My instinct is—rather than measure informativeness—measure a kind of “travel” of information from some expert individual to the top of a deliberation process. Ideally, some random expert that the organizers wouldn't even have known about.

From our process: I love that our leading user-asks-about-abortion related value comes from a woman who had an abortion against her parents will when she was 19, and didn't like how people talked with her about it. The “expert” turned out to be a lay person who was unusually articulate and had a very specific experience.

How to measure this?

• Were the facilitators of the process surprised and delighted by what emerged?
• Did participants think their contribution ended up in the right place in the output?
• How much did people's beliefs or values change, as they went through the process?
• Over many runs, what's the entropy on the backgrounds of people who contributed winning pieces of material?

If this framing seems interesting, let’s develop a metric together and compare some of our mechanisms on it. We could even create a democracy leaderboard, for how well different mechanisms surface unexpected, but brilliant input.