Comments on: Systems that Get Better the More People Use Them

By: Harry Miller

Harry Miller — Mon, 14 Jul 2008 19:17:45 +0000

Microsoft gets user input different ways. Each product is pretty much in charge of its own system, though there’s communication between divisions. Office has the best feedback mechanisms, because their Help is almost entirely online–they can get the usual Web statistics and easily collect comments. In Developer Division where I work (Visual Studio, .NET Framework) we have offline, local Help plus an online version on MSDN. Online, people can rate topics and leave comments. The writers look at the ratings and they also get every comment assigned to them in our work item tracking system (though this is pretty recent).

We also have the MSDN Wiki where people can add comments. Last February I went to a conference and videotaped interviews with our customers talking about how they used the product and how they got Help.

Harry Millers last blog post..Prioritizing for the Reader

By: bursa reklam

bursa reklam — Sat, 14 Jun 2008 07:47:45 +0000

thank you…

By: Tom

Tom — Fri, 13 Jun 2008 13:03:15 +0000

Kai, Thanks for your comment. By the way, this time I think you unchecked the Commentluv check box, and so that faulty link didn’t get pulled it.

Re Microsoft, supposedly they track user behavior (if you allow it). That’s how they discovered that most people only use 20% of the features in Microsoft Word. They used this behavioral analysis to come up with the Ribbon.

It’s not an automated example, and it leans more towards user analysis. The better web 2.0 examples are when a system automatically gets better from user additions.

By: Kai

Kai — Fri, 13 Jun 2008 07:15:00 +0000

Tom, thanks for raising some very interesting questions!

I guess part of the charm of O’Reilly’s definition is that it implicitly conflates use and care (rather than wear). Using a Web 2.0 system, it gets better because users take care of it while using it, whether they contribute actively or indirectly, automatically. (Never mind remuneration for these services… ).

And speaking of music: Yes, a canned music track doesn’t improve with use – but a song will when it’s played live and exposed to a community of fans and caretakers. I’m thinking especially of jamband phenomena…

When it comes to making documentation better with use, I think the ideas you point out are all practical in their own contexts. And I would add good old interviewing users to see if they find the help useful and how it could be improved as well as tracking click paths through a system to see what paths users actually take.

However, I think on a larger scale it has less to do with tools and techniques, but rather with the users and how they can achieve what they want or need to do. So maybe it’s as simple as *allowing* users to improve the documentation to their needs. I’m still feeding off of Scott Abel’s idea of allowing users to help one another at 2:30 in the morning.

I would also argue that O’Reilly’s examples translate badly to most of us technical writers: Google, Amazon and eBay all have sooo much traffic to play with and try out all kinds of improvements. For every example mentioned, I’d bet there’s five that have fallen by the wayside because they didn’t work due to failing technology or faulty premisses.

Speaking of oodles of users: What does Microsoft do to have their users improve their help? Anyone know?