Unlocking the Mystery of Web Analytics Dave Hodgins Data Coordinator & Web Developer Library Research Service, Colorado State Library /
About Me I am the Data Coordinator and Web Developer at Library Research Service ( I produce data and tools that empower libraries to improve the lives of all Coloradans.
Goals Spark interest in beefing up your web analytics strategy, and take away one or two keywords/concepts that you will research further.
Talk is cheap, but so are web analytics
Web analytics offer you valuable insight into your users’ behaviors, wants, needs, etc.
Web analytics are essential to successful goal/data driven decision making.
We use Google Analytics because It’s free It’s well documented It’s built & supported by one of the largest advertising companies
We use awesome open source tools like Laravel and Bootstrap… …because they make the process of producing cool and useful stuff… …quick and (relatively) painless.
We roll our own analytics because We want granular control over our data We want to accurately track “tricky” stuff like direct file access We can
Server -> MySQL -> Laravel, fronted by Twitter Bootstrap
Custom pivot table in LibraryJobline
Tip: Measurable ~!= Useful Just because you CAN track something doesn’t necessarily mean you SHOULD. Reduce signal-to-noise ratio by designing PURPOSEFUL data collection strategies.
A Brief History of Web Analytics Server-side Server logs traffic Log analyzer (poorly) parses data Analytics are…generous
A Brief History of Web Analytics Client-side JavaScript logs user interactions Data are pre-processed and parsed by Google Analytics are more accurate
The Mighty Google Analytics Tracking Code
A Brief Tour of Google Analytics
Lets Make This Better
Filters Filtered Views allow you to exclude or modify data before they are saved to GA. Since the manipulation occurs before storage, you cannot apply filters to historic data!
Filters Filtered Views are handy when you want to limit a user’s access to a subset of your data. Filters views permanently alter your data, so always maintain a raw, unfiltered view!
Tip: Filter Early and Often While GA provides methods for segmenting existing data, filters are often more convenient; but they cannot be applied to historic data. Anticipate future needs!
Filters Typical filters include: Exclude data that doesn’t belong to your site (apply to every view) Exclude internal traffic Segment by sub/domain Segment by content/services
Filter Options
Tip: Suck it Up and Learn Regular Expressions Regular expressions are like keyword searching using Booleans, wildcards, and placeholders, except with the additional pain of a hot poker to the brain.
Tip: Suck it Up and Learn Regular Expressions Regular expressions let you do things like… Find URLS that begin with 4 digits (e.g. WordPress permalinks): ^/\d{4}
Tip: Suck it Up and Learn Regular Expressions …and do things like like… Find URLS for content in the “data and tools” section of the site: ^/data-tools
Segments Segments allow you to group data without making permanent changes to your GA profile. Multiple segments can be viewed side by side, making comparisons easy. You can create custom segments using the same criteria as filters.
Tip: Filter vs Segment FilterSegment Historic DataNoYes Compare DataNoYes Permanent ChangeYesNo Restrict AccessYesNo
Events Events are user activities that do not generate pageviews. They might include: clicking a link, changing a form value, toggling dynamic content, etc.
Events Events are essential for tracking user interactions in situations like…a single page website, and/or dynamic (AJAX) content
Events vs Virtual Pageviews A virtual pageview is a pageview that you manually create and send to GA. They get lumped in with all the rest of your “normal” analytics data.
Events vs Virtual Pageviews Use virtual pageviews to track clicks and/or links to static files that cannot have the GA tracking code, such as…PDF, DOC, XLS, JPG, etc.
Campaigns A campaign is process for tracking EXTERNAL traffic. In it’s simplest form it is a series of parameters that you attach to URL’s and that are used by external sources.
Campaigns This is your querystring: This is your querystring on GA campaigns: article?page=2& 797/Parttime-25-hour- Concierge?ref= &utm_source=newjob &utm_medium= &utm_campaign=new%20 job%20 s
Campaigns Track external and/or 3 rd party content and links such as… – – Social media – Non-HTML content (PDF, Word, etc.)
Campaigns Campaigns are especially useful for social links because: Links within mobile apps do not sent referrer data Your links will (hopefully) be shared via other sites/networks
Tip: Use Campaigns to AB/Split Test Want to know which version of your newsletter/banner/etc. is most engaging? Campaign those suckers.
Campaigns NEVER use campaigns to track internal links
Campaigns Create campaigns using Google’s URL ics/answer/
Done!