Starting From Examples, Not Rules

The wrong way to build a rubric is to start by thinking about what should make work good—to sit down and write out the criteria you think matter, then test it. This almost always produces a rubric that's either too abstract or too rigid, neither of which actually works.

The right way is to start from examples. Find three or four pieces of genuinely excellent work in your domain. Find three or four pieces of adequate or mediocre work on similar tasks. Study them. Ask: what's actually different between the excellent work and the adequate work? Not what should be different. What is.

Write your observations down. Not as abstract principles, but as specific observations. The excellent writing has shorter sentences. The adequate writing has longer ones. The excellent design has clear visual hierarchy with distinct levels. The adequate design treats all elements as equally important. The excellent proposal opens by framing the problem in a way that immediately shows why it matters. The adequate proposal launches into solution details without establishing why we should care.

These observations are the seeds of your rubric. They're grounded in actual examples. They're specific enough that someone could recognize them in new work. They're not rules—they're patterns you've noticed.