In bigger organizations recruitment is an ongoing process. So even if you are leaving such an organization finding your replacement is not a tough task(Of course in companies like Google too if a key person is leaving it is a difficult situation!). There are buffer resources within the project. Also you can peep outside the project and get a person from 'bench'! While you are leaving the company you transfer knowledge to other fellow members of project or a new comer...in neither situation is it a big deal! Not a great resource crunch is felt.
Is it equally easy to handle attrition in startups as with giant organizations? Definitely not. It might take some time to find a right fit for the project (A wrong especially hasty choice becomes a bigger burden!). Search for another resource begins only after a present employee has resigned. Even if a perfect replacement is found immediately, the company has to bear salaries of both the resources at a time- the one serving notice and the new one. For bootstrapping startups especially this becomes heavy. On the other hand if no person is found to shoulder the responsibilities and tasks, the responsibilities are divided and distributed among rest of the members. They already have lot of tasks and a timeline to meet (This add-on may discourage them!).
Regular/higher attrition will obviously affect the release/delivery of product. The release is tightly bound with revenue...delay in delivery is delay in getting paid for work from customers.
Theoretically it is easy to chalk out plans like signing bond of minimum duration to serve and so on. But practically it does not work out well. Binding employees with such contracts can prove to be turn offs for joining an organization (may be 'bond' sounds very legal :) ). So rule of thumb- better leadership, better planning and better management to keep employees happy and satisfied with work!