This final presentation I attended at Computers in Libraries 2013 was one of the best. Jeff Wisniewski of the Univesity of Pittsburgh presented this talk on how to effectively use Google Analytics to get relevant results. It tied in beautifully with the metrics session I attended on day one and I wonder why they weren't in the same track. Both emphasized the need to have statistics that mean something, and to get statistics that mean something you need to establish clear criteria of success. Establishing clear criteria of success means you need to model out some set of individuals whom you are trying to target, either with your statistics or with your website, or for both.
In the case of Google Analytics many people (myself included) grab some numbers that look good and chart them out for reporting purposes. Most of the time there is some bureaucratic agency that is asking for those numbers, so you make sure you have them so they will be happy. However the numbers that are skimmed off of Google Analytics can be flawed to the point of uselessness. Unique visitors can be skewed by individuals who use different devices or clear their cookies. Average time on page can be skewed by one person who spends an age there because he got up to go to lunch or something. Bounce rate may be important or it may mean nothing. If you have a page redirecting people away from a website (like to a catalog or databases) you might really want a high bounce rate since you're running the website to provide access to those services.
Instead of relying on these canned numbers which are easy to obtain but likely meaningless, a better approach is to carefully determine what you want to look for and then figure out how you're going to get it. Wisniewsky used the acronym of S.M.A.R.T. to enumerate the criteria your goals should meet : Specific,
Measurable, Attainable, Relevant, and Timely. To get these goals we need to talk to users, find out what broad types our users fall into (high school students going to the web page for research, moms going to the web page to find out storytime hours, active citizens who want to know information about how their money is being spent, etc.) and then figure out where we want those people to be going.
Google Analytics can be used to map out pages you expect these categories of user to want to go to and the paths they are likely to follow on your website to get there. These pages and paths are called "goals" and "funnels" respectively. You can ask Google Analytics to provide reports for you then on how many people successfully navigated a path and attained a goal, and how much time they spent on a goal page, if the goal page is some kind of destination page that is worth spending time on. Following these steps will lead to a more comprehensive and useful analysis of your website that you can use to make it better and you can figure out what kinds of people are using your site and in what volume.
Then you have useful numbers that mean something, and the state can have the number of hits you had in 2013 if it really wants to know.
No comments:
Post a Comment