Sifting through the Twitter noise

Reading Time: 2 minutes

I was checking my blog stats last week and noticed and noticed that I had a sudden spike in traffic in the week of Thanksgiving (yes, I know, I should check my numbers more often, but I have a day job):


A quick glance at my favorite web analytics tool revealed that the culprit seemed to be none other than Twitter (with a decent bit of help from Techmeme):


Twitter is not normally anywhere in my top referrals, let alone Techmeme, so I was by now aglow with excitement. Furthermore, it looked like none other than the world’s pre-eminent ex-blogger (although he, er, appears to have started again), Jason Calacanis, is to be thanked for my brief elevation. A glance at the top pages being looked at confirmed which post:


But what did Mr Calacanis have to say about my lowly blog? The referral information in my web analytics data was not very much help – it just told me that the link came from Jason’s Twitter page at the URL How do I find out what he said? I hustled over to Twitter to look through his recent tweets to see if I can find the one about my post. This is where I hit a snag. Jason Calacanis is a prolific tweeter, sharing insights into the profound currents that swirl in his brain many times a day. So I had to trawl through several (to be accurate, thirteen) pages to find the post, which turned out to be this one:


It was only when I did this that I realized how much of a pain in the butt it is to locate a specific tweet that’s sent you traffic. The job is made much harder by the use of services like and  which make it impossible to determine whether a tweet links to your site simply by looking at it. But the lack of informative data in the referring URL means I have no other choice but to try out the various links in the hope that one of them will lead to my site. Which, when some of the links are like this, is a bit nerve-wracking:


Of course, given how infrequently my lowly blog is referenced on Twitter, this isn’t such a problem for me – and tweet-based links from less prolific Twitterers would be easier to track down simply by a process of time-based elimination. But if I were doing analytics for a larger site and I wanted to track and analyze the traffic being driven from Twitter, it would be a right royal pain in the ass.

It’s hard to see what Twitter can (or should) do about this. Even if Twitter were able to push meaningful referrer values through with clicks from, no such data would be available if a reader clicked through a desktop-based reader such as Twhirl. The problem wouldn’t yield to the kind of solutions used for RSS tracking (i.e. republishing your RSS feed through a tracking service such as Feedburner).

What do you think? If you’ve come up with a creative way of tracking Twitter referrals, let me know in the comments.

4 thoughts on “Sifting through the Twitter noise”

  1. The only thing I could think of in your case is a bit involved. Basically, you could use cURL and the Twitter API to export all of said individuals tweets– import that dump file into excel, filter by http:// and then check the links that way.
    Not perfect, but quicker than manually going through everything.

  2. I’m certainly venturing out of my realm here, but this seems like a great opp for TinyURL. They could have a data service to report on the destination URL along with the referring pages. If they could just point you back to the individual tweet — but if you at least knew the TinyURL that drove the traffic you can search Twitter for it.

  3. TinyURL is a pain wherever you find it – there’s so much information about a site in its URL and it’s irritating to have to go to a URL with no idea of what you’ll find behind it. Twitter could help by enabling links to be embedded in Tweets properly (so the real URL can be seen, but the length limit is still respected), but I guess they’ve optimised for simplicity, so that’s unlikely to happen.
    One approach would be for websites to run their own tinyurl-like scripts on their own domains to provide a quick link to their deeper pages. Again, that’s a long way off becoming a trend.

Comments are closed.