The trouble with Excel…. It’s just too bloody good!
We do a lot of work across the finance and insurance sectors and at every turn there are critical business processes that are built on Excel and running on desktop operating systems. We recently came across an Excel spreadsheet with 6500 lines of VBA code doing some very sophisticated work. That’s no longer an Excel sheet, it’s a business application.
Modellers, Risk Analysts and Actuaries alike simply love Excel. Some of those professionals are looking at or working with alternate solutions like MatLab, R, F#, C++ etc. Even those that are looking at alternates simply love to prototype and model in Excel.
So there it is…it’s a great tool that average and sophisticated users love and will continue to use industry wide for all sorts of business applications and processes.
But hear is the rub….
The solution to Excel batch processing
If you want to turn these sophisticated Excel applications into something that runs server side and can be run in batch mode and is fit for sharing across a group of users, you have a problem.
We have seen companies try all sorts from creating PC based “grids” with desktop compute resource being scavenged from the office floor to server racks stuffed with Windows desktop VM’s or large SharePoint farms running Excel services. Generally, these solutions are all flawed.
The other thing we see is that solutions like ClusterSeven are used to deal with the symptoms rather than the cause by providing governance, controls and audits for these business processes. But the answer to the root cause is to make Excel run in a mature sever side compute engine using server side technologies and operating systems and able to benefit from the mature development processes that exists today to safeguard these processes.