So I've been reading Sam Calagione's book "Brewing Up a Business" (which is great by the way), and he talks about starting with just a 10 gal system. He says the benefit was he didn't have to commit himself to one huge batch of a test beer, but he says he was brewing constantly.
I was curious how feasible it would be to start something like this, so I decided to do some number crunching:
To put it into perspective, let's say you get a full 10 gal. yield from a batch. This is 1280 oz, or 80 pints. If you brew 3 batches a day for 5 days, you will be generating 1200 pints a week.
If your restaurant has 120 seats, you are open for seven days a week, and you are doing an average of only 1.5 turns a day (a conservative estimate in a busy market if you are putting out good food), you are having 1260 people come to your restaurant in a week. Since everyone doesn't drink, and some will be families with children, let's assume that only 80% of your customers in a week will want at least one beer. This is 1008 beers, which leaves 192 pints, which means only 19% of your beer drinking customers will be getting a second pint. Ouch.