In several startups and projects, I’ve seen a curious tension between innovation and adoption — and nowhere more than TrustTheVote’s development of open election technology. Today’s specific example comes from a question I received recently from someone we met at GoGaRuCo: what are we doing about more fair voting counting algorithms, esp. approval voting, and using Web technology to help with ballot usability issues, and several other useful innovations that would be valuable changes in the U.S. election system? These and other innovations seemed to my correspondent to be valuable improvements, and in theory I had to agree. But in fact we do not have a long list of method-innovations on our tech roadmap for voting systems, much less the development work we’re doing now. Why?
Here’s the curious thing about innovation and adoption: we have and continue to talk to a lot of election officials — the people who will be making decisions about the adoption of TrustTheVote technology — and for voting systems, innovation is really not top of mind. We ask what they want in a new generation of election technology, and what would motivate adoption. Their answers almost uniformly diverge along the same lines of voting systems (caution on innovation) vs. election technology more broadly (enthusiasm for innovation … a story for another day). For voting systems, the message on must-haves is very clear: deliver voting system technology that correctly implements our current election methods, reliably, open, worthy of trust, with smaller scope for human error, and reduce the workload on our aging shrinking base of volunteer poll workers. Actually I could go on from there, but the point is that innovation in this case is not a driver to adoption, but (if delivered unwisely) a potential hindrance to adoption. But the reverse is also true: innovation in simplicity or transparency can provide some would-be-nice nudges toward adoption, if the must-haves are clearly met.
I’ve seen plenty of technology start-ups begin with a story interesting to potential early adopters, but with this gotcha: for the customer go beyond proof of concept and gain substantial value, they have to change part of their existing IT practices. And in most cases, they were very very interested, but not actually motivated to adopt, regardless of innovative technology and potential value. That’s what voting systems are like, but times 100. Any innovation is clearly in the would-be-nice category at best, because what election tech adopters are hungry for is to be consulted while we’re building a new system, to ensure that it actually fits what they do today, and makes their existing job easier. The current vendors have to varying degrees delivered voting systems that they thought their customers should like, and weren’t interested in making big changes when their customers found the products hard to use and required changes to the way they run elections.
And you know what? most election technology innovation has been done the same way, to date. Many people have tried their hand at open source election technology (recognizing the trust value of openness) with interesting innovations (which would address some real defects of current election methods), and the same attitude of “I know what should be valuable, and if these government folks could see the value, they will use my stuff.”
But at TTV, we’re doing it the other way, boring as it may be, optimizing on adoption to get some improvement in voting technology actually fielded and at work openly to start re-building trust in our computerized election systems. And believe me, I’d like TTV to deliver all sorts of cool stuff as well, like alternative voting schemes and counting algorithms, and other innovations for election officials to see, touch, and try, and decide for themselves whether it’s promising enough to put some effort into using. And an open, public technology base should really help, too. But for the near term, we’re focusing mainly on only one innovation in voting systems: a system that election officials and voters could use and actually like.