I have had a few characters now that were a bit behind in their main stat in return for other things. Totally worth it. And highly effective.
Of course, they were also full-casters that weren't sorcerers. Char-op may not just be about a few stats.
Certainly. Some classes are better than others, e.g. Wizard is pretty much unequivocally stronger than Sorcerer even though both are fairly strong due to being full casters. Within a single class, some subclasses are better than others, e.g. Battle Master is better than Champion (not that that's a difficult bar to clear, but it's a ready-to-hand example.) Within a single subclass, some choices will succeed more often than others, e.g. a Barbarian with 16 Str is going to succeed a lot more often than a Barbarian with 8 Str, because the former will be rolling attacks with advantage, and getting bonus damage, whenever they Rage, while the latter will not. Even within a single subclass and stat array, some choices are better than others, e.g. there's no reason to use a short sword as a Dex-based melee attacker when you can instead use a rapier, since the two are equivalent in every possible way except damage dice, and you can always reflavor what your weapon looks like (e.g. perhaps you use a "leaf-blade gladius," which has the
stats of a rapier but
looks like a special variant of short sword.)
Back in 4e, these differences were even easier to demonstrate, because you had quite a bit more depth of choice. As a Paladin, I generally favored playing a Dragonborn (no one who has interacted with me on this forum should be surprised by this fact), which worked well as a so-called "Balanced Paladin," pushing both Strength and Charisma as opposed to only focusing on one side or the other. Such a character has free choice of any attack powers, since Strength and Charisma were the two options for attack and damage, but tends to have weaker secondary stats and thus weaker "rider" effects (e.g. some attacks give Wis mod temporary hit points or the like.)
Buuuut...you could totally still have solid secondary stats if you wanted, because there were lots of ways to mitigate the difference, and that was generally my personal preference. I would take 16 Str and Cha, essentially the equivalent of taking 14 Cha as a Sorcerer in 5e, because I knew that that choice wouldn't be a big deal on the grand scale: I could choose an accurate weapon (such as a longsword) to mitigate the reduced attack, and quickly pick up feats that gave mitigating bonuses; I would usually take an At-Will attack that gave a hit bonus equal to the number of adjacent enemies, further erasing any accuracy issues; and I would try to pick up a lot of powers targeting Fort/Ref/Will, since those defenses are generally a bit lower than AC is. In exchange, I'd get slightly higher secondary stats, allowing me to use stuff like Lay on Hands more often, or to get more oomph out of my Healing Surges (since Dragonborn add their Constitution modifier to their surge value.)
Unfortunately, 5e has chosen to flatten almost every mechanical aspect of the game. Feats are one of the few areas that
haven't been flattened to hell and back. As a result, they're pretty much the last bastion of mechanics offering any form of depth whatsoever (them and multiclassing.) But since they ALSO made that depth actively compete with "get better at your basic stuff," we now have a dramatically worse situation than the "feat taxes" of 4e or even 3e, where you have essentially "ASI taxes." Don't do the
interesting stuff feats offer, because interesting is
less effective than the incredibly dull and boring +2 to your bread-and-butter actions.
In pursuing a simpler, more intuitive game (which is not a bad goal), they have instead produced a flatter, less-interesting game. Instead of making baseline competence guaranteed, so that people can choose to do whatever they like within the mechanical space
without worrying about optimization, they have created systems which actively force a choice between "do the optimal thing" and "do the interesting thing." It would be like if 4e had allowed players to take the bitterly-disliked "Expertise" feats repeatedly--no one would ever take anything else because +1 to all the stuff you do regularly is Just Too Good. And these were issues people called out
during the public playtest, almost a decade ago.
5e's design is significantly responsible for this problem. 5e COULD have been designed such that pursuing variety was rewarded, or so that basic competence was guaranteed, thus making any further choices purely a matter of what interests the player. But that's not what the designers chose to do, and now people complain that players follow the stuff the game's design rewards.