Write a Dashboard Spec Before You Build It
Drafts a build-ready dashboard spec with KPIs, layout, filters, drill-downs, and out-of-scope guardrails.
Dashboards fail in the spec, not the build. An analyst gets 'make me a sales dashboard,' guesses at the KPIs, picks chart types nobody questioned, and three revisions later the stakeholder still cannot answer their actual question. This prompt forces the hard thinking up front: the single decision the dashboard supports, the exact KPI definitions, the layout, the filters, and, just as important, what is explicitly out of scope so it does not balloon into a do-everything report that answers nothing. Tying every KPI to a source table and grain prevents the late-stage discovery that two tiles compute 'revenue' differently. Writing the spec before opening Tableau or Looker means review happens on a one-page document instead of after days of building. Use it at the kickoff of any dashboard request, or to rescue one that has sprawled, the out-of-scope section alone often saves a week. Treat the open-questions list as the agenda for your stakeholder check-in.
You are a BI lead. Write a clear dashboard specification so an analyst can build it without guessing. Context: [DASHBOARD PURPOSE]. Primary audience and how they will use it: [AUDIENCE AND DECISIONS]. Data sources: [DATA SOURCES]. Refresh cadence: [REFRESH]. Produce a spec with: (1) the single key question the dashboard answers, (2) 3-6 KPIs with their exact definitions and targets, (3) a layout outline top-to-bottom with the chart type and filters for each section, (4) required slicers and default filter state, (5) drill-down paths, and (6) explicit out-of-scope items to prevent scope creep. For each KPI note the source table and grain. Keep it build-ready and opinionated about what to cut. List open questions for stakeholders at the end.
What you can expect back
Key question: Which stores are below target this week and by how much? KPIs (source / grain): Sales vs Target % = sales / weekly_target (sales fact + targets / per store-week); Week-over-Week Sales change (sales fact / store-week); Stores Below Target count (derived / region-week); Avg Basket Size (sales fact / per transaction). Layout top-to-bottom: (1) scorecard row, four KPI tiles with red/green vs target. (2) Ranked horizontal bar of stores by Sales vs Target %, worst at top. (3) Line chart of regional sales vs target over last 12 weeks. Section filter: region. Slicers: region (default = manager's region), week (default = latest complete week). Drill-down: click a store bar to a store-level view with daily sales and basket trend. Out of scope: individual employee performance, product-level mix, real-time intraday data, forecasting. Open questions: Is 'this week' the latest complete week or week-to-date? Are targets fixed or revised mid-week? Should closed stores be hidden or shown as zero?
Illustrative example — your results will vary by tool and inputs.
Get sharper results
- 01Insist on one key question; a dashboard that answers five questions usually answers none of them well.
- 02Tie every KPI to a source table and grain in the spec, it surfaces definition conflicts before anyone builds.
- 03Write the out-of-scope list aggressively; it is the cheapest scope-creep insurance you will ever buy.
- 04Set default filter states, a dashboard that opens on 'all regions, all time' wastes the user's first ten seconds.
Adapt it for your case
Ask for an ASCII or grid wireframe of the layout to align on placement before building.
Ask it to expand the KPI list into full definitions with inclusion rules for your data catalog.
Append your tool, e.g. 'outline how to build each section in Looker or Power BI'.
Common questions
Why spec it before building?
Reviewing a one-page document is far cheaper than rebuilding after days of work; most dashboard rework comes from skipping this step.
How many KPIs should I include?
Three to six focused on the key question; beyond that the dashboard becomes a data dump and the signal drowns.
What's the point of the out-of-scope list?
It gives you a documented, agreed boundary to point to when new requests arrive, which is how dashboards avoid sprawling into unusable everything-reports.
You may also need
Choose the Right Chart Type for Your Data
Recommends the best and runner-up chart type for your message, variables, and audience with encoding guidance.
Interpret an A/B Test Result Without Overclaiming
Interprets A/B test numbers honestly, flags validity risks, and gives a ship or keep-running recommendation.
Write a Precise, Unambiguous Metric Definition
Produces a precise metric definition with formula, grain, inclusion rules, and resolved edge cases.
Explain a Statistical Concept to Non-Technical Stakeholders
Translates a statistical concept into a plain-language explanation tied to a real decision for non-technical stakeholders.