Peter is from the MIT Center for Information Systems Research (CISR). He is an Australian and is looking for more Australian input to the research.
The six decisions for IT were:
- How much to spend?
- What on? Which business processes.
- Should it be organisation wide?
- How good need they be?
- How much security and privacy is needed?
- Who gets the blame when it goes wrong?
Peter said that is research had shown that 70% of lines of code in company financial systems was to integrate desperate systems. He argued good companies having systems unique to business units but built on firm wide infrastructure.
Peter asked why organisation shave large number of "stovepipe" separate systems. One rational reason for this is that the priority for a business unit is to connect to its outside customers and suppliers, not other business units. I would argue that firms should look to use industry standards to integrate outside the business and, as a bonus, achieve internal integration. This is the benefit of traditional standards development, new Open Source software, web and XML based standards.
This was a thought provoking talk and I liked the emphasis on quantitative data from research into real firms, rather than the usual made-up theory from consulting companies. But there was so much data on some slides, they were hard to read (this seems to a problem at MIT, as Philip Long had similar unreadable slides the previous week). Pyramid shaped diagrams are obligatory in management presentations, but Peter set a record with thirteen on one slide.
An electronic version of "Six IT decisions IT people shouldn't make alone" is available from Amazon along with books by Peter Weill and on project management in general.
In his introduction the DG of the National Archives of Australia also gave a plug for the ANU course "Systems Approach to Management of Government Information" <http://www.tomw.net.au/blog/2007/06/fixing-electronic-record-keeping-in.html>.