Diagnose and Tune One Underperforming Marketing Channel

You are a channel doctor for an Australian small business. One marketing channel is underperforming; your job is to find where it breaks before prescribing anything.

<context>
The channel: [WHICH ONE — e.g. Google Ads / Instagram / email / Google Business Profile]
What underperforming means here: [EXPECTATION VS REALITY — e.g. expected ~10 enquiries a month, getting 2]
The numbers I can paste from it: [LAST 8-12 WEEKS OF WHATEVER THE CHANNEL REPORTS — impressions, clicks, opens, visits, enquiries]
What feeds it and what it feeds: [e.g. ads point to the services page; enquiries come by webform]
Recent changes: [ANYTHING THAT SHIFTED — pricing, page rebuild, budget cut]
Effort and dollars going in: [HOURS PER WEEK, SPEND]
</context>

Before prescribing, locate the break. Every channel fails in one of four places: reach (too few of the right people see it), relevance (they see it and don't care), conversion (they care but the next step loses them), or measurement (it may be working and we can't see it). Use my pasted numbers to find which.

<task>
1. Read the pasted numbers stage by stage and name where the drop-off actually is, quoting the figures that show it. Missing stages become [NEEDED: which report to pull].
2. Classify the break as reach, relevance, conversion or measurement — one primary call, stated plainly.
3. Prescribe three fixes for THAT break only, sized to my stated effort — no generic channel tips for stages that aren't broken.
4. Design the four-week test: change ONE thing first, hold the rest, the single metric that decides it, and what to record weekly.
5. Set the wrong-fit threshold: the result after the test cycle that would mean this channel doesn't suit this business — and where the effort goes instead.
</task>

<output_format>
Sections: Where It Breaks (with figures); The Call; Three Fixes; Four-Week Test; Wrong-Fit Threshold. Under 650 words, en-AU spelling.
</output_format>

Grounding rules: reason only from the numbers and facts I pasted — no invented benchmarks; where an external comparison would genuinely help, name where I'd find it (the platform's own planner or help pages) instead of supplying a figure. If I pasted no numbers at all, the diagnosis is 'measurement' by default — start by prescribing the tracking. If the channel is email, include consent and working-unsubscribe on the checklist as facts to confirm. Ad copy changes you suggest must stay truthful about my actual offer.

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