I have been playing around with Google Analytics for UMW Blogs, and I took a snapshot of the traffic thus far this semester. Since August, 24th until today, UMW Blogs has had half a million page views by 211,000 visitors, 135,000 of which were unique. What’s more, we’ve had traffic from just about every country on the globe.
I’m still thinking about ways to use some of these stats to actually say something meaningful, and while I am entranced by the numbers at times, I’m not sure if looking for meaning in scales and aggregates makes any sense at all—in fact, I wonder if it simply reproduces a model of data collection and quantification that is eating the heart out of assessment in most institutions. Anyway, there you have it, but what it means is something (if anything) else, I’m afraid.
I hadn’t looked at our GA in a couple months. I suspect some obsess over them far more than is necessary. The use we find in them is an indicator of what people read thoroughly, where they came from and what they searched for. So, that way, they are useful. Beyond that, like most stats, they probably are not that useful.
Those are insane numbers. Easily 10x the traffic going through UCalgaryBlgos in the same period.
I’m curious about how many are fed by Google – is the sheer volume of content being published publicly through UMWBlogs making a huge footprint in Google searches, driving hordes of people your way? Are these direct links, from people referring to the activity?
Over 90% of the sites on UCalgaryBlogs are non-indexed, where the blog owner has opted to hide their site from Google and other bots. So the only traffic they get is from people they send links to (their profs, other students, etc…)
I think stats are most useful when they’re showing change or differences.
So for example, you change the site, or add a widget, and the stats change. You can see the effect your intervention made.
Or with differences, if you segment the stats and notice different sorts of behaviour from different user types, which in turn might lead you to make a change to thee site to benefit just a particular subset of users. (The stats might also help you find those different subsets – and they may be different to the groups or user types you’d anticipated would use the site).
And of course, stats can hlp you see where traffic is arriving from, what search terms are driving traffic (and which posts Google might be considering authoritative for a particular topic, for example).
Headline stats – visitor numbers – are only interesting in the sense that they tell you people are hitting the site. Which is to say, thy’re not interesting… (although the numbers you post are impressive š
Our social network at the University of SĆ£o Paulo, http://stoa.usp.br/ (based on an old version of Elgg) gets 4k-8k/visits per day (about twice your traffic). Right now about 3/4 is via Google et al., and that number used to be even higher.
See the presentation embedded in http://stoa.usp.br/destoa/weblog/49289.html for more details (esp. page 7 and 8). I agree that the aggregate numbers are only useful to impress administrators and to track changes (did you see http://trendly.com/ ?).
If you scroll down to the last comment in the above post, you can see an analysis that I did of the “participation” (posts, comments, internal messages, etc.) of users in our system. As it turns out, we see very unequal, heavy tail distributions: a small fraction of users are responsible for a large fraction of the participation. I think these kinds statistics, “distribution of indicators over users” are a very useful complement to aggregate data. For example, you can easily spot your power users and adjust your site policies taking or not taking them into account.
@Ron,
You echo a lot of what Tony says as well, what people are reading, where they are coming from, and how long they stay may be of more import that the aggregte numbers I am obsessing oer. I’m afraid my own aversion has more to so with not knowing how to use them to make the site better, or further make faculty and students aware of what is being hit often as a resource on a topic.
@D’Arcy,
We have the opposite situation than UCalgary blogs in regards to open. i would think over 90% of our blogs are Google searchable and we have seen a major uptake in traffic from Google this semester. At the start of the semester less than 20% of our traffic was from Google, now it is just under 40%—that change in an of itself is fascinating. Referring sites is another 40%, and direct traffic around 20%. So, it seems a majority of traffic is coming from places other tha UMW Blogs, which suggests a strong case for openness when I think about it. But I can;t really gather what these people are doing with it, I think I need to dig deeper. One thing I discovered is that we are getting a ton of people re-using images uploaded to UMW Blogs, we have a ton of traffic from google images. Maybe this rabbit hole is worth getting lost in š
@Tony,
Yeah, you make a lot of good points, and I appreciate you deciphering my broad discounting of stats with some more intelligent approaches. I like the idea of watching change in traffic after we tweak stuff. Marth Burtis has been interested in doing some SEO work to see what happens when we actually optimize the work on UMW Blogs for search engines. Something that would be really interesting to watch.
Additionally, the segments stats is something I just noticed after you mentioned it. I have to find some examples of how people are using that, because this sounds useful and possibly illuminating. As for traffic, I’m finding a lot comes from various search terms that are, indeed, making certain sites “authorities” on topics like banned art, the 2008 financial recession, digital history, Venice, 18th century poetry, and contemporary poetry reviews. All of which Google seems to be making a big push for in various searches. I guess you have confirmed what I have been thinking in this regard, the aggregates say far less than some intelligent slicing and dicing of the information. And I need to read up on how to do this intelligently so that I don’t simply fetishize traffic numbers.
Recommended approaches and resources would be appreciated š
@Rev – The thing with hosting a large number of blogs is that you have very little control of the content that attracts visitors. On homeschooljournal 4 of the top 5 visited posts/pages are over 2 years old. The top one is in a blog that hasn’t been updated in almost 3 years. So, I agree it’s a bit of a quandary as to what to do with the stats.
These may give you some ideas, though they were written before the Advanced segment tools were produced:
http://www.google.com/cse?cx=009190243792682903990%3A2ke2vucb8nm&q=allintitle%3Alibrary analytics&siteurl=www.google.com/cse/home%3Fcx%3D009190243792682903990%3A2ke2vucb8nm
See also:
http://writetoreply.org/actually/2009/10/28/thinking-about-user-tracking-on-writetoreply/
http://writetoreply.org/actually/2009/11/22/measuring-website-usage-with-google-analytics-part-i/
@Tony,
Funny, when I read the most recent post on OUseful, it was all Google Analytics, so it is nice to know we are going through this together. And I can live off your experimentation in this realm š
Should know all link lead back to the Hirst, who has insatiable innovative thirst! Thanks Tony.