GitHub is essentially a web interface to a single open-source DVCS. The switching costs to using another DVCS such as Mercurial (which I personally prefer) or another Git hosting provider are minimal. There are no barriers to entry since Git is and always will be free. There are very few network effect benefits to using GitHub for paying customers who by definition want to keep their code private.
GitHub is a good service and I use it myself, but I'm now increasingly cautious about using it since GitHub are now under serious pressure to deliver pretty spectacular returns to their investors.
In the end I think this investment has less to do with GitHub and more to do with the fact that e.g. yields on 10-year Treasuries are currently under 2%. Investors are desperately looking for any decent returns on their money and throwing it a company like GitHub is just a Hail Mary pass.
I don't think GitHub is popular because it's just a web front-end over a DVCS.
I'm basically a mercurial user. Use it for work. Used it for all my personal projects. The commands and workflow seem much cleaner and make much more sense to me than git's. I switched to git for personal projects soley because of GitHub.
Partly because it has a very good user experience and is snappy (bitbucket is slow in comparison) but mostly because that's where everyone else is.
I'd go as far as to say that GitHub has done more for git than git has done for GitHub.
There is no WAY that anybody ever loved Pets.com the way that I love GitHub.
And I'm not the only one! Not by a long shot.
That represents 476k 'micro' subscribers in one year, 277k 'small' subscribers, 151k 'medium' subscribers. Cheapest business plan (bronze) would require 133k; most expensive (platinum) would require 16.7k subscriptions. Github announced 1 million users on 21 September 2011. I can't find any details about current #s or the % of paying users.
Remember that all these numbers double if GitHub only has 20% net profit (bear in mind this would be 2.5% higher than the insurance industry, which is an extraordinarily predictable and profitable industry, and oil/gas, the most profitable industry, sits at 25%).
So while not 'spectacular' it's not at all obvious whether they can generate the sales required to generate the return necessary, and the optimistic attitude of people posting here without any numbers to support their conclusions is a bit disturbing (not a reflection on parent's comment, but in general).
Looking at the numbers I'll make the following observations:
- the total market size of potential GitHub users has a fairly low limit (somewhere in the 10s of millions I would guess)
- GitHub has some popularist advantages but ultimately competes on price as far as I can tell (I've started moving to Bitbucket because I see no point in paying for something I can get for free, and the recent hack shows GitHub doesn't offer better security AFAIT); as soon as they gain a larger market share there's no reason to believe competitors won't want in on the action, and as GP points out, switching repository hosts is really easy
- while it looks like they'll be able to generate that 2% return over the next few years I'm struggling to see them generating a much more interesting return based on the numbers I'm looking at
- I have a feeling people are confusing feelings of "GitHub is cool and I like it" for implying that "GitHub should do well as a profit generating business for its investors"; they are also probably fantasizing about having $100m invested in their own startups and taking criticism as personal attacks on this 'plan'
It should go without saying that Treasuries are risk-free since the government can print dollars to pay back its debts. But in this economy, investors are willing to get negative real returns (i.e., after inflation) on Treasuries since they're more concerned about the return of their money than returns on their money.
(For fairly approachable commentary on the current macro investment climate, I'd recommend reading Bill Gross' monthly columns at http://www.pimco.com/EN/Insights/Pages/InvestmentOutlookOver...)
It doesn't look to me like the people who love and use github have been using it because it looks like it will be stable, and stick around for a while. They seem to have been won over by popularity effects and performance/cool tech factors.
So, if github fails, it may not really be as big of an issue. Github could fail as an investment or company, and still leave a running system behind, up as a non-profit even; the risk to its customers doesn't seem that big. And the potential rewards, depending on how they invest the money, can be really big. There's lots of good work they can still do, that everyone will benefit from.
If anything, I suspect this investment is to make Github more competitive with Bitbucket -- remember, BB's parent company is Atlassian, which has pretty deep enterprise experience they can leverage. You may like GH's issue system better than BB's free one, but a paying BB customer can integrate it seamlessly with Jira and other Atlassian tools.
Does GitHub do some things better than Bitbucket? Sure, but I don't believe that those differences justify a valuation approaching a billion dollars.
Maybe I'm missing something, but does GitHub support something other than Git?
> The switching costs to using another [...] Git hosting provider are minimal.
$ git clone git://github.com/jbarham/stathat-line-counter.git
$ cd stathat-line-counter/
$ git remote add bitbucket ssh://firstname.lastname@example.org/jbarham/stathat-line-counter.git
$ git push -u bitbucket master
> There are no barriers to entry since Git is and always will be free.
> yields on 10-year Treasuries are currently under 2%
I've worked in investment management. I know that investors with $100 million to throw around don't have a lot of good places to put their money. No pressure. Good luck.
The wrong part of that sentence is the word essentially.
GitHub is essentially a social/collaboration app for people who spin code.
The fact that it uses the git distributed version control system is an implementation detail.
The essence of GitHub is in the network of people who use it and how easy it is to do so. Not in the fact that its a (really really) good front-end to a particular tool.
I'll admit that I'm unenthusiastic about social media in general, so this aspect of GitHub is the one I use the least. I've always found their slogan of "Social Coding" to be a bit of an oxymoron as programming for me is primarily a solitary activity.
Generally I find the design of GitHub to be more refined than Bitbucket (e.g., the README's look better) but it's not something I'd pay for.
I'd argue that the real action for "social collaboration" on open source projects is where it's always been: on mailing lists. The project I've followed most actively the past couple of years has been Go, which is hosted on Google Code, but I very rarely need to go there since I'm on the mailing list.