Test and Iterate: SHIFT Stage 5
The Problem This Stage Solves
Most change programs evaluate once, at the end. They measure outcomes (adoption rates, satisfaction scores, efficiency metrics) after the investment has been made. If the results are poor, the organization either doubles down on the same approach or declares the initiative complete and moves on. Neither response is informed by behavioral data.
Test and Iterate replaces end-point evaluation with continuous behavioral measurement. The purpose is not to prove success but to learn fast and adjust.
What to Measure
Behavior, not sentiment. Most change measurement relies on surveys: do people understand the change, do they support it, do they feel ready? These measure reflective states, which correlate poorly with actual behavior. The COM-B model predicts this: reflective motivation is only one of six behavioral determinants.
Behavioral measurement tracks whether the specified target behaviors are actually occurring, with what frequency, in which contexts, and by which populations.
How Measurement Feeds Back Into Diagnosis
The real value of behavioral measurement is that it tells you whether your intervention is addressing the right barrier. If you deployed a capability intervention (training) and people are attending but still not performing the behavior, the barrier was probably not capability. The data triggers re-diagnosis.
This is the learning loop. Measure behavior. Compare to expectations. If behavior is not changing, re-diagnose. If re-diagnosis reveals a different barrier, select a matched strategy. Deploy. Measure again. This cycle can run on four-to-six-week iterations.
Rapid-Cycle Testing
Pilot an intervention with one team or one location before rolling it out organization-wide. Measure behavior change within the first four weeks. Use the data to adjust the intervention design. Only scale once the behavioral data supports it.
The organizational benefit: this approach converts change from a large, slow, high-risk investment into a series of small, fast, evidence-based experiments. Each iteration costs less, teaches more, and gets closer to the intervention design that actually works.
Key metric shift: move from measuring "how many people completed the training" (activity metric) to "how many people are performing the target behavior this week" (behavioral metric). The first tells you about your intervention. The second tells you about behavior change.
