Work out whether your free plan is paying for itself

You are a finance-literate operator reviewing a free tier nobody has questioned since launch. Below are my plan limits, free and paid user counts, conversion and cost data.

1. WHAT IT COSTS - table: Cost line (infra, support, onboarding, sales time, seats) | Monthly total | Per free user | Confidence (given / estimated / missing).
2. WHAT IT RETURNS - conversions to paid, referral or content value, with the evidence for each. Mark anything I asserted without data.
3. VERDICT - pays for itself, breaks even, or subsidises non-buyers. One paragraph, my numbers only.
4. OPTIONS - table: Option (keep, tighten limits, trial, reverse trial, kill it) | Effect on conversion | Effect on cost | Who gets angry | Reversible?
5. IF WE CHANGE IT - migration plan for existing free users: notice period, grandfathering rule, and the one message they get.
6. MEASURE FIRST - the two numbers that would settle this, and how to get them in a week.

Rules: no outside benchmarks. Do not invent numbers - write "missing" and say what to pull.

PLAN LIMITS AND PRICING: {{paste}}
USERS, CONVERSION AND COSTS: {{free/paid counts, conversion rate, infra and support costs, anything you only estimate}}