This session, lead by Sebastian Chan of the Powerhouse Museum was packed to the brim, and well worth it. Seb gave a great overview of Google Analytics, as well as other stats tools and talked a bit about how to use them to create a big picture of visitor analysis. My notes are not very organized, but might prove useful for some.
The Search “Problem”
Google deep links.
Now google has search inside this site right there on front of results page, when you hit a high ranked site right away (search for Powerhouse museum). Not doing this to the Walker Yet.
Google hi-jacks these search results, shows ads, can hurt.
When someone enters from a deep link off google, hard to gauge intention from visitors.
Funny, jobs seems to be a top result on many museum sites.
Retailers hate this, they want to funnel your experience through all their ads. (Amazon wants to try to sell to you at all points, google thwarts this)
Traditional Metrics
Come out of the advertising world
visits, pageviews, time on site, etc
Ask people what they’re using
Almost everyone is using some sort of javascript based client side solution
Look at google analytics
13k visits, if were using a logfile based solution, would be vastly inflated by bots, spiders, harvesters
Javascript tagging solution resolves some of this
When switching to a logfile based solution, with inflated numbers, going to tagging solution, hard to explain big drop to the marketing department or the director
Benchmarking tool in google analytics. Interesting, haven’t looked at this before.
Demo of how unique visitor counting works.
Time on site: Seb showing a site with time on-site of 0:54 seconds. Break it out, 23K of visitors show a time on site of 0-10 seconds.
How many visitors have a depth of visit (Number of pages visited) of 1 page. Tons! Every visitor that hits only 1 page, they count as 0 seconds.
If you deduct the people that only hit one page (getting rid of 0 sec visits), you see that the median time on site visits are much better.
Looking at content drilldown, now looking at comparing regions and how to use drilldown.
Site search overview. Google can track search terms on your site, whether or not you’re using a google search on your site.
Traditional Metrics.
RSS feeds are not hit in GA, but can in logfile analysis. But problematic, because RSS hits are not really visits.
Email metrics
Useful to gather addresses of those who read email, take those addresses, and use them as seperate sub-group that are receptive.
Do the same, with those who clicked through, can drill down to specific exhibitions, etc. Allows very direct marketing to receptive people.
Realtime Metrics
Reinvigorate. My buddy paul loves this. Lets you see see visits in real time, what they’re searching for, etc…
Looks like tons of traffic from google. Wow!
Lets you expand visits, etc. Really slick.
Lets you pick up trends very quickly.
Outside metrics
Quote from Seb: “You’ve heard of Alex? Alexa is crap. Don’t use it.”
Instead, use AttentionMeter. But, it’s US only. Free!
Works that ISPs sell anonymozed logs, attentionmeter uses that for data.
Looks pretty nice.
Quantcast lets you look up similar sites, how you compare, etc.
Compete.com is another one worth looking at. Compete is better than Quantcast, according to Seb.
New Metrics
Flickr Metrics
For powerhouse, metric is not views, but how many things were tagged. Part of their goal with flickr commons, getting people to tag stuff.
Technorati Authority: Dubious, not all that great. Only measure blog links.
egoSurf.com: lets you check the “ego score” of a person and/or a site. Neat.
Domaintools: Lets you look up how much you’re being used as a reference in Wikipedia, amongst other things.
http://whois.domaintools.com/enwikipedia/walkerart.org
Side notes:
Tabbed browsing messes with things. if you open a tab, how accurate is it when you open a new tab and come back to it later? Not very accurate.
Question I asked about RSS Metrics:
Look at people who are nerds.
Captures a lot of machines, too.
Best metric is people who use feed my inbox.
Edit: Also some great notes here.