**Some tips on getting the most out of your [[Metrics]], [[KPIs]], and [[OKRs]]**
# Ease
The harder it is to measure, the less frequently it will be measured. The less likely it will be effective as a metric. People respond to [[Incentives]].
# Clarity (Relevance)
A metric should be obviously relevant to the thing ([[Defining Area|area]] or [[Defining Project|project]]) they are intended to measure. If it's possible to avoid using [[Proxy Measures]], do so.
# Clarity (Objectivity)
Where possible, hard quantifiable, indisputable measures should be used over subjective, or "soft" metrics.
# Pairing [[Leading Measures]] with [[Lagging Measures]]
Leading measures are a good way to see if something is going off the rails before it happens... lagging measures are a good way to see if the thing you did about it was effective. [[Leading indicators measure strategy adherence and lagging indicators measure strategy effectiveness]].
# Standardize
Having a dashboard (something like a [[Work, Health, Love, Play Dashboard]]) or a common personal operating system (something like [[Periodic Reviews]]) will give you the needed structure to actually rely on your system of metrics.
# Think about what Behaviors are being Incentivized
People respond to incentives. If it is possible to [[Gaming KPIs]] and [[Rule Beating]].
# Pair Quantity with Quality
A KPI based solely on quantity incentives pushing products out the door fast, no matter what quality they are in. It creates a culture of go, go, go, which can be reckless to your bottom line.
A KPI based solely on quality incentives not producing anything at all. You have zero defects if you make zero shipments.
A good KPI pairs quantity with quality. This prevents (or dissuades) [[Gaming KPIs|gamifying metrics]].
# Consider Gaming Manipulations
If you are designing a measure and a metric, consider not only what incentives you’re driving, but also how those who are being measured could subvert the results via [[Rule Beating]].
# Measure the _real definition_ of success
It is not always possible to measure success directly. When using [[Proxy Measures]] , try to align them as closely as possible with what “success” would _really_ mean. Throughput is one aspect of a process, but if what’s being throughput is garbage, than you have an assembly line whose very successful at making garbage.
The best KPIs are those that, when green, _positively_ ensure that all is actually well - and not that someone is playing games. These aren't always easy to come up with.
****
# More
## Source
- [[Measure What Matters]]
## Related
- [[Gaming KPIs]]
- [[KPIs]]
- [[Goodhart’s Law]]