>It just needs to process transactions accurately and list a bunch of shops across a few countries.
Yes, I hear this kind of off-the-cuff assessment frequently from potential clients, and I turn them down if I'm not able to help them see the complexity often involved.
After all, opentable is a restaurant reservation system for the entire world.
>You'd be amazed how "cheap" it can be to hire a greybeard these days, sometimes less than twice a fresh Stanford grad...
This is true, and the open ageism among SV companies makes me incredibly sad.
Knowing how/where you can effectively add caching, or even pre-aggregated data into an existing system can be the difference between a system going under, and being able to keep the single-instance database server from collapsing your business while a new system can be established.
If you have tight, modular code following good practices (e.g. a relational backend, not an ORM), you have experience doing it before, there's no micro-manager, and with a clearly defined, functional spec, you can build robust things that scale well relatively quickly. We're not talking about a new Facebook here, merely copying 80% of the functionality of an existing product for a launch in markets with no competition, as greenfield. Also, it's not like a site like that would reach a million users in a couple of months, nor would it win design awards, but why should it? It just needs to process transactions accurately and list a bunch of shops across a few countries.
I do think that the global market for programming talent is like that for real estate here in Singapore: I pay around 2 grand a month, and my neighbours across the street something like 8-10 for the same space. It is what it is. You'd be amazed how "cheap" it can be to hire a greybeard these days, sometimes less than twice a fresh Stanford grad...