I have a big beef about web analytics. While analytics are useful, informative and quite frankly, absolutely crucial, the ease and access to free tools like Google Analytics has caused many to use analytics tools without really understanding them.
Here’s the one thing I hear often from skeptical marketing people who don’t fully understand social (an imaginary example): “I get Google Analytics reports, and less than one per cent off our traffic comes from Twitter so clearly it’s not working for us.”
So here’s the problem: analytics, while providing more precise and in-depth metrics than ever afforded to marketers, is NOT a perfect science. I could go off on several rants at this point, but for the purpose of brevity I’m going to focus on Twitter referrals.
The reality is most Twitter users do not access and use Twitter from Twitter.com. They use Tweetdeck or another desktop applications, and they also access Twitter often from mobile devices. When someone on Tweetdeck or an iPhone clicks your link and visits your site, this does not register as a Twitter referral, but in most cases as direct traffic. Direct traffic is essentially the “other” bucket when the analytics tool can’t figure out the source. This includes email clients, bookmarks and any other non-browser application.
While I can’t find an official number of Twitter.com vs Twitter app users, a recent(ish) study by Sysomos revealed that 42 per cent of the 25 million Tweets analyzed originated from non-official Twitter apps. That’s not even including the official Twitter apps (aka Twitter for mac + Twitter for iPhone/Android/BB, at the time). That tells me at the very least, a significant percentage of Twitter users are NOT using Twitter.com and therefore, to the untrained analytics watcher, not driving traffic to web sites.
Now I hate to complain about anything without a helpful solution, so here’s my advice:
- Tracking links – if you are pushing out content via Twitter or other social networks, you can use a tool to code those links (before shortening), you plug in a few tracking details and get a lovely URL that looks like this:
with all that fun stuff at the end telling Google Analytics information about the URL for tracking purposes. You can then take that ugly long URL and plug it into a service like bit.ly and distribute the shortened version.
- The “lite” version is to just track all your bit.ly links which does in fact tell you more details, here’s an example:
As you can see, bit.ly does in fact track different clients, (but still has a “catch-all” category).
I would use the first example if you’re pushing out marketing campaigns and tracking clicks and conversions as tying it back to the source is pretty important and the tracking URLs will allow you to keep tabs on a referrer right through to conversion. If you’re just pushing out content (i.e. blog posts), while you can still do the Google tracking, it’s not completely necessary if there’s no direct call to action.
Finally, again I will emphasize: this is not an exact science! Even with the tracking URLs–if I click on a URL in a tweet, read it and decide to re-share it, I’ll copy and paste the URL from the browser and cut out all that silly tracking mumbo jumbo (you think I’d know better, but I still do it!) and paste the original link in, thus crushing your clever measurement trick! If I do that I’m sure others do as well.
Track what you can, as much as you can, but remember it’s a trend, an indication, but not an exact number. Measurement needs to be an ongoing process and you can only really do so effectively when you measure consistently, benchmark and compare against your own results.