There isn't a lot of published information that I've seen (or can quickly find) on how washio and postmates validated their business model, but I definitely agree that starting fairly lean and validating your business first even if that means a lot of manual effort up front is a better approach than investing heavily in infrastructure before launch.

Of course, if you are very successful, that manual work will pile up quickly and managing it could be quite painful, but you can have worse problems than a successful launch.

Maff Rigby's 7-day startup recommendation is definitely a good one.