By now, everyone has heard of the problems with the app that the Iowa Democratic Party purchased to aid in getting the caucus results.

First, I don’t believe that there was anything sinister with the provider of the app.  Yes, the company that makes the app has connections to people working for several of the campaigns and the Obama Administration.  But the world of political consultants is rather small.  Political parties need data bases to track voters and contacts in the various precincts and counties; so they tend to have relationship with certain software providers who have divided that type of information.  And political campaigns need that same type of data; so they are likely to have connections to the same consultants.  Some of the features needed for keeping an up-to-date list of voters and tracking the likely Democrats in each precinct (and updating based on canvasses) would seem to be comparable to what is needed to run caucuses and get results.  In any case, the process includes a paper “back-up” that has to be verified by the local representatives of each campaign.  So, while mistakes were made, to support a conspiracy theory, you would need to imply a lot of plants in all of the campaigns in the right locations.

Second, what does seem to have happened as a common tech problem.  As a government attorney, I have seen multiple generations of case management systems.  While the programming is beyond me, successful new systems have several things in common.  After getting the basic parameters, the programmers design a program to meet the requirements.  A bunch of internal alpha testers then sit down and try to use the system.  Their comments on what works and what doesn’t work then lead to revisions designed to fix any bugs in the software and make the system more user friendly.  Then you recruit beta testers from the pool of people who will have to use the system once it goes live.  Again, updates are made based on the comments from the beta testers.  Then the system is rolled out gradually starting with some pilot counties or pilot units within the office.  This gradual rollout allows training of small groups of users and a chance to fix the system when the real world experience doesn’t match the testing.   It may take eight or nine months (or even longer for something going to every county in a large state) before the system is running in every county and every office.  While there are certainly target dates, there are no absolute deadline.  So if things go poorly in the pilot counties, you can take the time to fix the problem.  (For example, when my state went to an electronic filing system for court pleadings, it took three to four years to get to 100+ counties.  And this gradual process allowed the trainers to spend two to three months with court staff in each county and to offer multiple training opportunities to attorneys.   Since my practice at the time involved multiple counties, by the middle of the process, I was used to using the system in half of my counties and couldn’t wait for the rest of the counties to get the system.) Continue Reading...