Using the internet to redo politics, again

You know one thing that the internet is really good at?  Accessing large databases from anywhere.  That's Google, Facebook, Twitter, and just about any message board in the world.  Those are massively huge databases, and those websites are pulling content out of them on the fly.  Often people are adding content, and it's filing that into the database and feeding it to your friends when they scroll down the Facebook wall in nearly real time.

There's a database that's really important to elections: voter data.  Especially the public voter data of registered voters that they can get from election districts which has a name, an age, a political party, an address, and maybe a phone number.  When you volunteer for a campaign you generally end up doing one of two things, both of them connected to that data.  The first is make phone calls to people to ask them for money, ask them to volunteer, or ask for their vote in the election.  The second is to go out on the street and actually knock on people's doors and ask them to come out to vote, called canvassing.

Those work. They show your neighbors that the candidates have dedicated people working for them, and it really sticks the name of the candidate in a percentage of people's heads, and going house-to-house gets a certain percentage of people out to vote.

But the way that that data is handled is a disaster.  For both previous presidential campaigns I made phone calls at least once.  Both times I was handed a sheaf of paper printouts with the names of the people that I needed to call, and I called down the list and marked a box with the responses.  The given responses were usually "no answer," "refused," "willing to volunteer," and "willing to donate."

Nearly immediately I ran into a ton of problems.  The first was "That's not me, I just got this phone number."  That's not a no answer, and it's not technically a "refused to answer" either.  It might mean that the number was listed wrong, it was changed to another person, or that the person is lying and doesn't want to talk to me.  But that doesn't mean that the person couldn't be contacted through another number or by email, or that the lying person wouldn't mind getting emails.

The second was "I was just called by the campaign five minutes ago" or "I've been called five times by the campaign today" and one caller actually said "If I get another phone call tonight from you I'm not going to vote for Obama."  You know why that was happening?   Because in the database that the Obama administration was keeping listed these people as "need to be contacted" and multiple call locations would just print out the top of the list.  At the end of the day, all that data needed to be put back in the system so that people could be marked as "contacted." That meant tons of confusion about who needed to be contacted, who had been contacted, and who didn't need to be contacted again, and another round of volunteers who had to input the data.

The third was, "I'd love to volunteer or donate . . ." As the person on the phone with them though, there was no way for me to take a donation over the phone.  You're not going to be having volunteers taking credit card information, and without access to their email address, there was no way to actually follow up with that specific individual to remind them to donate when they aren't on the phone.  Once you say "go to the website" then you have to hope that they will follow through on their own.  The same is true for volunteering, where I would often check the "willing to volunteer" box on the sheet of paper, but couldn't give them any up-to-date information on ways to volunteer in their area without asking them to go to the website on their own.

A secure xml database with a every person's basic information input (no credit card information) that volunteers could access directly could have solved all of those problems.  First, it would have been much easier to change between phone contact and email contact, and incorrect numbers could be marked without necessarily deleting the person entirely off the lists (you could even email them to check if they'd moved or changed phones).  Second, marking responses could be done directly to the database to make sure that people weren't called five times in a row.  Third, a system that allowed flags for different kinds of follow up would have made the lives of the campaign directors much easier by tracking those requests.

Let's say you have that information in a database already.  Perhaps you create an iPhone app that can access one or two records at a time.  The app could allow a volunteer to contact the next person on the list with one touch, and lock out the entry for 15 minutes so that it wouldn't pop up for anyone else.  At the end of the call, the app could present questions for the volunteer with buttons as simple as "Yes" and "No" that would allow the campaign to control the information gained, such as "already contacted," "needs follow-up," or create flags and messages that later volunteers could follow up on.  That information can flow directly into the database, or it can be sent to a trusted supervisor anywhere in the country to be reviewed before being input. 

The flags could be used to create lists of people that need assistance getting to the polls, or who want to volunteer.  If the person contacted speaks Spanish, it could be tagged for follow-up by a Spanish speaking volunteer.  The system could keep track of issues that voters care about and call scripts could add specific information based on what the system knows about the political views of a particular candidate.  All of this could basically function in real time. 

With the prevalence of smart phones, a similar system could be used by canvassers going door to door, getting information to the campaign as people are contacted.  You could even have multiple rounds of canvassers hit the same neighborhoods at the same time without fear that you would knock on the same door multiple times.

And having that better data would improve the analytics of the campaign by a thousand fold. How many people need follow up?  What kinds of questions are they asking the volunteers who are calling?  How many people in a given location need rides the day of the election?  Are certain canvassers getting better responses, and if so what are they doing to be more effective?  Are certain people better at talking to people on the phones?  And you can have records of which person contacted which other person and perhaps even have the same people following up. Imagine the experience of knowing that "Jackie from the Obama campaign" is the only person who will call you, and you can ask her questions and she'll get back to you if she doesn't know the answers.  She speaks your language, and can arrange a local ride to your specific polling place for you.

Let's go back to the watch party I mentioned in the previous post.  People might be uncomfortable inviting other people over for a watch party in their homes, but what if they know that campaign has all the information on those people, and it will usually be the same people coming to events every time?  It's the same concept as Uber, you have a trusted group of people whose information is on file getting electronically matched together to create a better experience for the host and for the participants. By building a community of people with the same views, you get them more involved in the process.

From the perspective of the volunteer, this system also makes so much more sense.  You can make calls from anywhere, at any time, for just a few minutes or a few hours.  You can canvass in your own neighborhood for 30 minutes after work instead of needing to take four hour shifts.  You can even reward the volunteers who canvass and do the most calls with campaign swag that they can show off in person.  Perhaps the best caller and the best canvasser in an area can get VIP passes to a candidate event when Obama is in the area, and get to shake hands and get a picture?

Supervisors could also review new input from anywhere, even while out canvassing themselves, or even across the country.  The system would allow people in Maine to help out people in Hawaii if Hawaii volunteers got overwhelmed.  Programmers could even create different user interfaces, as long as the architecture of the database was properly structured, analogous to the way that Facebook on a desktop computer is different than on a phone, or the way that Tweetdeck allows different access to twitter than the native app does.

Finally, this kind of system could be useful even more for down ticket races.  What if Jackie called you back for the midterms to ask you to vote?  Maybe even for your mayor or city council person?  By building relationships between volunteers and voters you could drastically impact local races.  You could print up direct mailers specific to individuals to keep them informed about their specific local elections, and the issues that they care most about.

This is the system that the Democratic Party needs to build to mobilize the next generation of political supporters.  The paper and pen canvassing and random calls make the party look disorganized and out of touch with modern technology.  The party will also be moving into step with the current culture of personalized information technology and using technology to connect local volunteers together toward common local, state, and national goals.