The Control Model –
When to take corrective actions?
One of the challenges in establishing an effective project control system is to know when to make corrective actions. The ultimate corrective action is changing the baseline. Measurement and Evaluation without Action is useless, but Action without adequate Measurement and Evaluation is just as damaging. W. Edwards Deming has given us some guidelines based on statistics.

In his “Funnel Experiment,” he asks you to imagine balls dropping through a funnel toward a target on the floor. The balls represent weekly progress, and the target simulates the baseline budget created during planning. Do the balls always land on the bulls eye? No - perfection is unrealistic for most projects. Deming says that over a statistically significant period of time, the smallest variance from the plan (see red circle) results from monitoring against a consistent baseline. If corrective action is taken before you have statistical results, divergence may result. You should work with the team daily to meet milestones, but wait to change the baseline. Item-by-item compensation can destroy a system that is already “in control.”