In contrast to the first few consensus meetings which had sometimes lasted for a full day and had
been characterized by heated discussions, by the fall of 2004, conversations during the meeting were
cordial, lasting two to three hours. Facilitated by McMillan, participants focused on understanding
the reasons behind significant differences among the functional forecasts. Once the underlying
reasons for diverging forecasts were identified and discussed, the group would revise the proposed
consensus forecast by open discussion. Sales’ bottom-up forecasts that were slightly higher than the
other two forecast types were often justified by the sales directors based on their intimate knowledge
of upcoming sales deals or prospects; as a result, the higher forecasts were accepted by the group.