Anyone who has bought computer systems over the last 30 years has come up against the same dilemma – shall we go for a ‘best of breed’ packaged solution, or shall we build it internally, using the skills of the internal IT teams? It is the classic “Make or Buy” dilemma.

You talk about your requirements with IT. They look at the problem – it looks like a few database tables with a couple of front-end screens for data entry/update will do the job. Chuck in a couple of reports and hey - the job’s done. It won’t take more than a couple of months, tops. Surely that’s better than spending more money on what looks like an expensive solution from a vendor?

Do you think that any of your critical business processes are vulnerable to a spreadsheet error? Or perhaps, has the number of models (including artificial intelligence) proliferated to the point where it’s difficult to understand your risk exposure (never mind reduce it)? If yes, there is something you can do.

Creating and maintaining a Model Inventory is simple in concept but much harder in practice. Identifying the existing models to include can be challenging, especially given the fact that the line of business aren’t usually enthusiastic about documenting their model usage. Some model owners would prefer that you just go away, thank you very much. But it’s the “maintain” task that is really difficult. Keeping a model inventory accurate and up to date takes a lot of work. From chasing down various owners and users for updates to reporting on KRIs, it is a thankless job that never ends. Model Inventories are very resource intensive but there is an easier way that is better, faster and cheaper.

Earlier this year, Peter High wrote on Forbes.com about an interview he had with Clay Johnson who currently is the CIO of Walmart. Although the topics of discussion weren’t specific to end-user computing risk management, I found Clay’s lessons learned from his early experience at FedEx to be very applicable to my work focused on Model risk and EUC risk management.