Just before Twitter went down today (yup, it was down again), I noticed something strange. Whenever I clicked on any shortened link in my Twitter stream and look at the address bar of my browser, I saw a fleeting click tracker before it redirected to the final site. It looked something like this: “http://twitter.com/link_click_count . . .” For instance, here is the full URL redirect for one link I managed to capture:
http://twitter.com/link_click_count?url=http%3A%2F%2Fbit.ly%2F3omd6p&linkType=web&tweetId=3541772256&userId=12798452.
Others noticed this as well. When Twitter came back up, the redirects were gone. Maybe too many people were clicking on them.
Whether this was just a test or a preview of what’s to come, it suggests that Twitter wants to track all the links people click on the site, which is something you’d think it was doing already. The way that Twitter was doing the redirects was a bit clumsy. You actually ended up being redirected twice. First by the original URL shortener like bit.ly or ow.ly, and then by Twitter itself. While it only seemed to be happening on Twitter.com itself, the redirects worked for any short link including bit.ly, Tinyurl, ow.ly, and so on.
Why would Twitter be tracking links all of a sudden? It’s all about the passed links. First of all, those links are a treasure trove of data. By seeing which links get shared and clicked on the most, Twitter can tell where it is sending the most traffic, who is sending the most traffic, the most popular tweets, the most influential users, and more. All of this data would come in handy for Twitter’s planned analytics service it wants to roll out to business customers.
They can do a whole lot more too. First of all, they won’t be relying so much on bit.ly for click data on short links, even though bit.ly remains the default URL shortener on Twitter, and thus the biggest one, for now. With it’s own data, Twitter can move into bit.ly’s backyard and start showing the most popular links and what is being shared right across Twitter. In addition to ternding topics, it coudl also show trending links.
(Flickr photo by Gerlos).
I think twitter should worry about staying online all the time first before tracking what i click
Twitter sucks compared to places like http://anonboard.com
Its surprising that they werent tracking all this stuff previously.
Me too! My guess would have been this data would already be control. I’m curious how many outbound clicks are caused because of Twitter daily.
I would of never thought of Twitter making an analytics app, where they charge a monthly fee. Clever, but will anyone buy it?
Everyone loves stats. Of course there will be people who would pay for that. I probably would.
Yes, businesses will most definitely buy it. Everyone and their grandmother’s bank is wanting to know how to market successfully to twitter. The analytical data will be useful to internet marketers on a number of levels.
Or maybe, Twitter could add a digg element. and a leaderboard for top URL sharers. And then we’d unlock badges and be like foursquare. And suddenly Twitter can be EVERYTHING. Except it still won’t work a lot.
“yup, it was was down again” – there’s a mistake?!
I know that collecting data is powerful and very very important and interesting as well due to the fact I am a statistics lover. Anyway I agree with Skankazoid to make twitter stable before adding new features!
I think they are testing for a future monetizing program. That’s what it sounds like to me anyways..
James F.
Owner, TwitterBackground.com – Free Twitter backgrounds
[..]In addition to ternding topics, it coudl also show trending links.
trending and could have typos.
@David
Yes, the amount of information they are sitting on is a huge gold mine.
ADUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUHHHHHHHHHHHHHHHHH……..
You know… companies like Google would NEVER track what links you click on… EVER!!!
This wreaks of Sean Percival.
no comment
I’m surprised at this point that Twitter didn’t just merge Bit.ly into the overall website. Seems like the logical thing to do.
( Per Conan O’Brien’s bit )
“…TWITTER TRACKER!!!TWITTER TRACKER!!!TWITTER TRACKER!!!”
Looks like they tried to do what facebook has been doing on links from early on. Couldn’t stand the load apparently.
Instead of posting downtime, post about uptime!
It’s unsurprising (you could say entirely predictable) Twitter are collecting click traffic. That kind of data would be invaluable to marketers looking to find out how their Twitter-targeted promotions are faring, and Twitter would easily make money out of that.
What I am surprised about, however, is the way that Twitter went about it. The two redirects are messy at best. I would have thought that they would implement a proper (a la Google Analytics etc.) tracking system for this kind of thing in order to extract as much information about the user’s habits as possible in order to provide as much value to potential customers. Maybe they have now learnt their lesson and will implement it “properly” (i.e more robustly) next time around.
NOOOOOOOOOOOOOO THIS IS BAD!
This is totally normal, and done by lots of sites. Site makers would be incompetent if this _didn’t_ track clicks.
links would be first tracked by twitter and then by any short service / other url
btw i ‘m pretty sure if techcrunch stopped blogging about twitter for 1 week, their traffic would be plummeting
About time twitter is tracking clicks. I believe this can be the beginning of their business model. Tracking those who are sharing links and those in their commnity who are engaging and clicking allows them to build a very robust social graph…. add brand pages, track them…. add code on brands sites…. perhaps a twitter network of sites… it can happen..
@davehonig
What about the hundreds of API clients – how would they track those clicks? There is no guarantee that javascript would be available on these clients
replied to wrong comment – see below..
I wrote a post about the importance to Twitter of click-tracking for external links (including via API clients): http://tr.im/xewY
Haven’t the devs at Twitter heard of unobstrusive javascript? They could capture all the data they wanted in background calls from the browser (XHR, anyone), all without getting in front of the url. It’s faster, since the user doesn’t do two server hops in succession.
I think it’s abundantly clear that Twitter has used every penny of its 55M in funding to buy PR on all the mainstream media, while allowing its technology team to be run by novice programmers who have no idea what they are doing. I cannot believe people put up with this crap…
question is: how much did TC get?
That’s the first thing that came to mind when I saw this. Why not use background event tracking like all of the JS-based analytics packages do? It’s not like Twitter users have JS disabled.
Im guessing Erik didn’t do a spell check before he hit the publish button? HUGE amount of typo’s in that article.
They don’t give a shit about selling analytics, that’s chump change. They will use this data to improve their real-time search algorithm.
This is a really crappy and amateurish way of link tracking. Twitter should really know better.
i dont like this idea because i already feel like there is to much tracking going on anyways why on tweeter why now
It’s probably for advertising purposes.
Why didn’t they just do this from the start? then it wouldn’t cause any hostility when they decide to do it!
Twitter has always been tracking our clicks as they have Google Analytics stamped all over the site.
I would guess the new URLs are a sign of ‘better’ tracking methods internally for real-time benefits.
Dear Everybody!
Guess what! Just about every site on the internet tracks your clicks. Ever hear of Google Analytics? It tracks whatever you do. Plus, most corporate websites use tools like Eloqua, Marketo, Loopfuse or Tea Leaf to track your every move, where you go, where you came and what you’re doing. Even before you register, they tag you with an anonymous cookie and infer your company through reverse DNS.
Don’t be naive. They sky is not falling –or perhaps it has fallen already. Nothing to be alarmed about. Move along.
Absolutely spot on! If folks are worried they should have said so long ago or get off the grid.
I’m not surprised that twitter is tracking click. However, I am surprised that they haven’t done it before and used the information. As for the app, let’s see who actually buys it.
Or they could buy bit.ly …
… or any of the other services already doing something like this. Grow “google-like”, in the good ways.
Twitter does not have the function re-tweet. You can’t re-tweet as easily from your web based account than you would do from a platform like tweetdeck and seesmic.
@# and retweet were created by users. Twitter integrated@# and now they are thinking to do the same with re-tweet. Maybe what was seen are tests to address this.
Just a thought…
Not sure why people are in a huff about this move? It’s most likely related to their near-future plans to provide statistics and data for premium accounts.
I noticed the cumbersome address in my address bar for each of several links I clicked on. And then I couldn’t get to those links because of server problems at Twitter. The system didn’t work.
I never love twitter.
Good job Erick. I’m usually sick of Twitter stories, but this one’s pretty interesting.
Sneak peak at what’s next. Nice scoop
http://www.traderbots.com
Interesting article. I didn’t notice this on my end before things went down so I guess that they ran it on select users. Ryan
Umm…I imagine they were “tracking” things backend before, hey? All they’re doing now would be to start paving the way to offering some analytics data to users.
However, much of the more interesting data — inter-user comparison, “what URLs” are users at large clicking on, and as mentioned before, how other APIs are being used — would most likely not be tracked in public view. (That’d be just the sort of information, however, most useful to marketers/branders/etc. Tweet link clickthroughs are relatively boring, though seeing who follows/unfollows after a specific post would be relatively useful.)
Ahh! It’s the commies! Or the aliens! They’re following us around!! AAHHH!!
Quite frankly, I really hope they can figure out how to make their CSS load consistently before trying to add new features.
Makes API programming so much trickier, have to accommodate for slow or no response from Twitter servers.
Google already tracks your search results clicks!, but… I didn’t find an article about it on TechCrunch?
Oh Wait! it’s on stackoverflow: http://stackove...is-the-best-way
I can’t see any reason to believe that twitter wasn’t already tracking clicks, but accidentally made it public…wouldn’t be the first time things were accidentally made public
Interesting article, although we all knew this was next.
This is not good, it will definitely drive alot of people away even more to use third party tools instead of using the web version.
Such a move would be a welcome one. Trending links. Ha.
I think people are underestimating how much more difficult it is for Twitter to track links (across multiple clients using APIs) than it is for Joe Shmoe to track links on their website.
I can really see this as a way of opening the doors for advertisers to come in…
wait…if this was the case wouldn’t all the bit.ly data get screwed up – its referrer would be the twitter count page rather than the actual page.
Pete Cashmore is that you? (Peering through the binoculars…)