Latest · Interpreted · Published 11 Sep 2026 · GA Applications editorial
A dashboard shows the truth. Start by having one.
A dashboard does not create clarity; it displays whatever is underneath, including the confusion. If a metric has no written definition, if source records disagree, if nobody owns updates, a dashboard will present that chaos in prettier colours. The readiness profiler below decides the honest question first: is your data ready to be shown, or should you not build yet?
The readiness criteria are GA’s working standard from dashboard scoping; the cut-offs adjust per operation.
Recognition
The spreadsheet is not the problem
Every business has the spreadsheet: heroic, personal, load-bearing. The request “turn it into a dashboard” usually means something deeper — make the numbers trustworthy and visible to people who are not the spreadsheet’s author. That is a data problem before it is a design problem.
Visualisation is the last twenty percent of the work. The first eighty: define the metric, stabilise the record, assign the owner. A dashboard built on undefined metrics becomes a daily argument with nicer fonts, and arguments held in front of a chart erode trust in every number the business has.
None of this is an argument against dashboards. It is an argument for sequence. GA’s dashboard work starts with a named decision for exactly this reason; this article publishes the readiness check that comes before any build conversation.
Score before you build
The data-readiness profiler
Tick what is already true. This profiles the data, not the tooling — tool choice comes last.
0 of 8 complete
Seven or eight ticks: build. Four to six: fix the records first — a manual weekly review may be the right next step. Fewer than four: do not build yet; see below.
The do-not-build-yet state
Under four ticks, a dashboard project fails in a predictable way: disputed numbers, then distrust, then disuse, at full project cost. The correct intervention is smaller and cheaper — pick one metric, write its definition, stabilise one source record, review it by hand for a month. That is not a failed dashboard project. It is the first stage of a successful one, and any provider who skips it is selling you the argument, not the answer.
Sequence
The honest path from spreadsheet to dashboard
Four stages. Each is cheap, each is testable, and each makes the next one possible.
Define one metric
Write the definition in plain language: what counts, what does not, where it comes from, who owns it. Test it on two people; if they interpret it differently, it is not written yet.
Stabilise the record
One place of entry, one format, validation at capture. The spreadsheet can stay — the chaos goes.
Review by hand for a month
A manual weekly check against the written definition. This surfaces the edge cases any build must handle, at zero build cost.
Then build the smallest dashboard
One decision, one threshold, one owner, one view. Expansion is earned by use, not planned in advance.
Scope
What a good first dashboard looks like
One named decision (“do we have capacity next week?”), the number that answers it, the threshold that triggers action, and the owner who acts. Everything else is a later version.
The test is brutal and simple: if you cannot name the decision, you are not buying a dashboard — you are buying a screensaver. If you can name it, the smallest honest version is usually days of work, not months, and it teaches you what version two should be.
Evidence honesty
Sources, method and what would change this
- Inference
Readiness criteria
The profiler’s eight criteria and tick thresholds are GA’s working standard from dashboard scoping, not a validated instrument. Adjust the cut-offs to your operation.
GA Applications method
- Inference
Failure sequence
The “disputed numbers, distrust, disuse” sequence is GA practitioner pattern recognition; frequencies are deliberately unquantified.
GA Applications editorial
- Verified fact
Decision-first principle
The rule that a dashboard names its decision, threshold and owner before any chart reflects GA’s documented dashboard service method.
GA Applications service documentation
- Proposal
What would change this conclusion
If sub-four-tick builds in GA’s client work succeeded without record repair, the cut-off would be revised and the change dated here.
GA Applications editorial
Bring the spreadsheet
Send the heroic spreadsheet, or describe it. GA profiles its readiness and tells you honestly: build, repair the records, or run a manual review first.
“Do not build yet” is a legitimate outcome, and the cheapest one you will ever be offered.