Written By:
lprent - Date published:
7:52 pm, June 13th, 2009 - 7 comments
Categories: mt albert -
Tags:
The by-election is over apart from finding out the result. So I’m dragging my tired and abused body to the party to find out the result.
It was a long day and night. This is the second time I’ve been struck down with a cold, or in this case, some obnoxious common (not swine) flu right before an election. I cannot reccomend it. It makes setting up the backup on backup systems a bit of a nightmare as you puzzle over some arcane database relication tweak in a belated overnight work session.
But the systems worked fine with a few minor quirks. So did the election day campaign. I suspect that the turnout for the election was a bit light. I have an interesting feeling that it was lighter on the other side than ours. If it is then how interesting!
Cheers
Lynn
BTW: the auto-moderation is now off.
One third of polling booths counted (don’t know which areas), Shearer well ahead, Lee holding off Norman in second.
Bill and Ben Party currently ahead of United Future, Libz, Smack Kiwis, etc.
I’m curious, Lynn – what exactly are all these systems for? I didn’t know so much technology went into running a party campaign…
Congratulations Lynn, the campaign team in Mt Albert, and David Shearer. An excellent result! Will watch Shearer’s career with interest.
Now Lynn – get some rest.
Predictive targeting and operational control systems. We usually have scores of untrained volunteers prior to election in canvassing, and hundreds of people turning up on election day. No training and a very limited time to get any.
You use systems to fit people effectively into a cohesive system so you use their efforts efficiently. Making people understand that their efforts have a focus of purpose and are extremely efficient is extremely good at getting volunteers wanting to keep helping over time. It means that you don’t have to be an raging activist to participate in the grassroots democratic process at more than a voting level.
Because I come from an operations family background, my MBA major was in operations research, and because I’m a damn good programmer – I express my thinking in computer programs designed to enhance the effectiveness of volunteers. The code does that by regressive use of information for targeting. In other words using the same kinds of numerical analysis that has been used forever for correlations in many diverse fields. It also does it dispersed over the net leading to a much less hierarchical approach and one easier to do in the larger electorates we have today.
It is a very useful adjunct to the more traditional marketing techniques like message and policy. It also works a lot better with volunteers than trying to indoctrinate them with boring endless meetings.
Been slowly building systems and ideas to do this for a few decades starting from the principles that Anderton left behind in Labour a few decades ago.
I suppose that working on this site comes out of the same underlying rationale. It is a hell of a lot more satisfying than trying to change ideas in a political party. That is glacial.
If I understand correctly: deployment efficiency and impact audits in a feedback loop?
L
Yep. The more you understand about your electorate, the easier it is to concentrate resources where they are needed. The trick is to be able to use relatively small amounts of numerical data to sample reality. The more you do it, the more that you know what you don’t know and adjust your targeting to suit. I’m afraid that about half of what political people understand about electorates is bullshit when you actually test it.
Makes it a lot easier to run a coherent plan across the whole of an electorate when you’re doing it with real data, and on a distributed network.
Thanks Lynn,
Fascinating! I am astounded, all makes logical sense of course, but how very slick.