Political

Romney’s polling day technology meltdown: Orca

The usual post-electoral defeat search for explanations and people to blame has an added edge for the Republicans after Mitt Romney’s defeat earlier this month. Not only did Romney lose, he lost in all the states that were picked as being in serious contention, the Republicans actually lost ground in the Senate (when they had hopes of making gains) and the initial voting analysis shows the Republicans with a big problem: the parts of the electorate that are growing are the parts which vote against them the most heavily.

Add to all that the number of Republicans who seemed to genuinely believe the rhetoric about how the polls were skewed in Obama’s favour, and so were genuinely shocked when the defeats started rolling in, and it’s no surprise just about anyone and anything save for tricycles have so far been lined up for blame by someone.

One justifiable target is the computerised polling day system which the Republicans used, or rather tried to use as in fact it was a horrible failure. The Obama campaign had a similar problem in 2008 with Project Houdini, one of the mistakes its campaign made first time round. (A small digression: Karin Robinson wrote up a good critique at the time of my views on what Obama got wrong in 2008 and my list of positive things to learn from 2008 Obama campaign still seems a pretty good list. Liz Williams blogged it here.)

ArsTechnica reports of Orca:

It was supposed to be a “killer app,” but a system deployed to volunteers by Mitt Romney’s presidential campaign may have done more harm to Romney’s chances on Election Day – largely because of a failure to follow basic best practices for IT projects…

The goal was to put a mobile application in the hands of 37,000 volunteers in swing states, who would station themselves at the polls and track the arrival of known Romney supporters. The information would be monitored by more than 800 volunteers back at Romney’s Boston Garden campaign headquarters via a Web-based management console, and it would be used to push out more calls throughout the day to pro-Romney voters who hadn’t yet shown up at the polls. A backup voice response system would allow local poll volunteers to call in information from the field if they couldn’t access the Web.

But Orca turned out to be toothless, thanks to a series of deployment blunders and network and system failures. While the system was stress-tested using automated testing tools, users received little or no advance training on the system. Crucially, there was no dry run to test how Orca would perform over the public Internet…

In a final training call on November 3, field volunteers were told to expect “packets” shortly containing the information they needed to use Orca. Those packets, which showed up in some volunteers’ e-mail inboxes as late as November 5, turned out to be PDF files—huge PDF files which contained instructions on how to use the app and voter rolls for the voting precincts each volunteer would be working. After discovering the PDFs in his e-mail inbox at 10:00 PM on Election Eve, Ekdahl said that “I sat down and cursed, as I would have to print 60+ pages of instructions and voter rolls on my home printer. They expected 75 to 80-year old veteran volunteers to print out 60+ pages on their home computers? The night before election day?”…

Some field volunteers couldn’t even report to their posts, because the campaign hadn’t told them they first needed to pick up poll watcher credentials from one of Romney’s local “victory centers.” Others couldn’t connect to the Orca site because they entered the URL for the site without the https:// prefix; instead of being redirected to the secure site, they were confronted with a blank page, Ekdahl said.

And for many of those who managed to get to their polling places and who called up the website on their phones, there was another, insurmountable hurdle—their passwords didn’t work and attempts to reset passwords through the site also failed…

The full piece on Orca is well worth a read, as is the account of one would-be user of the system:

The end result was that 30,000+ of the most active and fired-up volunteers were wandering around confused and frustrated when they could have been doing anything else to help. Like driving people to the polls, phone-banking, walking door-to-door, etc. We lost by fairly small margins in Florida, Virginia, Ohio and Colorado. If this had worked could it have closed the gap? I sure hope not for my sanity’s sake.

A reassuring subtext for users of the Liberal Democrat Connect system is that the sorts of mistakes the Romney campaign made haven’t been made with Connect. It’s been rolled out well ahead of the biggest election day in the 4-5 year cycle, getting used first in council by-elections and then in an important round of local elections – important but not a national election polling day with its much higher numbers of voters. Similarly, on issues such as training the Connect approach has been very different, with extensive training available, including online videos.

That said, the most important lesson is always worth remembering – effective IT systems used by thousands of volunteers require much more to be got right than simply the lines of code tested in the office.

One response to “Romney’s polling day technology meltdown: Orca”

Leave a Reply

Your email address will not be published. Required fields are marked *

All comments and data you submit with them will be handled in line with the privacy and moderation policies.