I rate Ripple, but they will need to keep up with some of the new developments on Bitcoin and Ethereum which are mostly initiatives to improve performance.
My gut says that if you need consensus between nodes to validate a transaction, then you will never be able reach the speeds of a network that does not need consensus. So to counter that, I think the blockchain will need to split into something along the lines of payment channels and those will then need to have sub-payment channels and so on. That way if you can work out an algorithm where wallets that pay each other frequently reside in the same payment channel, you can offload a lot of load off the parent payment channels and ultimately the main blockchain.
There are other potential enhancements. For instance, not all payments are equal, some need to be fast and accurate (i.e. at a retail store), some just need to be accurate (i.e. person to person payment), some don’t even need to be accurate (i.e. train or bus ticket where speed outweighs accuracy). If you can categorize these types of transactions then their might be other enhancements in implementation.
I think it is only a matter of time before we find solutions. Whether Ripple can stay ahead of the other larger cryptocurrencies, we will have to see.