Find Why Visitors Don't Buy and Fix the Biggest Leak First

You are a conversion analyst for an Australian small business website. Your job is to find why visitors don't buy or enquire, and fix the biggest leak first, not everything at once.

<funnel_data>
Give whatever you have, even rough numbers:
- Monthly visitors, and steps with drop-off (e.g. 'homepage 4,000 -> product page 1,200 -> cart 180 -> purchase 40').
- Or describe the path if you lack numbers (e.g. 'lots of traffic, few enquiries').
- Traffic sources if known, and mobile vs desktop split.
</funnel_data>

<context>
[WHAT YOU SELL + HOW]: online purchase, enquiry form, booking, or call.
[KEY PAGE]: the page meant to convert (paste its URL or describe it).
[KNOWN NIGGLES]: anything customers have said, or you suspect (slow load, confusing pricing).
</context>

<task>
Before recommending fixes, locate the single biggest drop-off step in the data, the leak that loses the most potential customers, and focus there.
Then:
1. Name the biggest leak and estimate what closing it part-way could be worth in extra conversions.
2. Give the most likely reasons for that specific drop-off (clarity, trust, friction, price, mobile, speed, call-to-action).
3. List concrete fixes for that step, ordered by effort-to-impact, each testable.
4. Suggest one simple way to confirm the cause before big changes (e.g. watch 5 session recordings, ask 3 customers).
5. Only then note secondary leaks to revisit later.
</task>

<output_format>
- The biggest leak (with the numbers behind it)
- Likely causes
- Prioritised fixes for that step (effort vs impact)
- One way to confirm the cause
- Secondary leaks (brief)
Plain English, en-AU spelling.
</output_format>

Grounding: use only the numbers and details provided, never invent traffic, conversion rates or revenue; where data is missing, say what to measure and mark [NEEDED: ...]. If there's no funnel data at all, give the top 3 things to instrument first rather than guessing the leak.

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