The ERP we decided not to implement

The situation

A PE-backed distributor wanted a faster monthly close, cleaner data, and financial reporting good enough for its board and new owners. By the time I joined, it had already bought an enterprise ERP and was partway through implementation.

The problem

The five-year license had been contracted at $1.25M. Implementation was originally estimated at $600K; the custom-code path would have brought that estimate to at least $1.2M. License plus revised implementation put the projected five-year path at $2.45M or more. That is a projection, not the amount the company paid.

Even that path would have left extensive workarounds and a heavily customized system.

What I found

The software modeled the company’s commercial relationships as simple two-party contracts. The actual arrangements involved multiple parties with different rights and responsibilities, which the ERP couldn’t represent cleanly.

The implementation wasn’t the real problem. The project was implementing the wrong model of the business.

The decision

The question changed from “How do we finish implementing this?” to “Should we be implementing it at all?” I recommended that the CEO stop, and the project was canceled.

I then designed an alternative around QuickBooks, existing operational systems, cleaned and normalized exports, a governed data layer, and analytics.

The result

The CEO canceled the implementation after about $350K had been paid. I designed an alternative using systems the company already had, with full implementation estimated at about $50K. That was roughly 2% of the projected $2.45M or more for the five-year ERP license-and-implementation path. The five-year license had already been contracted.

These are projected costs and a design estimate, not a claim of realized savings.

Test whether a system can represent how the business actually works before committing more money to close its gaps.