How far each class actually moves the needle
The three products are the same hosting on three widths of address diversity, separated by exactly one position in the address. Every figure below is the monthly equivalent on a 3-year term. Class C gives every site a distinct third octet: non-consecutive blocks, the deepest European pool, the right answer for most networks at $2.98/IP. Class B moves the difference to the second octet at $4.55/IP, so two sites no longer share a parent range. Class A moves it to the first, and that is where the sequence ends: there is no fourth product, because there is no octet further left to buy.
The honest version: none of this is a ranking factor on its own, and we will not pretend otherwise. What address diversity removes is a shared liability and a visible pattern. Sites answering on one first octet share a reputation, share the fallout when a neighbour is blocklisted, and present an obvious shape to anyone who sorts your domains by their leading number. Class A takes that shape away completely; whether that is worth $9.00 a month per address is a judgement about your own risk, not a promise about rankings.
When the first octet earns its price
Three cases come up repeatedly on the European side. A client portfolio where the clients are direct competitors and the sites must not be traceable to one another: one shared leading octet undoes that, and separate branding does not fix it. A small, tightly themed European network where every site targets the same language and audience: the tighter the theme, the louder a shared address component reads. And a set of linking domains whose whole value rests on each one reading as an unrelated independent host.
If none of those describe your build, buy Class C in Europe and put the difference into content. The EU Class C pool is deeper, its ceiling far higher, and for most networks the third octet is already more separation than anyone is checking for.