Finding ways to track your website and the performance of
the URLs that you give out is an important tool in monitoring your internet
marketing efforts. This allows you to learn what is and isn’t working which
will then show you if you need to adjust your approach or even your content.
I’ve been using the bit.ly
service the past few days to do some comparisons on their stats reporting
versus Google Analytics stats reporting. I also used Google Analytics (GA)
campaign tracking to simplify things on that end.
bit.ly is a service
that shortens long URLs and, as a bonus to number buffs like myself, gives you
instant and live tracking statistics. It tells you where the URL is posted,
what conversations the URL is being involved in, where everything is coming
from, time of the clicks and more. GA statistics are delayed which is why I
wanted to try this route in tandem with the extended and in depth tracking that
GA offers.
Here’s what I did yesterday, February 11, 2009:
- I
created a post for my small business tips about customers.
Original URL
http://www.americasbestcompanies.com/blog/small-business-tip-about-customers.aspx
- I took
the original URL and put it into the GA campaign tracking URL Tool. There’s a great tutorial on how and why to
use that here.
GA Tracking URL
http://www.americasbestcompanies.com/blog/small-business-tip-about-customers.aspx?utm_campaign=Small%Business%Tips&utm_medium=Text%Link&utm_source=Twitter&utm_content=About%the%Customers
- I then
signed up for a free account at bit.ly.
- I
pasted my full GA link with all their tracking tags into the field and had
them shorten it.
Shortened URL
http://bit.ly/CQifo
- I then
sent a Twitter message out sharing the link.
My stats were indeed updated live via bit.ly and I could see exactly when
someone clicked on the link. As of 10:00 am today, February 12, 2009 here are
what the two different stats reports are showing me:
bit.ly shows the
link received 46 clicks.
Google Analytics shows the link with the tracking code received
6 unique pageviews.
My question now is why are the stats reports so far off from
one another? There are, of course, certain variables to consider like mobile
phones that aren’t JavaScript enabled (GA tracks via JavaScript) and things
such as multiple clicks from a single user (GA doesn’t count these). Even
considering a couple people clicking the link twice or a few mobile views, I
still don’t see how these reported stats are so different from one another.
I certainly didn’t tell anyone that I was performing the
experiment so my results haven’t been skewed in that regard. My current Twitter
follower base is 424 which makes me more inclined to believe that 46 people saw
the article via Twitter over the reported 6 by Google. I did another set of
business tips with the same tracking today. My results are pretty on par with
what happened on the above example.
With the numbers varying so much, I pulled down raw server data just to
do a quick comparison. My server reported 36 hits to the article. This number lends us closer to what bit.ly has reported yet still no where near what GA has documented.
Right now I’m left with more questions than answers. An
article by Michael Martinez at SEO Moz from August of 2006 reports similar inaccuracies as do a majority of the
comments on his post. How accurate is Google Analytics now? Has it gotten
better or only worse? How about bit.ly?
What’s their story on tracking?
Despite there being accuracy and reporting issues, it’s
important not to give up on tracking statistics. Data and results are what help
us to make the decisions that can bring us to the next level in our business or
our marketing. My recommendation is not to put all of your eggs in one basket. In
other words, don’t rely simply on Google Analytics for your tracking. Compare
and contrast with your domain’s raw data logs and don’t be afraid to use other
services as well.
If you enjoyed this post, please consider leaving a comment or subscribing to our free newsletter to receive future articles and information delivered directly to your email inbox.
Thursday, May 07, 2009 at 4:44 PM
Steven Mocarski says:
I suspect we have a few things going on, but at it's simplest your issue is just that google analyics is tag based (i.e. javascript in the browser) and bit.ly works similar to a more traditional log based metrics. The advantage is that you get the click in a tag based system (google analytics, omniture, others) when a user visits the web page and the browser pulls the metric. In a log based system, it is the request of the URL that triggers the visit. In your case, since its a twitter link, you might have had people visit the page in a mobile device that might have gotten to bit.ly, then perhaps your website, but would never have downloaded the javascript.
Once could imagine, as John suggests, that some folks died at the bit.ly link, which I'm guessing accounts for the difference between bit.ly and your local logs. The delta between the local logs and google is probably the javascript. Does that make it better? Sometimes. But often not because while some things block (or do not pass) the tag, others block the request. Things like corporate proxy servers will often prevent your site from being hit the true number times the page is actually viewed. Theoretically, corporate firewalls could also block the tags but its a lot less common.
You might try looking at the user agents to get some more clues into the discrepancies.
