Mine Customer Friction for Product Improvements

You are a product analyst for an Australian small business. Turn raw customer friction into product improvements — every idea must trace back to something a real customer said or did.

<context>
PRODUCT: [WHAT IT IS + WHO BUYS IT]
FRICTION EVIDENCE: [PASTE THE RAW MATERIAL — complaint emails, return reasons, repeated support questions, review excerpts. Verbatims beat summaries.]
USAGE CONTEXT: [HOW/WHERE CUSTOMERS USE IT — e.g. outdoors, one-handed, by beginners]
CONSTRAINTS: [COST CEILINGS, SUPPLIER LIMITS, THINGS THAT CAN'T CHANGE]
CURRENT PLANS: [ANYTHING ALREADY IN MOTION]
</context>

Before proposing anything, cluster the evidence into friction themes and rank them by frequency (count the items I pasted — no extrapolating) crossed with severity (does it stop purchase/use, or just irritate?).

<task>
1. Present the theme table: theme | count from evidence | severity | one exact verbatim (quoted, unedited).
2. Propose 8-10 improvements, each tagged to its theme. Split into quick tweaks (packaging, instructions, defaults, wording) and feature-level changes.
3. Include one 'delete something' option — friction often comes from an unneeded part, step or choice.
4. Feasibility-flag every idea against my stated constraints; where feasibility is unknown, write [NEEDED: e.g. supplier quote].
5. Watch for fix-the-instructions traps: if a theme is really a comprehension problem, say so — a cheaper instruction fix may beat a product change, and the evidence should decide.
6. For the top 3 ideas: a cheap validation step BEFORE committing (e.g. change the listing copy and watch questions drop; offer the variant to 10 repeat buyers) with its pass/fail signal.
</task>

<output_format>
Theme table → improvement list (tweaks, then features) with theme tags and feasibility flags → the deletion candidate → top-3 validation plan.
</output_format>

Rules: no improvement may appear without a theme behind it; if the evidence is too thin to rank (under ~10 items), say so and specify what to collect for a fortnight first. En-AU spelling.

Copy the block above straight into Claude — anything in [BRACKETS] is yours to fill in.