I don't think I'm so much deflecting the question as much as just highlighting that Github is a special case, so one should take extra care when translating their insights and lessons to one's business.
Github is a product for geeks (few of those take off), build by a founder team consisting exclusively of uber-geeks, with the exact right timing (wasn't possible before, and someone else would have done git hosting later - and did, but less well), in a winner-takes-all market (which, to be fair to them, they engineered - source code hosting didn't use to be a winner-takes-all market, but github's social features made it so). Moreover, Github lent itself very well to a viral spread model (doesn't work in most cases, and is hard to bolt on to an unsuitable product), and so had less need for sales efforts than most B2B products.
I'm not putting down Github's team (they executed brilliantly), or even the lessons that they garnered - those are worthwhile lessons to pay attention to and learn from. However, be careful when translating those to your startup, which is probably a very different beast and may crash and burn if you just apply Github's lessons directly!
I obviously completely agree with you (being a long time reader of your blog). The point is not that Github is a special case, but that all startups are a special case, and every founder has to figure out what works and what it doesn't for their own.
I have spinned up the contrarian touch a bit in order to exactly get this kind of reaction, and to raise this conversation, which I don't think we have enough.
Github is a product for geeks (few of those take off), build by a founder team consisting exclusively of uber-geeks, with the exact right timing (wasn't possible before, and someone else would have done git hosting later - and did, but less well), in a winner-takes-all market (which, to be fair to them, they engineered - source code hosting didn't use to be a winner-takes-all market, but github's social features made it so). Moreover, Github lent itself very well to a viral spread model (doesn't work in most cases, and is hard to bolt on to an unsuitable product), and so had less need for sales efforts than most B2B products.
I'm not putting down Github's team (they executed brilliantly), or even the lessons that they garnered - those are worthwhile lessons to pay attention to and learn from. However, be careful when translating those to your startup, which is probably a very different beast and may crash and burn if you just apply Github's lessons directly!