At the end of a rollout there's often one reassuring number: training rate 95%. It travels to the steering committee, the light stays green. The problem: the rate measures attendance, not behavior. It says nothing about whether users actually work in the new system the Monday after — or whether Excel keeps running.
The gap between "can" and "does"
People take the fastest path to a result. If the new standard process costs three clicks more than the familiar workaround, the workaround wins — quietly, without announcement. Over weeks, a second, unofficial system grows alongside the official one.
The training record says someone was in the room. It says nothing about what happens in the system afterward.
Why shadow systems stay invisible
Workarounds leave almost no trace in reporting. The transactions still run — just with manual steps before and after. Usage stats show logins, not quality. And no one volunteers that they're bypassing the expensive new process. The information only exists where the work happens: on the front line.
What you gain
Once real usage is visible, you can course-correct with precision: train where it sticks, simplify processes no one uses, and stop the creeping productivity loss before it hits the quarterly report. Adoption shifts from a gut feeling to a metric you can steer.