What I don't understand is that if soft-forks are backward compatible why would other miners reject it? Doesn't backward compatibility mean that new blocks are seen as valid by old clients? And if you wait everyone to update what is the benefit for soft-fork? Is there a con of Hard-Fork if it has 95% support?
The way that the soft forks work is that it makes something previously valid invalid. This means that it is backwards compatible, old nodes won't reject stuff made with the new rules. However, that also means that something invalid under the new rules can still be valid under the old rules. If miners and nodes are still running nodes with the old rules, then they will accept those transactions and blocks as valid, but not the new nodes. This can cause a split in the network and potentially a blockchain fork, and just be a major pain to resolve.
The point of the 95% signalling is so that everyone knows that it is safe to use the new rules because the miners are promising that they will use the new rules when creating blocks. It ensures that nearly all miners will enforce the new rules which means that the likelihood for a chain split is much much lower.