Monthly Executive Summary

Before you start, ask me about each [input] below — one question at a time, with ready-made answers to pick from; skip whatever you can already tell.

Write the monthly summary of advertising results for [client: if not listed, call roasr_list_clients and let me pick], for a reader who wants the outcome, not the mechanics.

1. **Load the RoASr method.** Call roasr_load_skill('roasr_frame'), roasr_load_skill('report_format') and roasr_load_skill('profit_math') and follow them exactly.
2. **Frame the data.** Call roasr_limits, roasr_list_clients and roasr_list_accounts: plan, niche, margin, KPI, account currencies, sync status and dataSince. Month: [month: last full calendar month (default) | current month to date | custom]. Audience: [audience: business owner (default) | marketing lead | custom]. Convert the month into exact since and until dates and take the previous period of the same kind as the comparison. If the two periods have different lengths, compare per-day averages and say so. If a window starts before dataSince, say so and narrow it. If an account has sync.state failed or never, name it and draw no conclusions from it. If currency.mixed is true, never sum money across accounts: report per account.
3. **Pull the totals.** roasr_metrics_summary for the month and for the previous period: spend, clicks, conversions, revenue, ROAS, CPA.
4. **Find what drove the change.** roasr_metrics_campaigns (campaign level, limit 50) for both periods, then match campaigns by identity. Name the 2-3 campaigns that most improved the result and the 2-3 that most dragged it down, by contribution to the change and not by name. Under campaign-level budgets treat ad set numbers as allocation, not as a ranking.
5. **Check the outcome is real.** roasr_metrics_actions for both periods: which bucket is the main result (purchase for ecommerce, lead or registration for lead generation, decided from the buckets), and whether it moved with total conversions. Buckets are not additive. A fall in conversions with stable clicks is a possible tracking issue: say so before blaming the ads.
6. **Pick the five numbers.** Ecommerce: spend, purchases, revenue, ROAS, profit if the client has a margin (revenue x margin - spend), otherwise CPA. Lead generation: spend, leads, cost per lead, change against the KPI target if set, and the number that matters most to this client. Show each number for the month, for the previous period and the change in percent.
7. **Decide what comes next.** Offer three decisions for next month. Each has: the decision, why the data points to it, the expected effect as a range with its assumption, and what is needed from the reader. Budget moves are phrased as "we recommend moving about N% of spend", never as done.
8. **Write for the reader.** For a business owner use plain words: no CTR, CPM or impression counts; explain ROAS once as "revenue per 1 spent". For a marketing lead you may add CTR, CPC and frequency. Where the data is thin or incomplete, say so in one line at the end.

**Output:** open with one sentence a busy person can forward: the result of the month and the direction, in numbers. Then the table of five numbers: indicator · this month · last month · change. Then "What worked" (2-3 items) and "What did not" (2-3 items) with a number each. Then "Three decisions for next month" in High / Medium / Nice-to-have order. End with one line of caveats from the frame.

**Output rules:**
- Recommendations only — never apply changes; RoASr tools are read-only.
- Profit appears only when the client has a margin; otherwise ROAS and CPA, with one line saying so.
- Every claim carries a number; no adjectives without data.
- Do not sum money across accounts with different currencies.
- Answer in the language of this prompt.