First, read this.

Second, I need to point out that this story is not true.  Do you hear me, internet?  NOT TRUE.  The Daily Currant is like The Onion, a satirical newspaper that pokes fun at things.

Third, even though it isn’t true, it is entirely believable, which is why so many people passed it around as true.  Partially because we’re happy to think that Bush just isn’t terribly bright, but more because the UI for voting is terrible.  Really, really terrible.

And it isn’t like we don’t know how to make good UI.  Collectively, we’ve created Facebook and Twitter and Mint and thousands of examples of truly unique interfaces that enable people to make incredibly complicated and detailed decisions.  In voting, you’re just picking one thing from a list of things.  A series of dropdowns would be better than what we have now in most voting markets.

The real fun of this story, though, is not in thinking about what better UI might look like, but rather the forces that are driving design in the world.  Overwhelmingly, design innovation comes from the private sector, the need to drive people to a decision point in an ever-increasing way.  The race to conversion, because it is the clearest metric that predicts revenue, means that companies are constantly recruiting for a combination of designers and data scientists to run every possible variation, chasing improvements as small as 0.1%.

Non-profits and the public sector just don’t have the same use for such small differences and as a consequence, we miss out on the big differences.  Voting software is built primarily by engineers concerned with security and redundancy, and because there is no concept of conversion, it could be the world’s most frustrating experience.  After all, as long as people persevere and vote for someone (regardless of whether it is who they intended to vote for), nothing changes.

It is, in a sense, the problem we have with any captive audience: as long as people don’t have a choice other than participate or don’t participate, and incremental participation isn’t considered as a success metric, then nothing changes.  Particularly because this is not software that is built in the public sphere, so things like Code For America have an uphill battle to get involved.

So what do we do about it?  For one, we need to employ some of the data science I talked about earlier.  Unless we test whether people are actually able to vote for the candidates they intended to, we have no way of knowing whether a UI is appropriate.  Indeed, intention congruence should be the new primary metric for judging voting software, followed closely by a subjective measure of the quality of the experience.  After all, if people don’t feel like voting was a good experience, then they won’t do it again.  Year-over-year participation, for example, should be considered.

And then, quite simply, we need designers who treat this problem like a commercial one.  Given the metrics, they need to design against them and have the opportunity to A/B test and refine in waves.  In a commercial atmosphere, this is standard fare, but in the public sphere, it is far less common.

I leave you with this thought experiment: how much is a vote worth?  That is, a single voter who is able to vote for the candidates of their choice effectively – how much should we as a country pay for that?  Let’s pretend you said it was worth a $1 to get an incremental voter through the experience.  ~126 million people voted in the 2012 presidential election.  That 0.1% change I mentioned early?  Its worth $126,000 each e