That transaction is 90 KB in size; presumably the exchange is aggregating deposit transactions from many customers. And its fee is about 1.3 satoshis per byte. As you can see here, there are currently lots of outstanding transactions offering a higher fee rate; as such, a miner would likely prefer to confirm 90 KB worth of smaller and more lucrative transactions, rather than this one.
The output address type doesn't have any direct effect on confirmation time or miner preference, as far as I know.
The size of the mempool in MB is not a particularly useful statistic in estimating confirmation time (until it gets below the block size, at which point you would roughly expect that the next block will confirm everything). Transactions don't get confirmed first-come-first-served; a low-fee transaction can languish indefinitely as long as more higher-fee transactions continue to arrive and be given priority.
Note that as a customer, this is not your problem. Your job was done once your deposit transaction was confirmed, and the exchange should have credited your account at that point. What they do with the money after that is solely their concern. It's quite possible that they've intentionally lowballed the fee. If a lull in transactions occurs sometime soon, maybe they get lucky and it confirms at this low fee. Maybe they don't care whether this happens anytime soon; after all, it's probably just money moving around inside their organization. If at some point it becomes more urgent, they can offer a higher fee (e.g. using RBF or CPFP). But none of that is any concern of yours.