Don't worry about it and don't use benchmarks to detect 'bottlenecks'.
It's a massively overused term and worth remembering that there is no always-relevant system bottleneck, only in certain uses - e.g. sometimes weak CPU/powerful GPU is the right choice (actually most of the time for most gamers - only a few games require much CPU, but most specs make sure they handle these situations as insufficient CPU performance is harder to work round than when it's GPU), sometimes the reverse is ideal. So if you're worried about it in a certain use then test in that use.
(OK, if you've too little RAM to even launch your OS or something then that's kinda a special case
)
Every system in every use will have some component that is the slowest part and is holding back the system somewhat, and some over-enthusiastic people would call this a bottleneck. If it's not restricting other components from doing their job then forget about it. Also, even if it is significantly restricting other components, this is only worth worrying about if you're getting inadequate performance. e.g. 150 fps instead of 300? Who cares. 40 instead of 60? Smaller drop, but more relevant. If your GPU usage is sitting around 100% (95 to 100 is close enough! You'll never get perfect 100%) then the GPU is not being bottlenecked, even if faster other components could make the system run slightly faster overall.
Also, sometimes it's hard to see what is limiting your system, you'll generally do better comparing with similar systems with a component different to see what the effects are. Monitoring usage may give you a signal, but not necessarily tell you anything useful, especially where software caps are in place.
Edit: The video does this acceptably, tweaking his components to see how it impacts usage etc. Notice how his score drops a bit going i7 to i5 without dropping GPU usage - this shows that there are more contributors to performance than just GPU power, even when the GPU is the most limiting factor - this is not a bottleneck though!
