Social Icons

twitter google plus rss feed

Pages

Friday, November 09, 2012

The day a Narwhal killed an Orca


The tale of an Iowa Narwhal hunter:

Election Day didn't begin well.

I got up at 5 a.m., showered, dressed, and headed to the bus stop. Upon arriving, I called DART (our local transit authority) to be sure I was getting on the right bus. As I feared, it turned out that the polling place I was assigned to was an impractical distance from any bus stop.

This is, after all, Des Moines.

Moreover, I was given the wrong information about which bus to take, and not told that the bus I needed to transfer to was an "on-call" bus (a kind of quasi-cab intended, sometimes on request and sometimes by running a regular but infrequent route, to supplement the bus system and enable commuters to reach destinations far off-route).

Fortunately, the lady who answered the phone was also wrong about what time the on-call bus would arrive. As a result, I got to the polling place only half an hour late.

But there were two poling places in the church where mine was located, and I initially went to the wrong one. A shame; the ladies working that precinct were sort of fun, and it would have been pleasant to have worked there even if there hadn't been several couches and easy chairs for me to have spent the day sitting in. But the people at the other precinct were nice, too. And as it turned out, my late arrival had no major consequences.

I said earlier that I'd explain after the election about Project Orca. Essentially it was a high-tech election day operation intended to overcome the "ground game" advantage of the Democrats nationally (the Democrats' Election Day operation was called "Project Narwhal," and the orca is the narwhal's natural predator).

When I was pounding the pavement back in Chicago, you had to go door to door with a precinct list, marking supporters with a "+," undecideds with a "0" and supporters of your opponent with a "-."  On Election Day, you would sit there at the polling place with a ruler and a pencil, crossing off "plus voters" as they voted. At various times during the day, "runners" would physically go to the homes of "plus voters" who hadn't voted to deliver a reminder slip and if possible personally encourage supporters to vote. Sometimes drivers were employed to give supporters a "lift" to the polls.

The polls closed at 6 p.m. in Chicago back then. At 5:30- sometimes earlier, if turnout had been light- even the "checker" left his post and "ran." The goal was at all costs to get as many of your supporters to the polls before they closed as possible.

Sometimes phone canvassing was used in place of door knocking, but the principle was the same: identify your voters, and at almost any cost get them to the polls.

Now, the science- and it is a science- of marketing has become so advanced that by studying a voter's demographics, it's possible to predict who he or she will vote for with 93% accuracy (yes, I find that scary, too). While canvassing is still done, especially in areas in which you expect many supporters to live, you have something to fall back on in the areas you don't get to.

Project Orca was a high-tech effort to simplify and revolutionize the whole election day process. It issued Romney poll watchers rosters of identified Romney supporters in their precincts. Next to each voter was a pin number and two boxes. One was checked when the voter voted. Every hour or so, you would go through the roster, check the second box, and either check off the names of those who had voted on your smart phone or tablet make a list of the pin numbers of supporters who had voted in the last hour. This was sent by smart phone'tablet ap or by texting to Mitt Romney's national headquarters in Boston. Poll watchers were also given phone numbers to Orca help hotlines, including one for legal help.

Once the numbers arrived in Boston, the theory was that a computer would decode them and remove the voters they represented from an electronic master-list at each local headquarters. There, phone banks would be in effect "running" voters who had not yet made it to the polls throughout the day on the basis of real-time information.

In theory, on Election Day we would be using using tactical nukes against Narwhal's ground troops. They would never know what hit them. In practice- as first tries of technologically complicated systems usually are- Orca was a disaster.

First, the people like me, who didn't have smart phones, had trouble all day logging in to Orca. Sometimes we couldn't log in at all, and Boston would tell us to just go back and check in voters while they worked on it. Eventually the lists of pin numbers we had not yet called in got so large that we fell hopelessly behind.

In the local phone banks, the callers were told repeatedly that the people they were calling had already voted. Part of the problem was the break-down in the system. But there was another, more basic flaw:the process of checking voters in, going over the roster and noting their pin numbers, making lists of their pin numbers, and texting those numbers to Boston was so labor-intensive and so time-consuming that it's doubtful whether Orca would have worked even if the technical glitches hadn't happened. By the time a voter's pin number reached the local headquarters, he would already have been contacted by the phone banks. Moreover, some poll watchers couldn' be there all day, which rendered the whole process rather pointless. Others weren't eventrained, and didn't have a clue about how Orca was supposed to work; the guy who gave me a lift from the polls to the Embassy Suites for the wake thought he had a bad list because nobody had told him that only Romney supporters were listed. He just sat there all day, doing nothing!

There was another problem with the Orca system, also a pretty fundamental one. If you made a mistake entering a pin number, there was no way to go back except to void the whole list and start over again. This was obviously impractical for reasons of time- and as the adage goes, "Garbage in, garbage out."

The only break all day from the bizarre combination of craziness and monotony was when a woman came into the polls wearing an Obama-Biden button, and I pointed out that that constituted electioneering. The Democratic poll watcher vehemently denied this, and we had words (Iowa election judges, like Iowa people generally, are nice, and don't like confrontation. It turned out that in the twenty years the Democratic gentleman had worked the polls, he had simply never seen a voter challenged for wearing buttons or campaign apparel in the polling place, and so he decided that it was legal).

It turned out that- amazingly- he was right, at least in 2012. The Iowa Secretary of State- who is, oddly, a Republican- had issued an opinion that for a voter (as opposed to a poll watcher) to wear such campaign paraphernalia was OK. "We don't like it," the Romney legal team told me over the phone, "but there it is." So we dropped the issue. The lady had taken the button off when I challenged her, and it didn't happen again. In the last analysis, it was no big deal (although I was warned at one point by headquarters that they had reports of people from the Democratic party inside the polling place openly encouraging people to vote a straight Democratic ticket, I personally saw no sign of such shenanigans).

At about 8:30- half an hour before the polls closed- I hit the wall. Numbers and names were swimming before my eyes- I'd been at this for over 14 hours- and I was having real problems doing even the simplest part of the process. At 8:45- the moment, as it turned out, the phone banks closed; it was too late to get anybody else to the polls anyway- I gave up.

I caught my ride in front of the church and headed for the "victory party" at the Embassy Suites. Ominously, the driver- the guy who hadn't been trained- didn't have the election returns on the radio. He was understandably disgusted, and understandably pessimistic.

Hanging up my coat in the lobby, I entered the ballroom. One glance at the TV screens- Fox News on one and a clearly delighted Scott Pelly on the other- sent me to the cash bar.

We were leading in Virginia, but narrowly. Ominously, we were actually losing in Florida, where the polls had showed us comfortably ahead. Even worse, we were barely ahead in North Carolina, where the Obama campaign had supposedly given up.

Ohio was close- but didn't look close enough. The general tone of the folks on Fox News, as well as the emotional weather of the room, matched my own foreboding. I finished my scotch on the rocks and chatted morosely with a few fellow volunteers. Then I sighed, got my coat, and walked out of the hotel about twenty minutes after arriving.

CNN called Ohio and the election for Obama just as I was leaving the convenience store across the street from my apartment building with my much-belated dinner. I'm afraid I didn't have much appetite.

As I said earlier, I'll post anon about where I think we go from here (warning: it's not going to be upbeat). Suffice it for now to say that I have never been as discouraged in the wake of an election in my 56 years of them.

ADDENDUM: Things were even worse in Colorado. And here's John Eklund's experience with Orca; being a techie, he's in a better position than I to comment on what went wrong.

Maybe, on further reflection, the Narwhal should be acquitted of the charge of murdering the Orca. Maybe the coroner's verdict should be death by natural causes- namely, congenital damage to its internal organs incompatible with life.

Or simply being born without a brain.

I kind of wish I hadn't read Eklund's account, because the numbers haunt me as much as they haunt him. The margins were so close that the failure of Orca might well have cost us Ohio, Virginia, Florida, Colorado- and the election.

ADDENDUM II:  The more I read about the Orca debacle, the more I think Eklund's fears are apt. It may very well have cost Romney the election. For more on the scale of the disaster, see this at National Review Online.

2 comments:

Harry Coin said...

Scott County (easternmost) Iowa R HQ wrote to all poll watchers at 4pm that we would not be getting voter lists and probably would not be getting any ORCA app either. The app came to us at 9pm on election eve, but didn't work all day across the county though we tried every hour. Some of us printed the 30+ pages that came in the email at 9pm, but even the phone method hung up on us.

The ORCA project proved one thing: Local equipment, local software is a must. Reports can go national, but too many races lost because of this, not just Romney to ever be risked a second time.

Robert E. Waters said...

Unbelievable.

It sounds self-serving now, but even a non-techie such as yours truly thought more than once in the runup about the things that can go wrong when an Election Day operation relies so heavily on untested tech, and on the stakes. Murphy was vindicated once again.

I entirely concur. The time to start building a national infrastructure of local equipment and software is now. Like it or not, something like Orca is going to be the wave of the future. But it may well have cost us the presidency (and, as you say, many state and local races as well), and we can never risk it happening a second time.

Reading between the lines, I have a strong hunch that the internal Romney polls said that we were going to win; hence, Mitt's being blindsided as he was by the actual outcome. We just didn't get our voters to the polls, and even a traditional Election Day operation might well have been good enough. In my own precinct- a heavily Republican one- easily a third of those on my list didn't vote. Extrapolate that across the country, and it tells the tale.

While I doubt that we'll hear much about this in the media, the more I think about this the more I think Mr. Eklund's fears are exactly right.

I think Orca cost us the election- us, and, as you say, good people up and down the ballot all over the country.