There is a lot of debate about the direction the designs have taken over time and this generation in particular seems to have very polarizing views.
Some designs, like the thunderbird, the dolphin and Primeape's evo seems to be well-received, others like the starters and the coin ghost thing that evolves into a skateboarder are controversial.

I think the issue is that the higher the number of themes they try to incorporate in the design, the higher the odds these themes will clash with each other. This is something that has happened in past generations too, like with Samurott and Duraludon.

Take Quaquaval for instance:
>duck
>peacock
>carnival dancer
>flamboyant
>strong kicks to jusitify the fighting type
>3rd stage starter so it needs to be more complex than a NFE

That's a lot of stuff to cram into a design compared to Psyduck or Ducklett. You also need to work with a relatively limited color scheme (most pokemon have 3-4 colors at most), it needs to be marketable and it needs to "look like a Pokemon" so so you can't have too many details like Digimon's obsession for random mechanical parts and anatomically accurate limbs.

At some point it becomes very hard to harmonize all this stuff in a Pokemon design.
Blaziken for example was just "fire chicken with a hint of taekwondo martial artist", Empoleon was "emperor penguin with a regal theme" and so on.

I think 3 is the maximum number of themes they can work with before the design is a clusterfuck.
For pic related the themes are:
>mantis
>stone axes
>"ancient" vibe
>Japanese vibe

As you can see it's a bit too much, which is why its design is a bit too busy compared to Scizor.

So rule of thumb is when a design feels off, count the number of themes it's based on and if the number is too high that's probably the reason.

Of course underdesigned mons are also a problem (like Dunsparce's evo, wtf?) but they don't stick out as much as the overdesigned ones.