Ethereum stores state of all contracts in a single blockchain. This is conceptually nice and convenient, but it’s not scalable. If you are running an Ethereum node, you’ll have to process logic of all apps using Ethereum and store all of their data on your computer.
Also it might be not cost-effective. If many apps will compete for scarce block space (and it has to be scarce for people to be able to run Ethereum nodes), only highest-bidders will be able to use it.
Ethereum developers want to address this issue by ‘sharding’, that is splitting blockchain into smaller semi-independent pieces, but it’s not clear if that’s going to work.
Lisk’s approach is very different: each dapp will run in its own sidechain, so they are addressing the scalability and incentive structure issues from the start. There is also a number of other technical differences.
It doesn’t mean that Lisk is better than Ethereum, it’s quite possible that their sidechain architecture is broken from security standpoint. But at least we know that 1) they are aiming for roughly the same market segment; 2) try to solve important issues; 3) do things differently.
It’s possible that Lisk and Ethereum will target different niches and will co-exist peacefully. It’s also possible that some another project will appear and eat their lunch.
If you consider participating in Lisk crowdsale, consider this: Lisk is a fork of Crypti done by members of Crypti team. People who invested into Crypti will be shafted if Lisk rather than Crypti becomes successful. So the thing is, these people can do it again if they run out of money, so they will start yet another cryptocurrency and Lisk investors will be shafted. But it’s also possible that this time it’s for real, but you should be wary of this.