Disclosure: my experience with developing computer systems is limited and outdated. And as for actually doing the systems work, including the software programming, I haven’t done any of that for decades. As a manager, I was fortunate to have wonderful, talented people for more recent though still ancient projects, without whom nothing useful would have been accomplished.
Just so you know, I did have some success, but the systems I worked on were not by any measure the size and scope of the website for Obamacare. Examples of systems I worked on: payroll, student accounting, appropriation accounting, process control, datacenter programs, court systems, law firm systems.
However, despite my being outdated, there are still a few of what we mathematicians call axioms, basic ways to manage and perform systems development that, I submit, never change. Here are a few:
First, the lowest bidder is usually the worst choice.
Second, if you bid the job out (request for proposals – “RFP”), you are likely to have a lot of headaches and dissatisfied people no matter what company you choose.
Third, despite the above, for any complex development project you need to get bids for at least two reasons: 1) choosing bidders and preparing the specifications for the system will get you involved in important details and 2) if you select the bidders well, even in the bidding process they will contribute to your information systems education.
Fourth, get the best people to do the job, even if they cost more.
Fifth, make sure those people don’t go off on their own to do whatever they feel like doing; i.e., define what you want accomplished and manage, manage, manage.
Sixth, do not hide behind a lack of knowledge about computers and computer systems; be involved, take some classes (especially on project management), get your hands dirty, manage, manage, manage.
Seventh, do not launch an entire system all at once. Instead, set up an appropriate test or “pilot” group or groups and get the system working. Get some of the bugs out before it is inflicted on an entire office, company, or nation.
There are other axioms. Pick up any book on information systems management for more.
Did I mention you must manage the project?
I recommend Steven Brill’s article in the March 10, 2014 issue of TIME magazine: “Code Red – inside the nightmare launch of HealthCare.gov and the team that figured out how to fix it.”
After you read the article, do you think that if our leaders in Washington had followed the above suggestions/axioms, the launch might have had fewer and less serious problems?
And this from a firm but rapidly aging and disappointed “Dem” – we needed our leaders to do a much better job on the rollout of HealthCare.gov.