The Manager
by
Weekly business review for ecommerce teams: reconcile the week into one set of figures, separate a real change from noise, and write five to seven ranked actions with an owner against each.
The weekly review usually spends its first half agreeing on numbers. Growth has one revenue figure, finance has another, and by the time everyone accepts a version there are ten minutes left for decisions. What gets assigned in those ten minutes rarely survives to the following week, because nothing wrote it down against an owner. This skill reconciles the week's revenue, margin, spend, cash and retention into one set of figures, compares each against the prior week and the plan, separates a real move from normal variance, traces the material ones to a driver, and writes five to seven ranked actions with an owner and a date on each.
What It Does
- Pulls the week's revenue, margin, spend, cash and retention into one set of figures.
- Compares each against the prior week, the plan and the same week last year.
- Separates a real change from normal week-to-week variance.
- Traces every material move to the channel, cohort or SKU behind it.
- Writes five to seven actions, ranked, each with an owner and a due date.
- Pushes the actions into Notion so the next review opens on what was assigned.
Use Cases
- Open the weekly meeting on decisions rather than on reconciling four dashboards.
- Give every function one action list instead of a shared doc nobody owns.
- Catch a margin slip in week two rather than at month end.
- Keep a written record of what was assigned and what actually moved.
Example Session
Run this week's review and give me the actions per owner.
Revenue held, margin did not. Three things moved enough to act on this week.
- Metric | This week | Prior week | vs Plan | Driver
- Net revenue | $412k | $408k | Minus 2% | No single driver
- Contribution margin | 31.4% | 34.1% | Minus 3.6pt | Shipping and returns
- Blended CAC | $46 | $41 | Plus 9% | Meta prospecting
- Repeat rate | 28% | 28% | On plan | Flat
The margin drop is not a pricing problem, it is fulfillment: expedited shipping on two late purchase orders plus a return spike concentrated in one SKU. Actions this week: Supply reviews both POs before Thursday, CX pulls the return reasons on that SKU, Growth holds Meta prospecting at last week's budget until blended CAC comes back under $43.
Why This Beats Prompting It Yourself
Four Dashboards vs One Set Of Figures: The Hidden Risk
A generic prompt summarizes whatever you paste into it. Paste four exports and you get a readout that averages four definitions of revenue and presents the result with total confidence. Worse, everything that moved is reported as though it mattered, so a 2% wobble in a noisy metric gets the same airtime as a three point margin drop. Nothing in the prompt distinguishes a signal from a week.
When You Still Need An Owner On Each Line
A review is only worth the actions that leave it. This skill ends on five to seven of them, ranked, each with a name and a date, written where the next review will open. It also says which moves were noise, so the meeting spends its time on the margin drop rather than on a revenue figure that did not really change. The figures underneath are one set, reconciled once.
Known Limitations
It reviews what is connected, so a cost line living only in a spreadsheet will not appear in the margin figure and the output says which inputs were missing. Variance thresholds are tuned to your own history, which means a business under six months old will see more moves flagged as real than it should. The actions it writes are proposals: nothing is assigned until a human in the meeting accepts them.
Install It
- Download the ZIP.
It is free and there is no account to create. - Unzip it into your agent's skills folder.
Claude Code reads ~/.claude/skills/, which is hidden by default: the command in the folder block opens it. Other agents scan their own directory, so drop the same folder there instead. - Ask your agent to use it.
Restart the agent if it was already running, then it picks the skill up with no config.
skills folder
Copy
~/.claude/skills/the-manager/
SKILL.md
references/variance-thresholds.md
reviews/
# one file per week, with the actions assigned
# macOS: create the folder and open it in Finder
mkdir -p ~/.claude/skills && open ~/.claude/skills
# Windows: paste in the Explorer address bar
%USERPROFILE%\.claude\skills\Frequently asked questions
How is this different from a dashboard summary?
A summary tells you what the numbers were. This reconciles them to one definition, separates the moves that are real from the ones that are noise, traces each real move to its driver, and ends on ranked actions with owners. The output is a meeting agenda, not a readout.
Which agents does the skill run in?
Any agent that supports the open SKILL.md format: Claude Code, Cursor, Codex CLI, GitHub Copilot, Gemini CLI, Manus, Grok and others load it unmodified. The format is portable, the location is not. Each agent scans its own skills directory, so you drop the same folder into whichever one yours uses. There is no config file to edit and no API key to provision.
What does it need connected to work?
Revenue and orders, ad spend, product and shipping cost, and a cash source. The more of those are connected, the fewer gaps in the margin line. It runs on a partial set and names the inputs it did not have rather than quietly working around them.
How does it decide what is noise?
By comparing this week's move against your own historical variance for that metric, not against a fixed percentage. A 2% revenue move is noise for most businesses and a real event for a few, and the threshold comes from your data rather than from a rule of thumb.
Does it assign the actions itself?
It proposes them with an owner and a date and writes them to Notion. A person in the meeting accepts or changes them. Anything that commits budget or headcount is flagged for explicit approval rather than written as a decision already taken.
What is included with the skill?
The skill itself, the variance threshold reference, the review template, and the Notion action format. It is free and security scanned. We re-publish the ZIP when the review template changes, so download it again if your weekly file stops matching the format.
An exec review is only as good as the numbers everyone agrees on.
Polar reconciles 45+ sources into one semantic layer, so the five to seven actions you assign start from a single set of figures.
<script type="application/ld+json">{"@context":"https://schema.org","@type":"FAQPage","mainEntity":[{"@type":"Question","name":"How is this different from a dashboard summary?","acceptedAnswer":{"@type":"Answer","text":"A summary tells you what the numbers were. This reconciles them to one definition, separates the moves that are real from the ones that are noise, traces each real move to its driver, and ends on ranked actions with owners. The output is a meeting agenda, not a readout."}},{"@type":"Question","name":"Which agents does the skill run in?","acceptedAnswer":{"@type":"Answer","text":"Any agent that supports the open SKILL.md format: Claude Code, Cursor, Codex CLI, GitHub Copilot, Gemini CLI, Manus, Grok and others load it unmodified. The format is portable, the location is not. Each agent scans its own skills directory, so you drop the same folder into whichever one yours uses. There is no config file to edit and no API key to provision."}},{"@type":"Question","name":"What does it need connected to work?","acceptedAnswer":{"@type":"Answer","text":"Revenue and orders, ad spend, product and shipping cost, and a cash source. The more of those are connected, the fewer gaps in the margin line. It runs on a partial set and names the inputs it did not have rather than quietly working around them."}},{"@type":"Question","name":"How does it decide what is noise?","acceptedAnswer":{"@type":"Answer","text":"By comparing this week's move against your own historical variance for that metric, not against a fixed percentage. A 2% revenue move is noise for most businesses and a real event for a few, and the threshold comes from your data rather than from a rule of thumb."}},{"@type":"Question","name":"Does it assign the actions itself?","acceptedAnswer":{"@type":"Answer","text":"It proposes them with an owner and a date and writes them to Notion. A person in the meeting accepts or changes them. Anything that commits budget or headcount is flagged for explicit approval rather than written as a decision already taken."}},{"@type":"Question","name":"What is included with the skill?","acceptedAnswer":{"@type":"Answer","text":"The skill itself, the variance threshold reference, the review template, and the Notion action format. It is free and security scanned. We re-publish the ZIP when the review template changes, so download it again if your weekly file stops matching the format."}}]}</script>