The Control Model – When to change the baseline?
The funnel experiment shows that we must have faith in the original project plan until we have enough statistical information to conclude that the baseline plan should be changed. However, every day the project team should measure schedule variance (SV) and cost variance (CV) against the baseline budget (PV), and make minor corrections.

See the diagram on the next page. When either variance approaches ±30%, and the project is 25% to 40% complete, there is sound statistical and motivational reasons to change the baseline. These conditions indicate that either the plan was in error, or that some unanticipated events have forced a direction different than planned. Until that time, continue aiming at the target.

What does 25% to 40% complete represent? They are zones of statistical information for medium to high risk projects. This is also the time when team members lose their faith in the plan, so it must be revised to tell the truth. Otherwise, they may just expedite and not follow the plan.