Gotta applaud this Chinese firm for using scientific experiments rather than gut feeling to decide on important policies.
However, please don't take far-ranging conclusions from this experiment. Call-center work is a fairly specific type of work with extreme values of "noise" and "distraction". For other kinds of work, the results might end up vastly differently.
(Also, the Chrome spelling correction sucks. Just sayin'.)
I was looking through it hoping to find some amount of detail on collaboration from home (Skype, Google Docs, etc). I guess call centers don't need this quite as much as engineers would. I'm not sure how applicable this would be for a job like I feel many corporate types on HN (like myself) work. Programming (or other engineering) at a startup is different from programming (or other engineering) at a corporation, and I need significant face-time with my coworkers and people outside of my team just to get the most basic work done. As an anecdote, I get significantly less work done on the one day a week (today) when 90% of my coworkers work from home just due to the fact that I often need to collaborate with them.
(Also, I've often wondered why Chrome's spell checker can mark a word as wrong, but when I copy it into Google search, Google knows it's spelled right. Should be the same system, I would think. The word uninstall comes to mind, I see that red squiggly right now.)
It's interesting that you say that because I have the opposite experience. When I'm at the office I find that people around me talking causes a significant distraction. I use noise-canceling headphones almost all day at work just to block out the noise.
In addition, I sit right next to a QA guy that works on my product and I can easy waste half a day dealing with his issues and questions.
I find that I get significantly more work done while at home because I don't have the distraction of coworkers tying up my time. I also find I tend to put in at least an hour more a day because I work when I would normally be commuting.
It really just depends on how much your job requires collaboration with other people, and what tools your company employs to make this possible. At my job, the first one is a lot, and the second one is next to nothing.
However, please don't take far-ranging conclusions from this experiment. Call-center work is a fairly specific type of work with extreme values of "noise" and "distraction". For other kinds of work, the results might end up vastly differently.
(Also, the Chrome spelling correction sucks. Just sayin'.)