I work in pharmacogenomics. We're building a prototype QC tool. It's a prototype in name only - it will be functionally complete and accurate. It will not, however, be used in any clinical trials. This is because it is not being built to GLP standards. GLP (Good Laboratory Practices) is part of a suite of government mandated organizational and operational practices to be used in areas viewed as critical to the public health and welfare, such as pharmaceutical research and development or aviation.

One of the key aspects to GLP is accountability. Every decision made, every action taken is recorded in a database. This leads to bloat and over documentation, but it does provide for a rigourous and well-understood procedure.

The invention of the spreadsheet is the single most important factor in the transformation of our corporations from being near-sighted to blind - unable to look beyond the six month return. Spreadsheets, when given the correct information, can pinpoint exactly which parts of your organization are being profitable and when - as quickly and as often as you update the data. The power of this tool is that it grants a lot of weight to any argument made in favour of shutting down non-profit centres because the model does not show either the intangible or the long term benefits of the department in question, but with the right data and the right model, the instant feedback it provides is invaluable.

How do these 2 things relate to my somewhat mysterious title? Well, I believe that our already bloated and corrupt governements could, in fact, be made to run leaner and more efficiently if we created a Good Governement Practices standard, and included a requirement to have a validated spreadsheet model as part of any new project.

Since it would be a reusable sandard, it could (somewhat counter-intuitively, I know) actually reduce the start-up cost and time of new projects. Since it would have a validated model with appropriate hooks into the database driving a spreadsheet, we could have daily reports if necessary to track progress.

The huge caveats to this endeavour are:
How do we create a flexible GGP tool that won't tank and fail 50% of the time like major ERP/CRM platforms do?
How do we decide on a valid model that can provide cost/benefit for intangibles and allows a long-term view?

I leave that as an excercise for the reader.

No comments: