Some things matter, some things don't. I'm not going to say what does and what doesn't (well, maybe one or two) - but here's the pragmatic view:
pre-1. What we hear is not all based on sound that goes into our ears - it is more dependent on the ear-brain interface. IE: how we interpret and react to what we hear. And the emotion that's generated.
pre-2. We have very bad auditory memory. Sound for human beings is mostly an in-the-moment things.
pre-3. The brain's interpretation can be based on many things, including mood, biases, desires, etc.
Remember those three points for the rest of it.
1. We know what we put into our systems. Unequivocally. We put a single waveform into our systems. No more, no less. We can define that. Easily. Nothing else goes into the system, so nothing else can come out.
2. We can very accurately measure what comes out of our systems.
3. The ideal / perfect system can only be as good as when 1 (above) = 2 (above). It can't do any more than that.
4. Therefore: If any change we make can not be detected as a change in the reproduced waveform, then it is not doing anything. That's an unassailable fact.
5. Granted, some inaccuracies (where 1 intentionally DOES NOT = 2) can be perceived as "good" - eg. even-order distortion, or frequency response inaccuracies that generate "warmth". Fine, but you can still see that in the generated waveform.
So that's looking at the whole system.
If we look at just digital components, it's even easier to determine if anything is going on. We can easily test the bits-in and bits-out, and perform a basic mathematical checksum to actually confirm if there is any change. No measurement required!
Many reasonable DACs (such as RME ADI-2 and many others) actually have "bitperfect testing" built into them, so you can see this very easily - all the work is done for you. You just get a "correct" or "incorrect" displayed on your screen.
So whether it's gone from your SSD, through an audiophile switch and to your DAC (or whether it's come from the cloud, through the Qobuz servers, bounced through 20 or 30 very (non) audiophile switches and routers in data centres around the world, through kilometers of cable in the streets, maybe even bounced once or twice to a satellite and back, through your ISP, modem, etc and through a cheap USB-IF compliant cable), the result is identical. Disagree? Too bad, we can prove it.
Therefore, I'm confident enough to go out on a limb and say things like audiophile switches, USB cables, ethernet cables, power cables, digital "cleaners", etc (and basically anything on the digital end) have absolutely zero effect. It's all in your brain. (Unless there is a flaw / problem).
Some of these things are actually flawed and quite laughable from a technical standpoint. eg. No matter what marketing guff anyone wants to carry on with - you can't "re-clock" a unclocked stream (asynchronous USB stream). The definition of asynchronous is "no clock data, no bearing on time". So if a so-called engineer can't even understand the parameters of what he is working with, what confidence should we have that the product is not actually harming something?
As we're all here to obtain the best sound possible - there are plenty of things that **DO** have a huge impact on the sound generated, for example speakers, room treatments, and positioning to name but three.
The sad thing is that attending the things that matter usually involves much more work than connecting up a sexy, oversized and impressive-looking braided USB cable with chunky gold plugs. Laughably, some don't even meet USB-IF specs. So most people take the easy route and pretend it makes a difference.
So if we're going to invest time / money / consternation on something to improve our sound, our efforts are best directed to those things that we know will make an audible difference rather than things that we can very easily prove will not make a difference.