Do we really need an EPM?

Here's the quiet routine inside most finance teams I talk to.

1. They log into the EPM.
2. They run a report.
3. They export to CSV.
4. Then they build the actual model in Excel.

This happens at companies that paid six or seven figures for the EPM. It happens with platforms that market themselves as finance-owned. It happens whether the system is fifteen years old or eighteen months old.

It isn't inertia

The usual explanation is that finance teams are stuck in their ways. I don't think that's it. Finance teams adopt new tools constantly, when the tools earn it.

The real explanation is simpler.

The model is where finance thinks.

If the modelling environment isn't a place finance can actually reason in (change a driver, test a hypothesis, see the lineage) then the EPM becomes a database, and the thinking happens elsewhere. Excel is where it lands, because Excel is the last place finance still owns the model end to end.

This isn't an argument for Excel

Excel has its own well-known ceiling. The point is different: what finance actually needs is narrower than what EPM tries to be, and deeper than what EPM actually delivers.

A fixable problem

The trade-off the industry has accepted (power or ownership, scale or visibility) was never a law of nature. It was a consequence of how modelling platforms were built.

That's a fixable problem. Worth fixing.