Pros and cons to everything said in this thread, but in general modules need a balance pass and a certain ammount of granularity added to them. Before taking to the soap box, remember that the effective volume of any class of module is 2 to the power of that module slot class times 2 cubic metres.
As in Class 3 is 2 ^3 = 8 cargo canisters, a cargo canister is a 2m long 1m diameter cylender... square it off its 2 cubic metres a canister, 16 cubic metres for class 3.
So take limpets for example:
A lot of "builds" end up putting a class 5 limpet controller in a class 6 slot. So 64T cargo rack, or a limpet controller capable of 3 times what works in a slot capable of a 2T cargo rack into a slot 32 times larger than the class 1 slot...
And thats before we ask about the other 32 cargo slots --> 64 cubic metres. To put that in perspective, the missing or wasted 32T when fitting a class 5 limpet controller into a class 6 slot is like an articulated truck's payload. When you are allowed out from corona lockdown, look at an atriculated truck, and imagine that carrying 32 IBC's (palletised 1m x 1m x 1m 1000L capacity tank) of water, thats how much space / weight is lost putting a class 5 module in a class 6 slot.
Limpets are a good example, but fuelscoops are even worse:
Every class returns significantly less than double the scoop rate of its predecessor, for ~3+ times the cost.
Frame Shift Drive's scale "properly" in terms of mass doubling with each class, but are similarly 3+ times the cost for each class and offer less than double the optimal mass and or fuel per jump. Exception being the class 3-->4 jump where mass and fuel allowances are greater than double:
Exploring the various sizes of ships modules available in ship outfitting typically shows diminishing performance as class increases, which strongly contrasts with the cost of these larger modules - which is exponentially increasing.