Power BI, Tableau and SAP Analytics Cloud cover the overwhelming majority of enterprise reporting needs, and covering them with a packaged tool is almost always the right call: faster to build, cheaper to run, easier to hand over. The interesting question is the residual minority — the requirements that keep fighting the tool.
You recognise them by the workarounds. A dashboard that needs three exports and a spreadsheet macro before the Monday meeting. A drill-down the tool cannot express, so analysts keep a parallel workbook. An operational view that needs write-back — approve this, hold that — which BI tools were never designed to do.
That third case is the clearest signal. The moment a 'report' needs to trigger action — an approval, a hold, a reorder — you are no longer building a report; you are building a thin operational application over your ERP data. Forcing it into a BI tool produces the workaround culture; building it as a small custom dashboard produces software people actually use.
The economics are better than most IT leaders expect. A focused custom dashboard — one audience, one decision, live ERP data, exactly the drill-downs asked for — is weeks of work, not months, when the data foundations already exist. The expensive version is the one built without governed KPI definitions underneath; then every number becomes a negotiation.
Our rule of thumb: packaged BI first, always — then custom, surgically, where the requirement involves action, unusual interaction, or an audience the packaged tools price out. The mistake is not choosing one or the other; it is failing to notice when the requirement crossed the line.