At Strata Santa Clara last week, we ran an Oxford Style debate. Mike Driscoll has a great writeup of the discussion, which got plenty of people talking—Ars Technica, despite feeling we were breathing our own data exhaust, admitted there may be something to it.

These have always been great fun; we’ve run them at Interop in the past, too. They work well for a number of reasons:

• They polarize a nuanced topic so the audience can hear both extremes, and ironically, realize that the truth is somewhere in between.
• They engage the audience in voting, so they have skin in the game. In my experience, the drop-off throughout a debate session is much lower than for panels or presentations.
• They have a “conclusion” everyone wants to tweet about. These are often picked up by the media (here’s the Informationweek writeup of our cloud debate.)
• They’re inherently sales-proofed. They make it really hard for vendors with an agenda to sell from the stage, since they’re busy debating a point rather than pitching their product.

Despite the success of the Strata debate, we set it up at the last minute, relatively speaking. And that may be a good thing: a debate works when there’s a hot topic—which often only becomes apparent at the event itself—and when you have smart celebrity types able to think on their feet. Debating experience helps.

Running an Oxford-Style Debate

These debates follow the Oxford format; if you’ve ever heard the NPR “intelligence squared” podcast you know this format. If you’re thinking of running this kind of event, do yourself a favor and listen to a couple of their episodes.

First: get a great topic

More than anything else, the key to a successful debate is a clear, controversial statement. “The sky is green” would work well; clearly everyone has opinions on it, and it’s polarizing. This is where you need to understand the audience’s mindset. We titled an Interop debate “Clouds are secure,” with great success.

The Strata topic was, “Domain expertise is more important than machine learning skill.” That’s a bit wordy, but fine for the audience. We discussed some other topics: “Big Data is Big Brother”; “Stats will Save the World”; “Science as we know it is over”; “Correlation beats causality given enough data.” Getting this step right is essential

Then: find great debaters

You need two teams of 2-3 people. One team is for and one against the topic. These people are the second most important thing. They need to be willing to put aside their sensible, balanced views of the world and instead take a rabid stance in favor of one extreme or another. In the aforementioned example of “the sky is green,” obviously few people think that’s the case. But a good debater embraces a position whole-heartedly.

It helps if some of these people are known to the audience—big names, or folks with big online followings. This makes everything they say more interesting, and provocative. It’s important not to underestimate the importance of entertainment in these things.

Then: run the event

Set the teams up on two tables, on either side of the moderator. You may want to put nametags in front of them, or slides with their names and Twitter handles for all to see. The moderator can also listen on Twitter or some other social network to solicit questions from the audience.

The moderator summarizes the proposition, often with some context and clarification of what it means. This should be as unambiguous and nonpartisan as possible, because it’s really setting up the basic rules for the debate.

The moderator begins by surveying the audience (a show of hands will do) to see how many people are for or against the motion.The moderator introduces and frames the topic. He/she then surveys the audience (a show of hands is usually enough) to see how many people are for and against the proposition.

Then the two teams get several chances to speak, usually in 5-minute chunks:

• They give their opening arguments
• They get to respond to their opponents in rebuttals
• The moderator gives them questions from the chair, floor, or social network back-channel
• They summarize their positions with closing arguments. Typically, whoever goes second for opening arguments gets the final word in closing arguments to even things out.

The chair then surveys again, asking anyone who wasn’t around for the first vote to abstain. The winning team is the one that moves the needle in their favor.