When it comes to experimentation, often the whole system product testing is seen as the pinnacle. However such tests usually only verify whether or not the product is working as intended. Carrying out detailed feature-based testing with such systems for exploration of a particular phenomenon is either not possible or very costly. On the other hand small-scale overly idealised tests can be conducted at pace and with much lower costs but the benefits come at the expense of reliability and unacceptable levels of uncertainty. In this talk the concept of High Performance Experimentation will be introduced which refers to targeted, intelligent testing, where phenomena of interest are isolated and first order contributors are accounted for with the right level of fidelity. An example of this approach developed to study complex turbomachinery vibration problems will be presented in detail.