6. Mistakes to avoid

The first potential mistake is not wanting to admit that improvements are possible, even after the system has gone into production. Many people believe that at that point "everything is done", and that any questioning, however small, could lead to disaster.

The second might be to want to "save time" and turn everything upside down, without recognizing that variations must be minimal and close to "uncontrollable". Patience is the key.

Then comes the risk of not knowing how to stop. And then, at a certain point, you have to admit that you're starting to go round in circles. It's sometimes this caricature of the "researcher who seeks" that you have to fight against. At a certain point, you have to admit that the variations observed in the result are so small, but above all so random, that you can consider yourself to have reached the summit....