Experience Refinement

Make competent products feel considered—feedback, waits, character, and finishes that earn trust without slowing the job.

Prerequisite: Customer Discovery. Core concepts: Emotional Design, Aesthetic–Usability Effect, Peak–End Rule, Surfaces, Flows, and States, Mental Models.

What it is

Experience refinement is the work of making a product that already does its job feel like it does its job—closing the gap between functional correctness and the felt sense of quality that comes from feedback, waits, character, and finishes that were clearly designed on purpose rather than left to whatever the framework rendered by default. This is not the same as Onboarding or Activation, which are about getting someone to value for the first time; refinement is what happens to a flow a user has already succeeded at, making the tenth time through it feel as considered as the flow’s designer intended the first time to feel.

The scope of this strategy is every surface, flow, and state a product has, with particular attention to the states teams forget: loading, error, empty, and partial. A product can have a beautiful “happy path” ideal state and still feel cheap, because most of a user’s actual time in a non-trivial product is spent in the other four states, and those are exactly the ones most often left undesigned.

Why it works

The aesthetic–usability effect is the empirical backbone of this strategy: people rate more attractive, more considered interfaces as more usable, and they forgive small rough edges in a product that otherwise feels cared for. This is not an excuse to decorate over broken tasks—refinement that raises perceived quality without raising task success is theatre—but it does mean that consistent, intentional feedback compounds into trust in a way that is disproportionate to the engineering effort it costs.

The peak–end rule explains why refinement should be targeted rather than uniform: people do not remember an experience as an average of every moment, they remember its most intense point and how it ended. This means a single unrefined wait, error, or completion moment can dominate someone’s overall impression of an otherwise well-built product, while polishing a low-stakes screen nobody lingers on returns very little. Refinement work is most valuable exactly where Micro Interactions, Loading Feedback, and Success Moments intersect with a flow’s actual peaks and endings—not spread evenly across the product like a coat of varnish.

Core Tools, Techniques, and Practices (TTPs)

• Micro Interactions: Small responses that make an action feel acknowledged, understandable, and complete
• Small Quirk: A small, repeatable flourish that makes a clear product interaction feel recognisably yours
• Loading Feedback: Make unavoidable waits legible, calm, and proportionate
• Success Moments: Mark meaningful progress with feedback that feels earned, then let users decide what comes next
• Perceived Effort Delay: When work takes time, show the real work; when it does not, respect the user’s time

Supporting Tools, Techniques, and Practices (TTPs)

• Pattern Alignment: Use familiar mental models where they help people act with confidence
• JTBD Copywriting: Write from the progress the user is trying to make, not the feature your product wants to describe
• Variable Reward: Add optional novelty around reliable value, without making users chase uncertainty
• Progressive Disclosure: Reveal the next useful layer of complexity when the user needs it, while keeping deeper capability easy to find
• Product Voice: One recognisable voice, with a tone that adapts to the user’s moment—celebratory at wins, plain and calm at failures

Insights & Metrics

Measure whether polish made work clearer and calmer—not whether people lingered longer.

1. Task Success Rate
   • Formula: (Successful task completions ÷ Total task attempts) × 100
   • What it measures: Whether people can finish the jobs the product claims to support
   • Why it matters: Refinement that does not raise success is decoration; frustration still wins
2. Median Time on Task
   • Formula: Median time from task start to successful completion (same task definition over time)
   • What it measures: Efficiency after feedback, copy, and disclosure changes
   • Why it matters: Good polish usually removes hesitation and rework; longer sessions are not a win
3. Error Rate
   • Formula: (User errors or failed actions ÷ Total relevant actions) × 100
   • What it measures: Clarity of affordances, states, and recovery
   • Why it matters: Spikes name moments that feel broken even when the happy path looks fine
4. Task Ease (Single Ease Question / SEQ) or Perceived Usability (System Usability Scale / SUS)
   • Formula: Post-task Single Ease Question (1–7), or post-session System Usability Scale score
   • What it measures: How usable the experience felt, not only whether it completed
   • Why it matters: Aesthetic polish can mask struggle in comments—pair perception with success and errors
5. Wait Abandon Rate
   • Formula: (Sessions that leave during a defined wait state ÷ Sessions that entered that wait) × 100
   • What it measures: Whether loading and “working” feedback keep people oriented
   • Why it matters: Opaque or theatrical waits become negative peaks people remember

Behind the Data

Find where the product feels clunky, inconsistent, or unintentionally loud.

Usability & recovery

• Which tasks have the lowest success rates—and what do people do just before they fail or retry?
• How do success and time on task differ by device, locale, or entry path?
• Which errors repeat most—and are they slips, unclear labels, or missing system status?

Feedback & perceived performance

• Where do waits feel longest relative to real duration—and does Loading Feedback match the work?
• After a Success Moment, do people continue with confidence, stall, or undo?
• Are Micro Interactions clarifying state, or adding motion that delays the next useful action?

Character without clutter

• Which Small Quirk or Variable Reward moments are noticed—and which are ignored or muted?
• Where does Pattern Alignment fail (people hunt for a control that “should” be elsewhere)?
• Does Progressive Disclosure hide power users need, or still dump complexity too early?

Further reading

• Aesthetic–Usability Effect (concept) and Peak–End Rule (concept) — Handbook working definitions before the primary sources.
• Microinteractions in User Experience (NN/g) — Feedback as system status, error prevention, and brand—not ornament.
• Response Times: The 3 Important Limits (NN/g) — Enduring thresholds for instantaneous, flow-preserving, and attention-holding waits.
• Beyond the Net Promoter Score: Measuring Perceived Usability (NN/g) — SUS and Single Ease Question as durable complements to success and error rates.

Agent skill

• Primary command / sequence: /productfeeling sequence — audit → states → delight → tone → critique (experience-refinement playbook; aliases: polish, craft pass)
• Core TTPs to load: Micro Interactions, Small Quirk, Loading Feedback, Success Moments, Perceived Effort Delay
• Supporting TTPs: Pattern Alignment, JTBD Copywriting, Variable Reward, Progressive Disclosure, Product Voice
• When the agent should use this strategy: “feels cheap”, “polish pass”, “microinteraction audit”, “loading states”, “delight without clutter”, “wait feels broken”
• Companion handoff: Impeccable — primary execution partner for motion, feedback, loading, and finish craft (polish / shape with intensity constraints from the brief); DocSlime — when refinement standards or quality bars belong in docs/DESIGN.md or docs/experience/
• Feeling north star this strategy serves: calm competence with earned character and legible waits
• Anti-goals: fake delays, animation that blocks the job, novelty loops that demand chase, delight that masks failed tasks
• Reference path: skill/reference/sequence.md