The wording that Nikon use, 'tested for 300,000 cycles' is vague, misleading and contains too little information to make an informed judgement. It is typical marketing speak. In reliability testing we use more clearly defined terms like 'mean', 'median' or 'expected lifetime'.
Presumably they intend this to mean the expected lifetime, or mean lifetime. If that is the case you can expect roughly an equal number will fail before 300,000 cycles and roughly an equal number will fail at greater than 300,000 cycles, depending on the failure distribution.
But if roughly half will fail earlier than 300,000 cycles the crucial question becomes how much earlier? This is where Nikon supply no information whatsoever. And it is this information which is of vital interest to the average camera user.
A more useful statement from Nikon would be one like this: No more than 5% of shutter mechanisms will fail in less than 100,000 cycles and the expected or mean lifetime is 300,000 cycles.
Formal reliability testing aims to provide more meaningful information of this kind. It is certain that a company like Nikon knows a great deal about doing proper reliability testing so they will have the full information. But they (like other companies) have chosen to release incomplete and misleading information. This is deceptive behaviour bordering on the dishonest. Sadly it is the norm.
Below is a short background to the kind of information they could provide.
See this link and this link.
Reliability data is normally modelled by a failure distribution pattern called the Weibull distribution. It typically describes the failure patterns of complex systems.
It has three parameters called the Shape (beta), Location (gamma) and Scale (eta) parameters. Reliability testing aims to determine the value of these parameters. Once they are known one knows a good deal about the pattern of failures that will occur during use and allows one to make statements like
'with a confidence of z% we can say that no more than x% will fail before y cycles',
which is usually what we want to know.
One can think of the location parameter as being roughly equivalent to the mean lifetime and the scale parameter as being equivalent to the standard deviation. But that is a very crude approximation that depends on the shape parameter.
I have gone on about this at some length because cameras are high value items that are critical to the professional life of many people. We have a right to expect honest and complete statements about their expected lifetime so that we can make informed choices and have reasonable expectations. Although I have mentioned Nikon in the context of this question it seems that all camera companies are equally guilty.