Dayparting Opportunities

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.

Find out whether [client: if not listed, call roasr_list_clients and let me pick] has hours or weekdays where conversions are clearly cheaper or dearer, and whether that is strong enough to act on.

1. **Load the RoASr method.** Call roasr_load_skill('roasr_frame') and follow it 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. Focus: [focus: hours and weekdays (default) | hours only | weekdays only]. Weekday window: [weekday window: last 8 weeks (default) | last 4 weeks | last 12 weeks | custom]. 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: do the work per account with accountId.
3. **Pull the hours.** roasr_metrics_breakdown with dimension hourly. It is a snapshot as of the asOf date inside the window, not a sum over many days: state asOf and say how many days of data it likely represents. Do not assume a timezone. RoASr does not report one, and depending on the platform the hour can follow the account or the viewer. Show the hours as labelled and ask me to confirm the account timezone before any schedule is proposed.
4. **Pull the weekdays.** roasr_metrics_daily over the weekday window. Group days by weekday; for each weekday compute pooled CPA (total spend / total conversions) and ROAS, never an average of daily CPAs. Drop the last 3 days: late conversions are still arriving. Require at least 4 full occurrences of every weekday.
5. **Apply the noise floor.** An hour or weekday gets a verdict only with at least 30 conversions in its group. Merge neighbouring hours into blocks (for example night 0-5, morning 6-11, afternoon 12-17, evening 18-23) to gain volume, and say that you did. A gap under 25% against the overall CPA is "no real difference".
6. **Check what could fake a pattern.** One promo day, an outage or a budget change can create a weekday effect: look for outlier days and say if one day drives the result. For lead gen, hours with leads that sales never reach may still be bad; mention that quality is not visible here.
7. **Decide.** If a block is 25% or more better or worse, supported by enough conversions, it is a candidate for a schedule or bid change. Give a rough size of the effect, for example "removing the worst block would save about N% of spend for about M% of conversions". Say that campaigns with few conversions often do better without a schedule because the learning phase needs volume.

**Output:** open with one sentence: whether a real dayparting opportunity exists, and where, in numbers. Then two tables: hour blocks (block · spend share · conversions · CPA · ROAS · verdict) and weekdays (weekday · spend · conversions · CPA · ROAS · verdict). Then 3-6 actions grouped High / Medium / Nice-to-have, each with the expected effect and your confidence (high / medium / low).

**Output rules:**
- Recommendations only — never apply changes; RoASr tools are read-only.
- Never state the timezone as a fact; ask me to confirm it.
- "No real difference" is a valid answer; do not invent a schedule on small volumes.
- If the client has no margin set, talk ROAS and CPA, not profit.
- Answer in the language of this prompt.