**7. Success with restrictions**

Assuming the predictive model is of good quality, the responses and its degree of achievement can be evaluated against priorities of the projects' definitions. Within the most experimental setups, responses targets are optimistically defined and with a high degree of safety. Sometimes not all of these targets can be fully achieved. Therefore, and as described in the optimization, responses can be weighted again according to their priorities and their targets, in order to find a sufficient compromise. This compromise could mend not only the factors and responses that need to be adjusted but could also mean that conditions that act as disturbances need to be compensated. Potential disturbance factors are, for instance: Relative air humidity, temperature, water content in raw materials, using different machines or bad machine calibration.

Also predictive models are only as good as their data. Even if factors are sufficiently enough arranged to describe the target functions, there are still a lot of things which could impair the prediction model such as bad pruning of the model terms, bad distributions or deviating experiments, missing factors, bad measurement *(calibrated)* equipment.

In addition to this, the approach is always to minimize the work, time and budget with efficient designs *(fewer experiments)* and fewer factors, so that not all important factors are implanted or the effect of factors are underestimated, and are, for instance, of a higher order than assumed. So, if a process is not linear and linear designs are used, the predictive capability of this model is very limited. If necessary due to interaction-, squared- or cubic factor-terms, the design could be complemented step-wise. The design of higher-order processes, complex designs are not recommended at the beginning, since these drastically increase the number of experiments. Complexity can always be reduced by focusing only on a small process space (Fig. 18).
