[mattmanser]

That comment jarred with me too, in reality what they did is the total antithesis of what Paul's quote embodies.

There are a few options for seeing if people will buy before you build. The often quoted one here (and it's also used in the Lean Startup as an example) is have a sales site, have a 'buy now button', but go through to a 'we're currently in beta, email me when it's ready'.

Then buy google ads or whatever your sales pipeline is and see if people click 'buy now'. If no-one does, you've got a problem.

If you've got a big ticket item, you can do something similar but be upfront that it's not built. Talk to clients, see if they want what you're thinking of making and suggest a price to see if they say yes. Again, if you can't find anyone to say yes, you've got a problem.

And for some projects, Kickstarter is another obvious method.

[jacquesm]

Sell before you build: you go through what you could automate by doing it by hand. Automation is optional. If you actually sell automation then that is of course impossible but if you sell a product that you could create by hand just as well as through a computer program then you can sell right away.