Many teams read “data capability” as being able to spit another sheet on Friday afternoon. The more often they are asked, the stronger they look. The shared drive then stacks “final” and “final 2,” and none of them answers this week’s real question: who should handle this swing. Shipping a report is processing. Analyzing the problem is the capability.
Report stack
Problem chain
A report stack’s mark: every new question opens a new column; nobody maintains the old ones. A problem chain’s mark: the question changes, you still walk the same KPI, the same detail, the same owner. If a document will not open, you can only keep stacking sheets; see If a number cannot open a document, what is the dashboard for. Treating systems as a warehouse exhibit also tempts people to ask for “one more full extract”; see Treating systems as a warehouse exhibit vs a decision entry. Self-serve extract on an unsigned wide table lets everyone stack another private pack. Capability looks stronger; the yardstick fragments.
An operations dashboard without a problem chain only moves the stack from email to the browser. Without a dashboard, the path breaks waiting for someone to deliver a sheet; see With vs without an operations dashboard: how the decision path changes. Visible efficiency comes from fewer duplicate sheets and more closed loops; see Visible efficiency a data capability actually delivers. Why you need an operations dashboard: Why an operations dashboard is not optional. Write “how many problem chains closed this week” into the data team’s review. That is closer to capability than “how many charts we shipped.” Output orientation rewards the stack.
At acceptance, ask whether this question will need a new sheet next time
If the answer is always yes, you still delivered a report factory. Analysis capability is also not staffing a pivot-table-only role. The role must translate operations language into a definition sentence, and translate document gaps back into an operations action. A pure “pull-data clerk” locks the team back on the stack. DaXi asks to accept against a real weekly-meeting topic: pick a swing on the spot, walk to the document, write the action owner. If that path fails, you are still shipping reports. Alerts that push problems to people are the night version of the chain; see Alerts pushed to people, so exceptions rarely sit overnight. To say how many new sheets you add each week, submit on the Data service page. How engineering walks from KPI inventory to a chart you can ask: From brief to launch: how a data project actually runs.
Turn new-sheet requests into a problem ticket: what must be answered, which existing chain fails, who owns the action once it works. If that cannot be written, do not ship a sheet. Habit changes, then the stack stops.
Analysis capability is not staffing another pull-data clerk
Analysis capability is not staffing a pivot-table-only role. The role must translate operations language into a definition sentence, and translate document gaps back into an operations action. A pure “pull-data clerk” locks the team back on the stack.
Self-serve extract can exist, but it must land on signed KPIs. Opening an unsigned wide table is an invitation for everyone to stack another private pack.
Write “how many problem chains closed this week” into the data team’s review. That is closer to capability than “how many charts we shipped.” Output orientation rewards the stack. Phase one freezes three to five signed KPIs. New questions prefer the old chain, not a new column.