The Analytics Desk
Skills · Subagent · Command · PDF
.claude/the-analytics-desk/zip
skills/06
├ analytics-audit
├ consent-and-data-map
├ dashboard-spec
├ data-dictionary
├ metric-review
├ tracking-plan
agents/01
├ analytics-config-reader
commands/01
├ /dashboard
FIELD-GUIDE.pdf9pp
English
One payment · No account · No subscription
- ✓ Download link in your inbox in seconds
- ✓ Secure card checkout · VAT handled
- ✓ 14-day money back, no questions
- ✓ Every future revision included
- ✓ Runs in Claude Code and claude.ai
Built for Claude first, our recommended agent. Skills follow the open Agent Skills standard (SKILL.md), so they also run in any agent that reads it; subagents and slash commands are Claude Code formats.
The Analytics Desk
Decide what gets measured, before it gets built
six skills·one subagent·one command·9-page guide
What gets measured, decided before anything is built: the tracking plan written from real questions, every field defined with an owner, and no tile that serves no decision.
Executive summary
- The problem
- Dashboards get built before anyone asks what decision each tile serves, so a ratio that improves when you spend less sits next to a traffic line that falls for good reasons. Once a tile exists it is defended, and the wrong one steers the budget long after anyone remembers why it was added.
- What it does
- tracking-plan derives every event from a named business question, metric-review puts each proposed KPI through seven tests, dashboard-spec writes down the decision behind every tile, and the config-reader subagent reports what actually fires.
- Why take it
- Measurement is cheap to decide and expensive to undo. Deciding first costs a document; deciding later costs a rebuild and the decisions made on the wrong numbers in the meantime.
For analytics leads and the marketer who is asked for a dashboard by Friday.
In the first week you will
- 01Run /dashboard on the next dashboard request and return the spec before a tile is drawn
- 02Put the KPI somebody proposed through metric-review and get primary, guardrail, diagnostic, rejected or blocked
- 03Send your tag manager export to the analytics-config-reader subagent and see what fires, what is duplicated and what is orphaned
Contents of the Kit
- 01Six installable agent skills: tracking-plan, data-dictionary, dashboard-spec, metric-review, analytics-audit, consent-and-data-map
- 02The analytics-config-reader subagent: reads a tag or event configuration export and reports what fires where, what is orphaned and what is duplicated. Read-only
- 03A /dashboard slash command that opens a dashboard specification from one sentence
- 04The vanity test as a scorecard: seven questions a proposed metric answers before it earns a tile, with the decision it serves named and the guardrail it needs
- 05Blank templates for the tracking plan and the dashboard specification
- 06A 9-page Field Guide on specifying measurement before it is built, and plugin manifest included
One skill, as it ships
SKILL.md
tracking-plan
- Use when
- Use when a product, feature, funnel or campaign is about to be instrumented, when an existing event list has grown without a plan, when someone asks "what should we track", or before any developer ticket for analytics is written. Refuses to add an event that no question needs.
- Requires
- The product, flow or surface being instrumented · The questions the business needs answered · Owner per area
- Will not
- No event without a question · Never invent property values, expected volumes, conversion rates or thresholds · Never rename an existing event silently
- Returns
- Tracking plan: {surface or flow} · v{n} · {date}
One of six skills in the kit · All of them install the same way
The contents, in full · six skills, one subagent, one command
- 01tracking-planEvents, properties, owners, derived from the questions the business needs answered. No event enters the plan without a named question behind it, names follow one convention, properties carry types and allowed values, and every event has a surface, a trigger and an owner.
- 02data-dictionaryEvery field in the warehouse or the analytics tool with its definition, grain, unit, calculation and owner. A field nobody can define is recorded as undefined, and the dictionary emits a handoff block the Reporting Engine's kpi-dictionary reads, so collection and reporting use one set of names.
- 03dashboard-specOne dashboard written out before it is built: the audience by name, the decisions that audience makes and when, the question behind every tile, the source field for every number, and the action each reading triggers. Metrics that serve no decision move to an appendix with the reason, and the spec carries an owner and a retirement date.
- 04metric-reviewA proposed metric against seven tests: can it fall and would falling be bad, what decision changes and at what threshold, is it a ratio that improves when you do less, how cheaply can it be gamed, is there a causal mechanism, does it depend on an attribution model, can it be computed from fields that exist today. The verdict is primary, guardrail, diagnostic, rejected or blocked.
- 05analytics-auditYour configuration export in, a verdict per business question out: answerable, answerable with a caveat, or not answerable, each with the exact reason (event missing, property missing, grain too coarse, retention shorter than the period, consent excludes the segment). Defects are cited to an object in the export, never inferred from silence.
- 06consent-and-data-mapWhat is collected, which consent category gates it, what happens when consent is refused, which systems receive it, how long it is kept and who owns it, as you report it. Every gap is written as a question for a qualified adviser: the skill maps, it does not rule on legality.
- 07analytics-config-reader (subagent)Reads a tag manager container export, an event configuration or a property settings dump and reports what fires, where, on what trigger, what is duplicated and what is orphaned. It reads and reports; it changes nothing and fills no gap with a guess.
- 08/dashboardOpens a dashboard specification from one sentence about who it is for and what they decide. It names the audience as a role or a meeting, and stops to ask when the request says "the business" or "stakeholders".
Think Like a CMO calls marketing return on investment a broken speedometer: an efficiency ratio that improves when you take a slice off your own top line. The End of the Click adds that sessions and pageviews now fall for reasons that are good for the business. Both arguments land on a job almost nobody schedules, deciding what gets measured before anything is built, and the Analytics Desk is that job as tools. The kit stops where attribution begins: incrementality is a separate discipline, and putting it in the same box as a tracking plan is how measurement projects stall.
Pairs with
Think Like a CMO explains which gauges earn their place on the dashboard. The End of the Click explains why the traffic numbers that once stood in for intent now mislead. The Reporting Engine takes it from there: the Analytics Desk decides what is measured, the Reporting Engine reports it.
Before you buy
01Do I need Claude Code?
No. Every skill uploads to claude.ai as a zip. Subagents and slash commands are the Claude Code extras.
02I am not a programmer.
Install is copying folders into a .claude directory. The Field Guide in each kit assumes nothing beyond that.
03Will it work in my agent?
Skills follow the open Agent Skills standard (SKILL.md), so any agent that reads it can run them. Built for Claude first, our recommended agent.
04Do the tools invent numbers?
Never. Every figure is supplied by you, confirmed by you, or labeled as an assumption in the output.
05How does it arrive?
Instantly. Pay by card and the download link is in your inbox within seconds, and on the confirmation screen.
06Is this a subscription?
No. One payment, lifetime access to your download, every future revision of what you bought included.
07What if it is not useful to me?
Write to info@conclusionmp.com within 14 days of purchase and you get a full refund. No questions, no forms.
Ready when you are
The Analytics Desk, $12.
Instant download, one payment, every future revision included. Or take The Retail Shelf, four items for $29, and this kit works out at $7.25.