Posted by: Denny Cherry
.NET, Social Commentary, SQL Server
If you were on twitter in the last couple of days you may have seen some links to a blog post titled “MySpace’s death spiral: insiders say it’s due to bets on Los Angeles and Microsoft” in which the author tries show that because MySpace used the Microsoft stack (ASP.NET, IIS, Windows, and SQL Server) that this was a major cause of MySpace’s demise.
As someone who used to work for MySpace’s infrastructure team (specifically I was one of the Senior level Database Administrators) that worked there, I can say that it was not the Microsoft technology stack which was MySpace’s problem. What caused MySpace’s demise was a piss poor UI (User Interface) and because of this poor UI design the slow page load times. It really didn’t matter what platform that the site was built on, crap code is crap code.
No before you assume that I’m a Microsoft Fan Boy, I’m not. You should flip through my blog and you’ll see that I’m actually quite critical of Microsoft and specifically the SQL Server product for which I’m a Microsoft MVP. I have a reputation within the Microsoft SQL Server product group of telling them the truth about their products, both the good and the bad.
The Los Angeles Area
The author also claims that they failed because there weren’t enough .NET programmers in the Los Angeles area which could work in the .NET startup mode. This is just crap. I’ve spent most of my career working at startup shops, all of which were .NET shops and there were always enough .NET programmers who could work at startup speed. What MySpace needed was a management team that understood that programming new features requires more than a few hours to do it correctly, and requires a solid scalable test environment to test in. The developers at MySpace weren’t given either one of these to work in. While the production database environment (for example) had several hundred database servers, the testing environment had only a single database server. Rolling changes wasn’t really tested before they went into production, and there was NO load testing that I saw at all. The basic excuse was that they couldn’t generate a production work load, so they didn’t load test it.
A major failure of the MySpace management was that they couldn’t retain the talent which they had. They didn’t appear to value the employees which they had. They were constantly hiring .NET developers and they were losing .NET developers just as fast if not faster than they could hire them. In the 6 or 7 months that I worked there I was the 3rd or 4th DBA that left the company. I’ve never seen DBAs come and go as frequently as they did there. During my tenure at MySpace I personally never felt like anything other than another cog in the wheel, with the exception of my last few days when they begged me to stay on board.
One of the points in the article which is valid (sort of) is that Los Angeles is a large area, so commuting to the office can suck (it would take me at least an hour to get to work, longer if I took mass transit). There is an easy way around this, allow your employees to work from home, and this problem goes away. However MySpace (more specifically the parent company Fox Interactive Media) has a strict no working from home policy. Killing this one policy would have made working there a lot easier as the employees could have lived anywhere in the area, only coming to the office when it was absolutely needed.
Were there problems at MySpace? yes.
Was the problem Microsoft? No
Was the problem being in Los Angeles? No
Was the problem the management at MySpace? Yes
Could these have been overcome? Yes
Will they be, allowing MySpace to make an amazing comeback? My guess would be probably not. At least not without a major house cleaning.