Measure for reporting, then measure for action

Recently I’m finding it useful to split some of the data-driven initiatives I’m working on into two parts: measure for reporting, and measure for action.

Measuring for reporting is the production and presentation of accurate enough data that tells what has happened.

It allows us to to have a discussion about what we’ve done and how we did it.

It’s automated and runs reliably in production, so it’s always available when you need it.

Examples include company revenue numbers, team performance and cost attribution to business units.

Measuring for action is data that is provided in near real-time, so it can be used to drive an action or a change in behaviour.

It facilitates a short feedback loop, so you can quickly see how your actions affect the measurement.

Examples include software observability, A/B testing data and comments on PRs indicating how your change will affect your cloud costs.

I first heard about this framing on a talk about driving carbon reduction, but I think it’s a nice way of framing any data-driven initiative, from growth strategies to FinOps and more.