Weekly metrics commentary that separates signal from noise
Writes the narrative for a weekly or monthly metrics report from a table of current and prior values plus team context: what moved, whether it matters, the likely reasons you know about and what to watch.
Prompt in Data by AISkillHub Library
What it does
Metrics reviews often read as a list of every number that went up or down. This prompt writes commentary that helps leaders focus: it identifies the few movements that matter, compares them with normal variation where you supply history, links them to known events from your context notes, and is explicit when the cause is unknown.
**Use it when** you send a recurring metrics update (weekly business review, monthly board metrics, support or marketing performance) and want a first draft of the written commentary.
**What you get** - A three-bullet headline summary. - For each notable metric: what changed, how it compares with recent history and target, known contributing factors from your notes, and whether action is needed. - A list of metrics that moved within their normal range and need no discussion. - Questions to investigate, where the cause is not known. - Items to watch next period.
**What it will not do** - It will not invent explanations. Causes are attributed only to events in your context notes, and otherwise the movement is listed as unexplained with suggested checks. - It will not call a change significant without history to compare against. If you give only two periods, it says variation cannot be judged. - It will not recalculate your source numbers. It works from the table you paste, and flags any values that look inconsistent (for example, a total that does not match its parts).
**Example** Given 8 weeks of history for trial signups that usually vary by about 10% week to week, a 6% dip is placed in the "within normal range" list, while a 35% jump in support tickets that coincides with a release noted in the context is called out with the release as a probable factor and a question on whether the ticket tags confirm it.
How to use it
## How to use
1. Open `prompts/weekly-metrics-commentary.md` and copy everything below the title line. 2. Replace each `{{variable}}` with your own content. Paste long material (notes, transcripts, tables) directly in place of the variable; keep the surrounding tags so the model can tell inputs apart. 3. Start a new chat in ChatGPT, Claude or Gemini and paste the filled prompt as your first message. 4. Read the output against the quality checks at the end of the prompt before you use or send it.
## Variables
- `{{metrics_table}}`: a table with metric name, current value, prior value and, ideally, several earlier periods and the target. - `{{period}}`: e.g. `week ending 25 Oct 2026`. - `{{context_notes}}`: known events: launches, outages, campaigns, holidays, pricing changes, data issues. - `{{audience}}`: who reads it, e.g. `exec team`, `board`.
If you do not have a value for an optional variable, write `none` rather than deleting it, so the prompt still reads correctly.
## Reusing it
To reuse it often, save the prompt (with the variables left in) as the instructions of a Claude Project, a custom GPT or a Gemini Gem, then paste only the inputs in each new chat.
Files
Save each file at the path shown, relative to your project root.