The Google Analytics Beta: Improving Profits through Web Site
Analytics by Ross Lambert
Web site analytics, for those who might not be familiar with the term, is the
tracking of various performance metrics for a given web site. The metrics
themselves can range from the simple (and relatively useless) count of "hits",
i.e. requests for a given resource such as a single web page, image file, etc.,
to the measure of far more complex interactions. These complicated interactions
can be totally arbitrary; for example, you might want to know the number of
orders from visitors who came from a search engine and scrolled at least
halfway down a long sales page.
That assumes, of course, that you can figure out how to configure all that
tracking, interpret the results and afford the monthly fees for the providers
of the service. The cost issue is apparently solved: Google Analytics (http://google.com/analytics
) is currently free in its beta version, and early indications are that it will
remain so. However, a word of caution is in order: The Terms of Service
referenced on the Google Analytics home page seems to indicate that Google can
and will make use of your site's data, at least in aggregate form (that is,
mixed in with everybody else).
In many minds Google is starting to become a Big Brother-like presence on the
web, hence its motives are suspect pretty much by definition. Personally, I
consider my site's aggregate data a fair trade for the value I will extract
from their software, but you will have to make up your own mind. If you're not
bothered by Google knowing as much about your web site as you do, then Google
Analytics looks very promising. It is a smart, easy-to-use implementation that
hits the sweet spot of web analytics.
The Sweet Spot: Easy Yet Powerful
The sweet spot I'm referring to is really the point where most of us live. We
don't have the technical know-how to configure the most complicated tracking
scenarios and even if we could, we don't have the analytical savvy to make any
sense of the data. Google has found the sweet spot by making tracking
configuration quite easy, and providing pre-cooked role-based reports that
provide lots of information you may not have even realized was readily
available. In short, you can get an awful lot of strategic data for very little
effort.
Configuration
Let's walk through setting up a simple and common scenario: We want to know how
well our sales letter is converting web site visitors to customers. Where
Google Analytics shines is how much valuable data it automatically gleans from
just such a simple test.
Google calls a tracking scenario a "profile". Although you can include URLs
from many web sites in a single profile, it is easiest if you organize things
such that a profile is fundamentally the same as a web site.
As part of setting up your profile, you provide the URLs of all the pages for
which you want data. Google then provides you with a JavaScript snippet to
include on each page. The snippet is self-contained and requires no editing. It
contains a Javascript include and the following line:
urchinTracker();
It really couldn't be easier.
You can put the snippet anywhere inside the body tags of your web pages.
Next, you want to specify a "goal". The goal in our case is sales; we know that
the goal has been achieved when the customer reaches our thank you page.
Therefore the URL associated with the goal is that of our thank you page. More
sophisticated goals can involve defining a "funnel" of multiple pages; this can
be extraordinarily useful in identifying a weak spot in a more complicated
sales process.
At this point our setup is finished! You then need to just let your site run
and accumulate statistics for at least 24 hours.
Reports
When you return and select View Reports, you will see an amazing array of
statistics at your disposal. The first thing you'll notice is a pop-down menu
with several roles, namely Executive, Marketer, and Webmaster. Each role has a
suite of pre-cooked reports likely to be of interest to someone in that role.
We'll focus on the Marketer role; when you choose this option you'll see the
Marketing Overview by default. It includes four charts:
1. A line graph showing raw page views over time 2. A pie chart showing the
proportion of returning versus new visitors 3. A world map showing the
geographic distribution of visitors 4. A pie chart showing the visitor counts
based on the referrer, i.e. Google, Blogger.com (for my blog), etc.
The Marketing Summary report is a numerical chart that shows the top five
referrers, the top five keywords used by searchers, and the top five campaigns.
A campaign is indicated by a code that you attach to a URL. Even so, by default
you get several campaign totals. These default campaigns are:
Organic: Indicates visitors referred by an unpaid search engine listing.
Referral: Indicates visitors referred by links which were not tagged with any
campaign variables.
Not set: Indicates visitors referred by links which were tagged with campaign
variables but for which the campaign variable was not set.
Direct: Indicates visitors who typed the URL directly into the browser.
The next report of interest is Overall Keyword Conversion. Since we have
indicated a goal of "sales" and linked it to our thank you page, the Overall
Keyword Conversion report is able to tell you which search engine keywords
result in the most sales. This is a really useful and potentially profitable
report.
The Campaign Conversion report shows which campaigns are creating the most
sales, the Conversion Summary produces total visits and total goal percentages.
Finally, the Entrance Bounce Rate is an interesting report that also has
valuable data, even in our simple scenario: It provides the list of pages for
which customers land and then leave right away. For some pages, our product
download page for example, we expect a 100% bounce rate. For others it can
illustrate a weak or problematic page.
Google Analytics provides an astonishing amount of data for very little
effort--and no cost (so far, anyway). Although there a few advanced reports
missing from its arsenal, it makes the bulk of the web site measurement you'll
want to do very easy indeed.
About the Author
Learn more about web analytics at http://midnightmarketer.com.
Ross Lambert founded Midnight Marketer, a newbie-friendly community of web
marketers. He is also the author of Sonic Page Blaster (http://spbsavestime.com)
and Ross's Guide to the Masters of Marketing (http://saleslettergenius.com).
|
