Very interesting... You could then start to crowd source content problem-solving. Even internally, people frequently don't have immediate access to analytics to reflect upon the content they create themselves.

One side of the curtain is already opened - 1.usa.gov (bit.ly) URL shortener stats. It gives insight into sharing of .gov info on twitter/facebook/etc. Opening all of it, will be even more insightful.

This would be useful to have real discussions about the value of websites. Numbers alone are unimportant, but they're a critical assesment tool. This can also identify if different sites have different/overlapping/the same audiences. If there is to be consolidation, it's important to help users of one site transition to resources at another.

That would definitely be a better approach, in my mind, than just setting targets for consolidating websites and/or domain names. Stats could help justify existence. Pair it up with required user feedback for sites and pages (five-star ratings, for example).

this is one of the best ideas i've seen. even if we can't agree on which metrics we should use for performance metrics, it would be fairly easy to "tag" all new .gov websites so that we can collect some universal data from here on out. might be difficult to retroactively implement, but not impossible.

Just have to say that I'm glad Dmitry gave a shout out to 1.USA.gov here. No one had such a broad view of how government information is being consumed before USA.gov released the 1.USA.gov data, and it's been fascinating. The more data like this we can get, the better! I'm wondering, can we open up USA.gov search data (after accounting for privacy issues, of course)?

It certainly has a nice air of transparency to it. It's not hard to imagine that a lot of people would think it was bad for their fiefdoms, but then that's true of any good idea.

Would this be page visits for all of the web pages? How would this be shown? I have suggested that this would be great data to put as a dataset on data.gov - is this what others have in mind?

The open source analytics system Piwik has the ability to collect realtime metrics and make them publicly visible. This might be TOO much transparency, but I'm sure it could anonymize aspects of user data.

At HUD, we posted our stats for years. Anyone could see them and use them. No one did. But I'm still for posting them. My thing about "analytics" is that we need to come up with metrics that matter. And it doesn't have to be 50-60 items. 3-5 metrics that actually show how fast customers find our top tasks, how fast they use them, how many drop out before completing them (so we can figure out where we have problems)...that would be huge. If we could come up with a few standard measures (usability tests, stats that track a user through a task, etc.) and throw in tailored customer satisfaction surveys about those specific tasks...that would be something we could use and something that would show customers that we care about their success. Win-win. This doesn't have to be hard, but it does have to be thoughtful.

A lot of the analytics tools require some bit of code added to the website itself - getting that uniformly added across .gov site would be a challenge. And many of the better ones are fee-based.

However, there would be a lot of value if, say, Google (thinking of Google Analytics, but if any other stats provider could do that, it would be wonderful) could create an interface that would simply capture data from all .gov sites and present that publicly.

It would be valuable for public and govt - we're constantly trying to identify benchmarks to make sense of our own numbers. We look at the online data available from Alexa, Compete, and Quantcast - but that's top level domain only, and it's not always transparent as to where their data is coming from. So I can compare Compete to Compete numbers - but I can't look at the Compete traffic and confidently compare it to Alexa search patterns.

A bad website can generate great numbers by making people go in circles. Search data needs to be incorporated to tell what people are looking for vs. the path to get there.