Prove Incremental Value.
Correlation can tell you what happened alongside a program. Test-and-control discipline helps determine what the program actually caused.
Correlation can tell you what happened alongside a program. Test-and-control discipline helps determine what the program actually caused.
Incrementality matters whenever an organization is deciding whether to scale, stop or change an intervention. It protects investment decisions from stories that sound plausible but are not causal.
What would we need to observe to confidently say this program caused incremental value rather than simply occurring alongside it?
A useful test starts with a clear treatment, a credible control or comparison group and a defined outcome. The design should be simple enough that business leaders can understand what is being tested and why.
Engagement metrics can be useful leading indicators, but the ultimate test should be tied to the behavior or economics the program is intended to change.
A test is not successful because it produces a statistically interesting number. It is successful when the evidence changes a decision: scale, redesign, stop or test again.
Short, practical perspectives on strategy, customer growth, decisioning and execution.