Note: Since publishing this piece, it has zoomed around various activist lists and generated a lot of controversy. In particular, the well-respected social scientists of the Analyst Institute showed me research proving that self-reporting of voting was absolutely unreliable — a fact that I suspected and noted, but didn’t have the data to prove. And Becky Bond of CREDO showed me a study that found peer-to-peer GOTV appeals to have zero effect on turnout.
In short, this makes my hypothesis nothing more than a hypothesis. Provocative, I still hope, but not gospel. It may be a hypothesis worth further study, but it’s certainly not worth supplanting traditional, proven GOTV tactics with personalized emails. So let me be extra clear in what I wrote below: My point isn’t to abandon traditional GOTV tactics. My point is to add a new tool to the toolbox. Campaign managers should ask their volunteer willing to phone bank or knock doors also to send a personal email endorsement to their lists. Key word being also, not instead of.
I trust no one saw my post and tore up all their phone banking lists. If you did, better get back to reprinting those lists!
The original post is below. Note that I changed the headline, to make sure no one misconstrues it.
As election day 2010 draws near, political campaigners are gearing up their GOTV plans. In the coming weeks, volunteers will be asked to phone bank, knock doors, pass out literature, plant lawn signs, and perhaps even stand on street corners waving banners at passing cars.
I’ve got a new idea — let them send emails.
Sound boring? Or ineffective? In my experience, it’s not. In a recent totally un-scientific analysis, a couple of friends and I found that sending out GOTV emails to our personal lists was by far the most efficient and effective use of our time.
The election in question was the Sept 14 New York primary. On Sept 13, I sent out an email to 225 New Yorkers, reminding them of the election and asking them to vote for Eric Schneiderman for NY’s Attorney General. My appeal was short and sweet — four brief paragraphs, starting with, “Just a friendly reminder that tomorrow, Tuesday, is Election Day in New York, and you should go vote!” I followed that with an exhortation to vote Schneiderman.
I didn’t expect much — perhaps one or two courteous replies, and a few more “don’t spam me!” complaints. Instead, I was pleasantly surprised by the response. Six people — Laurie, Bob, Patty, Danika, Julia and Matt — individually emailed me to say that they had voted for Schneiderman because of my email. Patty also told me she forwarded it to her entire New York list. Not a single person complained.
Inspired by this “turnout,” I contacted two friends who also sent out similar personal endorsement emails on Sept 13, and I asked for their “response rates.” Their results confirmed mine:
• Andrew sent his email to about 400 New Yorkers, “a mix of friends and cultural / business / political acquaintances.” He got five responses saying that they had voted on his recommendation; and one person who acknowledged forwarding the email to her New York list.
• Elana sent her personal endorsement email to 75 New Yorkers, and posted it to her Facebook page. She got 16 responses saying that they had voted on her recommendation. Her email was also forwarded to a high-profile magazine editor, who blogged and tweeted it.
Between the three of us, our emails reached 700 New Yorkers and generated 27 votes (or at least, people who reported voting because of it). If the numbers are accurate — a big if — hat’s a 3.8% conversion rate — a tremendous success for any GOTV activity.
In terms of votes per hour, it was an efficient use of time.
• My email took about 90 minutes; 15 minutes to write the email, 45 minutes to collect addresses from my address book, and 30 more to send (in batches, so spam filters wouldn’t nail it). I earned six votes.
Rate: 4 votes / hour.
• Andrew spent two hours on his email. He earned five votes.
Rate: 2.5 votes / hour.
• Elana spent four hours on her email, most of it researching down-ballot candidates. She earned 16 votes.
Rate: 4 votes / hour.
Averaged out, our rate was 3.6 votes / hour.
As someone who has run countless phone bank and canvass efforts, these numbers strike me as surprisingly efficient. When compared to the tactic of sending volunteers door to door, or phone banking, or lit dropping, the personal endorsement email seems like a totally viable tactic.
Not that all of those other tactics aren’t legitimate GOTV activities. They certainly are, and they should be used as part of a holistic GOTV strategy. My point isn’t to abandon traditional GOTV tactics. My point is to add a new tool to the toolbox.
In short, campaign managers should ask every volunteer willing to phone bank or knock doors also to send a personal email endorsement to their lists. And the campaigns should provide volunteers with support to make it work. Specifically, they should:
• Explain to volunteers the power and value of their personal email list
• Set aside time in the campaign office to do it, especially during periods when phone banking and canvassing aren’t effective (late night, middle of the day)
• Give volunteers template emails (plain text, plain English, low-pressure tone) so they don’t get stuck on the writing — but ask them to use their own words if they can
• Set up a “technical support” structure for folks who need help using their email clients to send out personal blasts; have your more computer-savvy volunteers be tech support for the less savvy
• Ask volunteers to report “response rates”
Of course, I fully admit that my un-scientific analysis could be inaccurate. Self-reporting vote counts are never reliable. My friends may have reported to me that I swung their vote, even if they would have voted for Schneiderman anyway. Other friends might have voted on my suggestion but not thanked me (I never asked for a report from anyone). And the impact of Patty forwarding my email to her list can’t be measured. My best guess is that these corrupting factors even each other out, and thus I did generate six(ish) votes for Schneiderman. [Note: after this was published, another friend emailed me this: "PS: By the way I just read your post and you can add me to your response rate. I am embarrassed to admit that I hadn’t really been following the primaries. Your email prompted me to make it to the polls 3 mins before closing (literally) ... and voting for SChneiderman who I voted of because of you."]
Also, I realize that this tactic won’t be appropriate under all circumstances. A high profile Presidential campaign probably doesn’t room for last minute persuasion. Very local races probably won’t attract volunteers with enough personal contacts in district to make an email effective. So, under what circumstances will this tactic work?
1) You need the right races. This tactic is probably best in a statewide, low profile campaign. Most volunteers do not have social networks confined to one State Senate or Assembly district, or even one U.S. Congressional district. So ballot initiatives, gubernatorial races, mayoral, AGs, comptrollers, US Senate primaries, and “at large” citywide races are good opportunities for this tactic.
2) You need the right volunteers. Not every person will have the same ability to reach out to a like-minded social network and activate them. Elana, Andrew and myself are all professional political or social change activists who are involved in our communities. We have more email addresses of like-minded voters than the average Joe — and more willingness to contact them. But, so will most campaign volunteers. They, too, tend to surround themselves with like-minded political friends and acquaintances, and they usually have enough commitment to contact those folks with a low-pressure email.
3) You need the right support. Navigating email programs may be difficult for less tech-savvy volunteers. If you don’t have someone who can walk a volunteer through using their BCC function, your efforts will be frustrated. But if you’re using volunteers for GOTV work at all, then you probably do — because someone is training them to phone bank, knock doors, etc.
4) You need the right sender. Meaning, volunteers must send these emails, not the campaign. Research has conclusively shown that campaigns sending GOTV emails, either to their list or to bought / rented / swapped lists, does not make one iota of turnout difference. The only thing worth trying is personal emails, authored by real live volunteers, sent out authentically to their social networks.
Variations on this tactic have been used for centuries, of course. A friend-to-friend endorsement is often seen as the most valuable coin in politics. Whether it is or not is debatable — at least one piece of research suggests it is not. Nevertheless, many campaigns and organizations have systematized this kind of contact into formal programs. Here in New York, the Working Families Party has long delivered “Friends & Neighbor” mailings to their volunteers’ lists, and they are about to launch an equivalent email tool. The League of Young Voters pioneered the concept of user-generated “ballot guides” (still very much active at theballot.org), which are intended to spread to friends and allies. And the geeks at the Democratic Party are building a ballot guide tool that looks more like a social media platform.
My suggestion is much simpler than all of that. There’s no new technology to adopt, no privacy issues to navigate, no websites to visit. Just plain old email, to plain old friends and acquaintances.
So, campaign managers, next time your volunteers are standing around waiting for the canvass to start, or hanging around after the phone banks have closed, plop them down in front of a computer and let them send emails!
posted in Campaigning & New York City & Online Organizing
I’ve got a big think question for all the organizers and social change agents out there – will we take advantage of the “new Internet,” or will it pass us all by?