Because the test is conducted by man, there will be a level of subjectivity. But if a test params and judgement criteria are agreed upon, then it should be good. It's like a auto magazine shoot-out comparison tests. Even with the same cars tested, it's not that Automobile, Road & Track, Car and Driver, Consumer Reports, etc. will select the same winner. However, as we know how they are tested and how they are judged, we can consume the feedback intelligently.