Thursday, December 13, 2018

What it is really like to work on a campaign: making the technology work

Just about every volunteer who had experience on an election campaign during our recent work in Nevada made some version of the same remark: "It's so great that door knocking no longer means juggling paper lists!"

They were so right. I put in a couple decades working on elections when organizing a canvass meant dividing up reams of paper, training people to mark their paper lists carefully, recapturing all that paper from volunteers, and setting up the grubby, smeared returns for attempted bar code scanning if not manual transcription. Today's VAN (Voter Action Network) database coupled with the MiniVAN canvassing app for phones and tablets has completely changed the door knocking experience, entirely for the better.

And yet, and yet -- the dirty secret of the volunteer operation I worked in last fall was that fully one third of the results that came back with earnest, eager canvassers were so tangled or incomplete that they did not provide useful information about interactions with voters.

Volunteers had to learn three operations to use the MiniVAN.
  • 1) They had to sign into the program and download the list (turf) they were assigned to work on. This was a purely rote activity: there was only one right sequence of necessary actions and once we learned how to limit volunteers' access to any byways that would get them off course, they could usually accomplish this, at least once. (They had to learn the sequence just in case the app tossed them out once they were in the field -- a rare but not unheard of glitch.)
  • 2) They had to record each actual attempt to reach a targeted voter at a door by marking "not home" (the most frequent result), "deceased," "moved," etc.
  • 3) If they actually contacted the targeted (listed) voter, they had to mark whether the person supported our candidate strongly, less strongly, was undecided, or was for her opponent. This was the core information the canvass was seeking; the campaign aimed make sure supporters voted.
All this had to happen concurrently with walking around unfamiliar streets, trying to find the correct houses to knock at, and, when finding a voter, being prepared to persuade her to vote for our candidate. But those activities weren't about the MiniVAN app.

The most common reason that volunteers' data came back as gibberish was that they failed to perform either step 2) above (not recording attempted door knocks including failed ones) or step 3) above (having conversations but not recording the preference the voter revealed.) Day after day, a considerable fraction of well meaning volunteers would come back, try to tally up their results, and either give up or realize something was very wrong and try to reconstruct what had happened, not very accurately.

I believe the prime responsibility of a volunteer program on a campaign is not to waste volunteer's time. And after all, we did actually want the informative data that volunteers might be able to develop through all this walking about. You can win elections by using good data. We tore our hair, tweaked the training, tried different ways to emphasize what operations were involved -- and never really overcame whatever was keeping so many of our folks from succeeding with the MiniVAN. They were endlessly willing and not at all dumb, but we couldn't make the system work better for them or for the campaign.
Shortly after we won on November 6, having helped elect Democrat Jacky Rosen to the US Senate, I got around to reading Atul Gawande's reflections on why digitization that he believes in is failing in hospitals he knows so well. Medicine delights in what computerized records can offer -- and the process is failing some of its best practitioners. (I've explored Dr. Gawande's wise insights frequently on this blog.) He writes:

Something’s gone terribly wrong. Doctors are among the most technology-avid people in society; computerization has simplified tasks in many industries. Yet somehow we’ve reached a point where people in the medical profession actively, viscerally, volubly hate their computers.

... Artisanship has been throttled, and so has our professional capacity to identify and solve problems through ground-level experimentation. Why can’t our work systems be like our smartphones—flexible, easy, customizable? The answer is that the two systems have different purposes. Consumer technology is all about letting me be me. Technology for complex enterprises is about helping groups do what the members cannot easily do by themselves—work in co√∂rdination. Our individual activities have to mesh with everyone else’s. What we want and don’t have, however, is a system that accommodates both mutation and selection.

This rings so true to me based on my experience with getting volunteers up to speed on the MiniVAN. We were trying to make a pretty simple technology work in tandem with what are complex human interactions when door knocking results in conversations that persuade. Maybe (some) people have great difficulty switching back and forth between the necessarily routinized task of recording a limited range of information and the demanding art of talking with human beings. Those people would be the ones who we never seemed to be able to help get the technology to work.

And soon enough, perhaps both the apps and the humans will be changed enough by our ongoing interplay in all of contemporary life so that the particular difficulties of 2018 will become quaint. Like Gawande, I don't know if that evolution will make us zombies or superhumans. Guess we'll have to wait and see to find out.

1 comment:

Celia said...

I agree with Gawande's observations. The clinic I use now provides each practician with a scribe who handles the data input (most I've met are medical assistant trained) while my doctor or the nurse talks and examines me. Getting a scribe for a canvasser seems unlikely though. Thanks for your hard work though.