A Product Manager's Proposal for the Unity08 Process

As a Product Manager in Silicon Valley, the three questions I ask about any problem are:

1. What do we want the customer to experience?


2. What technical innovation is necessary to achieve that?


3. What are the opposing systemic forces that need to be overcome?

For Unity08's nomination process, I believe the goal is to:

Engage a broad spectrum of American citizens
in selecting an optimal, bipartisan Presidential ticket.
through effective presentation of all relevant data

To me, the key word is "engage." If the Unity08 experiment is to succeed, we need to find a scalable way to actively engage a diverse pool of delegates in the hard work of interviewing and analyzing potential candidates; otherwise, it will become a mere popularity contest.

So, how does effectively enfranchise a large (and constantly increasing) community composed of diverse -- even contradictory -- viewpoints?

A. Self-Organizing Caucuses

We do it the way nature does it: through "spontaneous self-organization." In political terms, we enable the easy formation -- and self-policing -- of innumerable "caucuses." The goal is to create little "micro-communities" with their own governance and agenda, so that individuals who share a similar mindset can work to refine that into specific questions, positions, and policies. This gives new joiners an easy way to find a "home" in the vast see of Unity08 delegates, which in turn rewards those communities that are both welcoming and clear in their communication.

B. Nominated Questions

Having established these caucuses, we would then leverage them to nominate *questions* for the various candidates. This would need to be an iterative process, as we would want questions to be refined, merged, and split in order to best reflect the mood of the population as a whole. Once we've obtained a mostly-orthogonal set of questions, there would then be an internal poll (perhaps using Approval Voting) to pick a reasonable set that all candidates must address.

C. Remixable Answers

Here's where it gets interesting. In addition to written answers, I would encourage candidates to submit video responses of arbitrary length -- and not just to the mandatory questions, but any other questions they felt were important to them (or the caucuses). These would then be available on, e.g., YouTube, or even CSPAN.

And if the candidates were brave enough to grant appropriate rights, I would want to give Unity08 delegates permission to *remix* those videos into condensed position summaries -- though those remixes would need to be clearly labeled as such, follow certain guidelines, and only be available to registered delegates (unless approved by the original speaker). This would give the uber-wonks the opportunity to streamline and customize their candidate's message for those who lack the patience to listen to the whole thing, without necessarily reducing everything to sound bites. Importantly, any such remixes would need to include links back to the original source material, as well as to any rebuttals the original author chose to make.

The goal, ultimately, is to create the tools for a sustained, passionate, yet civil dialogue around all these matters, by using modern technology to involve ordinary Americans in the process.

D. Approval Runoff Voting

The final way to engage voters is to adopt a "March Madness" approach to voting, where the field is progressively winnowed down to 16 candidates, then 8, then 4, 2, and finally 1. This allows excitement to build (a la American Idol), and encourages delegates to learn more and more about fewer and fewer top contenders. This should also mesh well with the ongoing influx of new delegates. Each voting period should be multiple days (including a weekend), to maximize opportunities to participate.

The best mechanism for this (IMHO) is what I call Approval Runoff Voting, where at each "tier" voters can select as many candidates are there are "seats" available in the next round. For a given tier, this is the exact same process used for electing County Commissioners in most states, so it shouldn't be that unfamiliar; plus, individuals who don't care about the complexity can simply select their single favorite. However, unlike Instant-Runoff Voting (IRV), we have a better chance of a weaker "consensus" candidate surviving the initial rounds. Plus, the desire to attract multiple votes should discourage unsportsmanlike sniping among the candidates.

I believe that this approach has the best chance of providing a positive user experience for all Unity08 delegates, by maximizing their opportunities to be heard while minimizing the risk of negative interactions. In addition, I believe it would be relatively straightforward to implement this with fairly minor modifications to existing tools.

-- Ernest N. Prabhakar, Ph.D.