by Beth Kanter
http://beth.typepad.com/beths_blog/2007/04/screencast_trea.html
I have been knee deep researching and thinking about Web Analytics in general and Google Analytics in particular for a third screencast in a series I’m doing for NTEN. (The other two included my tagging and widgets
screencasts). The research has been going slowly — partly
because web analytics is a very complex and geeky topic. I’m
phobic of anything that might remotely appear to involve math or that
makes me feel stupid.
Here’s my first draft or what might be called a “treatment.”
I would appreciate any suggestions, improvements, pointers to
other resources or if your organization has a story to tell about web
analytics.
Possible Working Titles:
1. Analytics This!
2. Web Analytics As Simple Gifts To Measure Mission
3. Zen and the Art of Web Analytics
Scope:
This screencast will demystify web analytics and use Author and Analytics Evangelist Avinash Kaushik’s mantra of simplicity to illustrate some of the useful features and reports in Google Analytics, a free but powerful web analytics software tool. Kaushik’s thinking is based on “Occam’s Razor Principle” (which boils down to a poetic way of saying “KISS“) The screencast will show some practical examples of how at least one nonprofit organization’s web site, the Idealist,
is using the software in practice. The screencast will include a
companion wiki with resources to aid further explanation.
If you’re curious about Occam’s Razor, here’s a more detailed description or see the simplicity page in wikipedia.
I was happy to discover Laura Quinn’s recently published and very good article on TechSoup, “A Few Web Analytics Tools.”
She not only gives an overview of the different choices of analytics
software tools available, but also provides basic definitions for the
data one might collect. Since this screencast will only
show how one tools works, Google Analytics, her article provides the larger context and summarizes the pros/cons of the complete range of tools.
Screencast Audience and Learning Objectives
The
audience for this screencast is nonprofits that know they need to
analyze the success of their Web sites, but aren’t sure where to
begin. Or they might have some understanding of metrics and
have gone as far as setting up a free google analytics account, but are
not sure what to do next. The nonprofit’s web site isn’t overly
complex and has a clear marketing strategy in place. The
nonprofit will most likely not have a full-time IT or web manager staff
person. Or, if they do, the nonprofit will not have a
full-time Web Analyst, although might work with a web analyst on a consulting basis.
Learning Goals
- To understand the definition web analytics
- To understand how to identify actionable reports from the complex sea of information collected in an analytics tool
- To demonstrate how to get started using Google Analytics, navigate filters, and use goal-setting features.
- How can Google Analytics reports data help improve your Web site’s performance
- To show a practical example of how Google Analytics is used by a nonprofit organization
What follows is a very rough draft for a script.
Act 1: Analytics This!
A: Definition
It was not too long ago
that no one understood what that term meant. Let’s take a
look at the “Official Definition” from the Web Analytics Association, an association of web analytic professionals. (They host Web Analytics Wednesdays around the world and you can hear the definition described by their chapter in Brussels)
Web Analytics is the objective tracking, collection,
measurement, reporting and analysis of quantitative Internet data to
optimize websites and web marketing initiatives.
What is the visual metaphor? Pilot’s dashboard?
It’s a process as Bruce Clay outlines here and I have simplified here:
GRAPA
-Step 1: Goals for the web site as guides to the data collection
-Step 2: Research questions to frame your data collection (why/what around your outcomes)
-Step 3: Analytics software tool to collect data
-Step 4: Pick reports to answer your research questions
-Step 5: Action that improve your web site performance or marketing campaign effectiveness.
Hmm .. I think of Grappa, a
fragrant grape-based pomace brandy of between 40% and 60% alcohol by
volume (80 to 120 proof), of Italian origin. In Italy, grappa is
primarily served as a “digestivo” or after dinner drink. Its main
purpose was to aid in the digestion of the heavy meals. Grappa may also
be added to espresso coffee to create a “coffee-killer” The espresso is
drunk first, followed by a few ounces of grappa served in its own glass.
Resources
Defining Web Analytics by Neil Mason
How To Use Web Analytics, Part 1 by James Maguire
Top-Ten Web analytics blogs from Avinash Kaushik
B: Coffee Machine or Plantation? Why Simplicity Matters
“Web Analytics packages are sold as if it’s an
automatic coffee maker. In fact, it is more like buying a coffee
plantation. You can still get your cup of coffee (eventually),
but your going to have to stick your hands in a lot more manure than
you ever knew.”
Tom Cunniff – Yahoo Web Analytics Forum
C: What happens when you look at the software as the end, not the beginning?
If you jump into using any web analytics tool without the above
5-point process, you are likely to get overwhelmed with
data. You won’t be able to think. You can’t make meaning.
You can’t find the forest through the trees. There are no
insights. (Needs photo of someone holding their head or screaming)
So,
how to deal with the complexity? Well, get to that in a
minute. You’ll face that problem with any tool you ultimatley
use. So, let’s select a tool first.
Act 2: Selecting An Analytics Tool
There are a number
of web analytics software tools available ranging from simple web
counters, web hosting stat tools and more powerful (and expensive)
software packages. Google Analytics offers more functionality than
your typical web site counter and even better yet it’s free.
Laura Quinn, in her recent article on TechSoup, notes:
Because
the Google Analytics package is in an indefinite beta stage, some of
the experts we consulted with cited occasional problems. Several
reported difficulty in getting Google to show up-to-date stats, while
others noted a very occasional loss in historic stats for an entire
site. Google’s customer service supports this product primarily through
automated emails, so you may have little recourse if you encounter
problems. The method by which this tool monitors traffic results in
lower numbers (such as fewer visitors, and fewer page views) than some
other methods. Also, keep in mind that Google offers its product for
free because it makes money by watching you; by using Analytics, you’re
agreeing to let Google store your information and use it for aggregate
reports.
Nevertheless,
Google Analytics is widely used and widely liked. If you’re building a
new Web site, or have a bit of HTML knowledge, Google is a great free
option for surprisingly robust analytics.
(Would be great to get a voice over in Laura’s voice)
The Official FAQ: How Does Google Analytics Help Me?
While setting
up Google Analytics account is quick and painless, using some of
advanced features isn’t all that simple or easy to do if you’re not a
certified Google Analytics partner, one of the outside consulting companies that
provide tailored professional services for using the product.
So,
if your nonprofit web site has complex tracking and analysis needs, you
may need to work with a specialist consultant to set up your account and
train staff to use it. Several of the larger nonprofit organizations
we interviewed for this screencast have gone that route.
Act 3: Google Analytics That! An Introduction
You can start exploring google analytics to see if it is right for
you and we’re going to use to demonstrate some of the simplicity
concepts offered by Avinash Kaushik, author of Web Analytics: An Hour A Day. Note that the proceeds from his book will benefit the Smile Train and
Doctors Without Borders to assist in their efforts to make our world a
better place.
Show the sign up procedure and how easy it is insert the secret code
to get started tracking your Web site’s statistics.
It’s deceptively easy though and it is important to get your program set up correctly the first time. Google Analytics: Get it Right the First Time by Michael Harrison
This YouTube video shows how to exclude your internal traffic (min 2:13) by using a filter.
The 3 Most Frequently Asked Google Analytics Questions
Act 4: A Quaker’s Approach To Using Google Analytics
(Production Note: I wonder if “Simple Gifts” is in the public domain?)
Avinash Kaushik says that some reports are more important than
others and it comes down to a few essential questions. All the
metrics you collect, ask the So What? Test three times and if doesn’t
lead to action, you are wasting your time! (Recycle my slide from my training webinar to create an example – get an absurd example meaningless data from google analytics)
(See Six Key Reports in this interview as possible alternative framework with integration of metric definitions and identify pages in book where the how-to is.)
1: How many visitors/visits/unique visitors during a given period time?
These metrics are important because every other metric you need will be based on one of these.
Let’s define these:
Visits. This metic
shows the number of visitors to a particular site or page. This
is a one time or one-session browsing a web site. It doesn’t mean
you’re unique, it doesn’t mean that you haven’t been to this web site
before – maybe in the same day! Closing your browser or
leaving the site ends the visit. It’s arbitary how much time has
to pass before it is considered a new visit, half an hour is often
used. (Show the browser view of what one visit to a web site
looks like; show what two visits look like and show what three visits
looks like.)
Unique Visitors. This is the number of site visits
by different users. It is typically determined by a cookie or
your IP address or a combination. It isn’t a perfect science, it
is an estimate. It is really important to understand how your
software tracks unique visitors.
Let’s say I visited the NTEN web site 4 times, Alan Benamer visited
3 times, and Laura Quinn visited 5 times. We’d have three
unique visitors and 12 visits.
These statistics let you answer some basic questions:
- How many visitors are visiting our web site?
- What is the overall trend in the number of visits for each page of our site?
- What is the overall trend of overall visits to our site over time? (Week, Month)
The trend in the overall number of visits to your site over time can
give you insight into your site’s popularity. Comparing the
number of visits to each page is also a good way to identify which
parts of your site are most useful to your visitors.
2: From where are your web site visitors coming from?
Understand effectiveness of your acquisition strategy.
The more you know about your web visitors intent the better.
For example, you can check the referring url and look at the messaging
there. Understand what lead the visitor to your site.
Look for surprises in the referring report.
- Do the expected sites show up?
- Who are the uknown friends?
For approximately 60%, you should see a specific url. The
other 40% may be coming from email or direct. It’s a great
visiting card that someone is bringing to your site. Also look
at your unknown friends. Are there links from blogs?
Are there sites creating links from the goodness of their hearts?
Follow those links and find out who they are. Why are they
sending you traffic? What is the call to action? If it
someone you don’t have a relationship with, maybe this is a good time
to do that.
What we are seeing here is a shallow dive. The next step is to
do a deep dive where you can go into each web site that is referring
traffic and see what the quality is. How long are they staying on
your site?
- Inferring intent from search keywords
What percent of my traffic is from search engines? And, is it
enough? According to research, approximately 80% of web users use
search engines to find sites. What is your benchmark? So, how
does your compare?
Are the expected search engines showing up? Are we
overleveraged on one search engine? If so, take the proper
strategy to make sure you’re showing up on the search engines.
Take a look at the keywords. What are the quality keywords that are bringing traffic to my web site? (long visits)
3: Focus on what your nonprofit wants your web site visitor’s to do:
- Why does your site exist? Helps you focus on the data you need to answer that question only.
- What are your top three strategies?
- What should be happening on your web site?
- How are you acquiring traffic?
- What are the customer problems you are trying to solve?
From these questions, you need to pick the three most critical
metrics to measure success goals. The goals provide the
critical context to understand the performance of your web site.
Show example of Idealist tracking the different language versions and geographic distribution
4: Understanding web site visitor behavior
How do they get in?
You will be using a “top entry page” report. The analysis
questions are “What’s the real value of the home page?” What
percentage are actually entering from the top? What page is
creating the first impression of the site What are myvisitors
looking for? It is also important to understand where
people are entering from search engines.
What content do they consume?
This is the “top viewed” pages on the web site. You look at
the 10-20 pages. You will be shocked at what you find here. You
want to know what content your visitors are consuming and is this the
content you want them to consume? The site overlay report can
help you learn what people are interested in.
Page Views. The number of times any page was viewed
regardless of who viewed. (Show a date range and indicate that all
those views could, in theory, come from the same person clicking over
and over again.) This is an important metric if, for
example, you’re using GoogleAdsWords, which pays based on a percentage
of page views which result in a click through for an ad is what pays.
How do they navigate? What’s catching their fancy?
Show the funnel report from idealist.
More detail
ClickTracks Web Analytics Education Series, Virtual World: Feb 26 & 27, 2007
“Unleash the Power of Web Analytics.”
While written from a business or commercial perspective, Eric Enge has an article called “9 Ways to Make Money on Analytics”
which includes tips and suggestions for thinking about how the data can
help you think through improvements that may lead to more donations or
revenue.
Web Analytics Course from Site Logic
Interview with Dave Amos, Idealist
How does the information from Google Analtyics inform your decision-making re: web site campaign?
We’ve
been using Google Analytics since November 2006, so I would still call
us new to the software. Because we’re still new, I think we’ve only
started taking the statistics it’s collected, determining what they
mean, and acting on that information.The most obvious thing we can use is the very regular pattern of
the week. Mondays or Tuesdays are always our best days (unless there’s
a holiday), so if we want a timely blog post to see the most people in
the shortest amount of time, it makes sense to post it in the beginning
of the week.
I also keep fairly
good track of the languages of visitors of our site. We have an
extensive volunteer language program trying to translate the site into
as many languages as possible. It’s nice to see which languages are
good candidates to be added next, or to see if the work done thus far
has seen an audience.
Our
resource centers (like the Nonprofit FAQ, HR Resource Center, etc.) can
also use the statistics to improve the quality of their site structure.
If they see that lots of people are clicking on a particular section of
the resource center, they can decide to make it even easier for a
visitor to find, since it is obviously popular. On the other hand, if a
great resource exists but nobody seems to go there, they can try
highlighting it in a different way to determine if it was a findability
issue or the content just isn’t as engaging as they thought. 🙂 This
hasn’t happened yet in practice, but it’s something we’re looking at in
the future.
GA, and stats in
general, also kind of serve as a basic feedback tool. We have stats
(not on GA) that we can compare these stats to to determine if this
week was better than the one last year. If it was, was it better
outside our normal growth pattern? Could that mean it was because we
were having start-up meetings around the world and people were drawn to
the site? GA can help you get perspective on events like that.
What specific feature(s) do you find most valuable?
One
feature that I don’t use enough but has enormous potential is the Goals
and Funnel Analysis. It’s awesome to see where visitors “drop out” of a
certain process, like signing up. One of my colleagues in Argentina has
started to use it for Idealistas.org and she’s inspired me to take
another look at it and set up some goals of my own.
Top
Content is another feature that is extremely valuable. It’s really
simple, it just shows you how much traffic any particular page on the
site gets. But when you want to see if this resource center home page
is seeing traffic or determining which of the header navigation links
get visited the most, it’s the tool to use.
Other
features I use frequently are: Languages, Geo Location, Geo Map
Overlay, and Browser. Very basic tools, but useful for an international
site with a wide variety of visitors.
Any words of wisdom to other nonprofits are just beginning to use a tool like GA?
I
would caution against drawing strong conclusions from some of the
statistics offered on GA. The Goals and Funnel Analysis feature is
pretty safe, but trying to understand why your site had higher than
average exits this week can be a dangerous and unproductive guessing
game. It could have everything to do with a new headline and article on
your home page, or it could be something completely different. Stats
are better for long term trends or very short events (like measuring a
“digg effect” or getting Tech Crunched). It’s harder to find
correlations between incremental site changes and the ebb and flow of
web traffic.
Also… this is
kind of easy to figure out, but Google Analytics only updates it’s
stats once or twice a day. That means there’s no excuse for compulsive
stats checking! It’s likely going to be the exact same set of
statistics when you check it again five minutes later. 🙂
If you have a story about how you have used web analytics and
discovered actionable information and would like to leave a voice
message, please do! If you have a screenshot you’d like to contribute, you can send it to this group in flickr!