However, better predictability can be achieved if the system is linearized. Linearization of the system makes the link between cause and effect more direct. In mathematics, linearization is brought about with a small variations analysis. By proceeding by small steps, non- linear systems are transformed into linear relations. Small variations analysis is the mathematical equiva- lent of what Braybrooke and Lindblom (1970), with their disjointed incrementalism, and Quinn (1980), with his logical incrementalism, propose for the organiza- tion. These are step-by-step approaches which allow managers to proceed marginally. By doing so, the man- agers are in a position to observe the direct effects of their actions. As a consequence, they can modify the thrust of their effort in an almost continuous manner and make the necessary reorientations. Proceeding in- crementally improves the predictability of what man- agers do. However, when the organization is in its chaotic domain, prediction can only be in the very short term, because in the long term, even the slightest variation can have a great impact on the organization