One problem with the "perfect well approach" is that people see the result and will immediately dismiss it, because it is completely unrealistic - it's really hard to get people on board if you use that as a benchmark - I know that's not the point but this is how ppl react to it in my experience (especially if the performance engineer is younger) ... I find the "best in class" approach better in that regard, since "people have done that" makes it less academical and team members are more willing to listen - because everyone wants to know how it was done ... and that is a good starting point ... but of course the "best in class" approach may not reveal certain improvement potential.