This document discusses analyzing website data and metrics to better understand user behavior. It recommends segmenting content by type and user goals. Qualitative data like user interviews and personas can provide context to interpret metrics. Goals and funnels should be based on these user tasks. Different metrics may apply to different content like engagement, informational or funnel pages. Visualizing data through reports on traffic sources, landing pages, and top pages can help communicate results.
10. BEWARE THE
OUTLIER
Average Time on Page:
This is the average amount of time that all users spent on the
specific page within the date range.
Simple, but remember this information can be easily skewed
by the extreme boundaries of the results.
For example, say you had 10 people come to a Web page and
this is how long they stayed on the page:
11. BEWARE THE
OUTLIER
Person Time on page
1 2:23
2 2:12
3 1:39
4 40:44
5 1:49
6 2:01
7 1:33
8 2:09
9 1:17
10 1:56
12. If you add up the total amount of time you get 57:43 and then
divide that 10 and you get an average time on page of about 6
minutes and 3 seconds.
13. THE BOUNCE RATE
MYTH
We link to a LOT of third party sites from our websites:
Catalogs
Discovery tools
Overdrive
Databases
LibGuides
Book review sites
14. THE BOUNCE RATE
MYTH
A user who visits your homepage and then visits one of
these places has
15. WHEN THE COOKIE
CRUMBLES*
New vs Returning users relies on cookies.
âą Up to 66% of users user multiple devices
âą App. 30% of users deleted cookies deleted in a one month
period
*I stole this title from the comscore article of the same name
because I liked it
http://www.comscore.com/Insights/Blog/When_the_Cookie_Cru
mbles
19. FORBES MAGAZINE
SAYS:
âIf You Think A Web Analytics Tool Is Enough, Think Againâ
(and if Forbes says it it MUST be true)
http://www.forbes.com/sites/davidwilliams/2012/04/16/if-you-think-a-web-
analytics-tool-is-enough-think-again/
21. YOUR SITE
Needs to have SMART goals
Specific
Measurable
Attainable
Relevant
Timely
22. THESE GOALS SHOULD
COME FROM TWO
PLACES
Business objectives
âą Yours or your parent organizationâs mission and goals
User goals and objectives
âą But how do I know what users want?
25. USER INTERVIEWS
AND PERSONAS
A persona is a fictional person who represents a major user
group for your site.
Personas help you identify major user groups of your Web
site.
You select the characteristics that are most representative of
those groups and turn them into a persona.
26. USER INTERVIEWS
AND PERSONAS
Interview representatives of you key demographics
A. Can you take me through your typical online routine? You
open up your web browser and where do you go first?
B. E. Youâve been given an assignment to write a term
paper on climate change with at least 4 sources cited
throughout the paper.
C. i. What is the first thing you do?
27. USER INTERVIEWS
AND PERSONAS
âą How often do you go to the library website?
âą When do you go to the library website? What do you do when
youâre there? (Take participant to site)
âą What immediately draws your attention?
âą What features do you like/dislike?
âą Did you find the Website easy to use?
âą Why or why not?
âą If not, what would make the site easier to use?
âą What information did you expect to find but did not?
What links are the most important and should appear on the
main page?
28. USER INTERVIEWS
AND PERSONAS
Use the answers to develop personas that represent your users
Use the personas to develop a list of critical tasks for each
demographic
29. STACEY
freshman
the source of the information can come from
anywhere and is really irrelevant; she just needs
to know where to look
uses often commercial search engines like Yahoo
and Google to locate information
she doesnât distinguish between the online
catalog and the library web site; for her, the web
site is just the way to get to the catalog with some
additional information mixed in
Unaware of library tools and resources
needs help with the research process and is
overly-confident about her own research skills,
which may not be as strong as she thinks
30. spends a lot of time at the library to study for
midterms and finals; wants to be able to check easily
what the library hours are
she wants to easily connect to wi-fi and uses self-
print for all her assigned readings
does not rely heavily on the libraryâs print journal
collections; commonly uses e-journal collections
uses primarily books on reserve and rarely checks
out a book or two from Hillman collection; expects
Hillman library to have all the textbooks required by
instructors and is surprised if they are not in the
collection or are not placed on reserve for students
uses chat or calls the front desk to get help from
library staff with locating a reference
uses her smart phone to find the call number of a
book on reserve, see if a book is checked out and
when is due back, or look up the syllabus, the course
number, or the name of the instructor
31. USE PERSONAS TO
IDENTIFY TOP TASKS BY
AUDIENCE
What specific things does Stacey want
to do on our site?
Set up GOALS and FUNNELS around
these tasks
32. GOALS
Goal : the page a visitor reaches once
they have completed a desired action
Funnel : The (optimized) steps along
the way to the goal
38. GOALS Name the goal something
intuitive. In this example
it might be âClass
Registrationâ
Choose
whether or
you want
the goal to
be active
(on) now
Most library scenario goals
will probably fall under the
âURL Destinationâ type
40. *TIP
Exact match has to be EXACTLY the same as the URLâŠ.even
leading or trailing spaces will cause it to fail
41. add the URL(s) of the page(s)
along the path a user would
GOALS
take to get from the
homepage all the way through
to the thank you page
42. MAYBE DIFFERENT
METRICS FOR
DIFFERENT CONTENT
Set up different profiles for:
Engagement content (blogs, guides, reviews)
Funnel content (get me toâŠdatabases, ejournals, catalog,
etc)
Informational content (hours, policies)
43. MAYBE DIFFERENT
METRICS FOR
DIFFERENT CONTENT
Time on page:
Engagement content (blogs, guides, reviews): hopefully high
Funnel content (get me toâŠdatabases, ejournals, catalog,
etc): hopefully very low
Informational content (hours, policies): hopefully low