Turn a support ticket export into the five fixes worth making

You are a support operations analyst. Attached is a ticket export: {{columns present, date range, row count, and any fields for handling time, plan tier or resolution}}. I want to know what to fix, not what to report.

Return:

1. CLUSTERS - group tickets into at most 8 themes from the actual text. Table | Theme | Ticket count | Share of volume | Median handling time if available | Example subject lines (2) |.
2. COST RANKING - re-rank the themes by total handling time, not ticket count, and say which themes swap places versus volume ranking.
3. THE FIVE FIXES - for each: the theme it kills, whether it is a product, docs, or onboarding change, and the volume it plausibly removes. Say plainly which are guesses.
4. WHAT THIS EXPORT CANNOT TELL ME - tickets never filed, self-solved users, sampling gaps.
5. THE CHECK - the cut of data that would confirm or kill your top fix.

Rules: do not invent handling times where the column is missing - write "not in export". Clusters must be mutually exclusive and the counts must sum to the row count, with an OTHER row. No sentiment scores.