Progamming for MBAs

Business by on November 22, 2009 at 10:35 pm

My first job out of MIT was as a consultant for Bain & Company. Bain, like other strategy consulting firms was fueled almost exclusively by MBAs and recent college grads.

Excel was used extensively for casework. New consultants received several levels of Excel training and nearly everyone developed Excel expertise. SPSS made an occasional appearance in casework, but Excel was the tool of choice when the data set could be manipulated in Excel.

Super Models
The models and analysis conducted using Excel ranged from simple cash flow analysis through to analysis of data exported from ERP systems and models of thousands of chain store financial statements. Consultants that were unable to pick up Excel rarely went far in the company.

In many ways, skilled consultants used Excel as a basic programming tool.

1,048,576 >> 65,536
I rapidly became an Excel-geek, and I’ve carried my Excel-fu with me throughout my career and have used it to solve numerous problems. I elected to upgrade to Office 2007 solely because of the 1M row limit. I still find Excel quicker than MySQL and/or Python for a number of one-off analysis tasks. Data is far easier to explore, reformat and summarize in Excel.

Now add clusters
Even today, I regularly run tasks that max both of my CPUs in Excel. I have a good sense where the limits lie and avoid them. Microsoft has been testing Excel 2010 running on a cluster of machines. Nearly all the analysis that I do with Excel is exploratory in nature, and Excel provides an excellent solution for the vast majority of these problems. Microsoft is definitely mapping out a path towards the next step which truly excites me:

And then the cloud
Doing all of this in the cloud would truly be interesting. I and nearly all MBAs have no interest in running local HPC clusters. It gets really interesting when you can run a version of Excel that has access to the computational resources in the cloud. I’d happily pay computational fees to use Excel to process vastly larger amounts of data.

I’d also expect that much of the analysis conducted in the financial community that uses MBAs (and some PhDs) for algorithm development will likely remain in Excel rather than going through an implementation stage with a development team.

Google is of course on the way to doing something similar. For now they have targeted the casual user, but I’m sure that will change as Apps develops.

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 2.5 License. | Dave Naffziger's BlogDave & Iva Naffziger