If it can't be reused, it's probably not optimized.
Stage 3 · Exploration
Improve the system, not just the UI.
Flow simplification, interaction pattern refinement, content and feedback clarity checks, and cross-journey consistency reviews are explored at a structural level. AI generates state and flow variations quickly; designers evaluate against the optimization brief. Component behavior proposals are prioritized over one-off layout changes.
Stage 4 · Blueprinting
Make UX intent unmissable.
Edge cases are resolved. Error states, empty states, and feedback messages are optimized for clarity. System response behavior and data dependencies are aligned. The experience blueprint and UX acceptance criteria remove the ambiguity that causes implementation gaps between design intent and shipped product.
Stage 5 · Execution
Ship improvements users can feel immediately.
UI is polished within design system constraints. Interaction tuning, accessibility checks, and DS refinements are documented. Loom walkthroughs replace live handoff sessions. The measure of execution quality is whether users experience less effort — not whether the screen looks different.
Stage 6 · Validation
Did experience actually improve?
Task success is validated against the criteria defined in framing. Confusion and hesitation patterns are analyzed. Qualitative feedback is synthesized with AI assistance. The output isn't just a list of findings — it's a UX impact assessment and an iteration recommendation that feeds directly back into the workflow.
Stage 7 · Trade-off Review
Be explicit about what improves now versus later.
A UX trade-off log documents improvements shipped, known UX debt, risks accepted, and follow-up triggers. Decisions made under constraint are recorded so future teams understand why something is the way it is — and when the conditions exist to revisit it.
Stage 8 · Learning
Raise the bar system-wide.
Local fixes are generalized into patterns. Design system components are promoted based on validated behavior. Experience principles are updated to reflect what was learned. Each optimization cycle raises the floor for the next one — the loop doesn't close, it compounds.
Experience optimization is not polishing pixels. It's systematically reducing user effort — over time, at scale.
