Amongst the many different advantages of modern technology, there is one in particular which never ceases to surprise us due to its simplicity and power: the possibility of performing real-time tests (sometimes complex) on content whilst at the same time reducing risks to a very low level.

MVT (MultiVariate Testing) is available to almost everyone and its efficiency needs no further proof, provided that it is implemented correctly. With MVT it is possible to select variables, their possible values, combine them in different creatives, set up suitable frequency capping (rotation of displays), experiment, measure, interpret, validate the best combination and then put it online.

Does this mean, however, that a manager’s role will be reduced to just a few simple configurations beforehand? Is “mechanical” MVT an exact Science to such an extent that it can be automated and trusted 100%. This is far from being certain.