Something Is Very Wrong with Bit.ly’s Click Counts
May 21, 2009 45 Comments
I love the URL shortening service bit.ly, as I’ve written before. It’s a tremendous service, provide wonderful analytics along with the basic URL shortening feature. The service recently moved to make click counts much more visible, which is really helpful to see at a glance what got the interest of people you shared the link with. Search Engine Land’s Danny Sullivan has a nice write-up about it.
But something isn’t right with the counts I’m seeing from bit.ly. That, or something is seriously wrong with WordPress.com’s traffic stats.
Here’s what I mean. On May 19th, I tweeted this:
The first bit.ly URL is to Dennis Howlett’s blog post. The second bit.ly URL links to my post Newsletters Are Still Viable? How I Approached My First Newsletter Email. This post was from last November.
Fast forward. Bit.ly dutifully tracked the clicks on my shortened URL. Total count? 125 in the past week:
Hey…that sounds pretty good, doesn’t it? Plenty of clicks to the ol’ blog.
But then check out the number of views WordPress.com recorded for this same blog post the past week:
Say what?!!! Bit.ly is telling me the post got 125 hits. WordPress.com is saying it got 11 hits. Let’s do the math:
125
– 11
114
How can the numbers be so far apart? I mean, that’s not a rounding error. That’s a canyon of difference. Is bit.ly borked? Is WordPress off? This isn’t the first time I’ve seen these kinds of differences.
If you’ve got any hypotheses or have seen this yourself, I’d love to hear about it. Particularly if you’ve seen the same thing for a different blog platform, like Blogger or Typepad.
Update: There’s a discussion of possible causes on FriendFeed.
I’m @bhc3 on Twitter.
About Hutch Carpenter
I'm Not Actually a Geek 
From my experience, I would be quicker to say it is WP that’s borked. There have been times on my blog where an entry will have 9 unique comments, yet WP will show that blog post as not having views or only one or two.
I’ve been having the same problem. Bit.ly gives me a high click rate when I link to YouTube videos and yet YouTube shows a much lower view count.
I love using bit.ly but the numbers just don’t add up.
The simple explanation is that the click count is artificially inflated by bots. On the WordPress side the count is probably computed using Javascript, which bots do not execute, hence the discrepancy.
The wordpress figure is more accurate.
Fred – you’re right. WordPress does a good job with the integrity of their traffic stats.
Bit.ly actually counts people using Twitter clients and other third party tools. They show up as “direct traffic” in your stats. Nonetheless I also have the impression that Bit.ly stats are a little inflated. I don’t think they accurately filter reloads.
Yeah, I’ve noticed the disconnect between bit.ly click counts and my WordPress view counts to know thaqt bit.ly is overstating traffic. In some cases, like here, by a lot.
They probably logging bot traffic too on bitly that WP disregards… though that is some gulf. For my part, I trust WP stats as we have them on over 45 domains in conjunction with Google analytics – they pretty accurate.
Rob – that seems to be the consensus about the bit.ly numbers. Bots are killing the integrity of the click counts.
To be honest one should not be concerned with numbers from either. thats just a “parlor game” that is wasting your productive time. instead worry about providing good content you are proud of. The rest will take care of itself.
Thanks Chris – good philosophy. Still want to know numbers though.
Hutch, do you have Google Analytics on your blog? Would be interested to see if that falls closer to the Bit.ly measurement or the WordPress measurement. Also, contrary to Fred’s comment here, I believe that ads and pages like WordPress are NOT invulnerable to bots. Many sophisticated bots run JS and actually render the page. Anyone heard of cligs (another shortener service)? – it apparently differentiates between bots and humans – don’t know how good it is. The fact that one person had 9 comments on their blog but WordPress registers 2 leads me to believe that WordPress isn’t a good way to measure the accuracy of bit.ly since it’s highly flawed itself
David – I don’t. I have wordpress’s analytics tracking, which is quite good. Most people seem to think the wordpress numbers are closer to the truth. You’re right about wordpress and bots – it does a great job weeding them out.
I’ve had this issue on different platforms.
I have a bit.ly link pointing to a Blogger Blog with Analytics on it. Bit.ly says I got 125 clicks on a single link yesterday, but Analytics says I have 24 visits for the entire day. I have other bit.ly links pointing to the Blog, and other links elsewhere, and a tiny bit of regular traffic.
But even if my ONLY traffic was from this bit.ly link, 125-24 is 101.
I have bit.ly links pointing to my column on examiner.com. Their reporting tool shows 85 hits for one day, but I have multiple bit.ly links pointing to the column. One bit.ly link alone shows 380 clicks, another 180, another 75.
Examiner.com is a large site and I get traffic from a number of other sources.
Again, if we ignore all of those other sources and just imagine the bit.ly links are providing traffic, that’s hundreds and hundreds of hits compared to 85. The discrepancy is not small. I’m considering using a different tracker for a while and see how it compares with the other stats.
Great info Sean. I’m seeing similarly bloated click counts. In that FriendFeed discussion to which I link, someone suggests trying tr.im.
We’ve had huge discrepancies between bit.ly and Google Analytics. Stuff like bit.ly registers 600 clicks, and GA registers 180. That’s huge, no?
Oh and my point about Linkbee is that they’re only going to pay you for clicks that people can view, and that they won’t report robots to you at all since they won’t get money for those “clicks,” so their analytics seem more accurate. The question with them is, do they underbid the clicks??
Hi, We are having the same problem. Our blog is on Movable Type, and I chart the stats with Google Analytics. We recently starting sharing blog posts via twitter, and our bit.ly stats showed that we got 900 clicks on a particular link that was retweeted by a person with thousands of followers. But our GA stats show the usual amount of traffic for that: around 600 visits, with NO discernible uptick at all. I haven’t seen any theories that can account for a discrepancy this big.
I agree bit.ly seems too high, and I just started using it today.
Does anyone have experience with tr.im?
What you’re seeing are total decodes, as opposed to total click-throughs measured by JavaScript on the page. Decodes can be caused by bots, and by applications, like browser plug-ins, which expand the underlying url without causing a click-through.
If you download a browser plug-in that automatically expands short urls, for instance, it looks a lot like a human user to an analytics program.
Absent JavaScript on the page, it’s hard to distinguish between a decode and an intentional click-through. At the end of the day, Bit.ly complements rather than replaces JavaScript-based analytics utililties like Google Analytics or Chartbeat.
Rex Dixon
Community Manager, Bit.ly
Hi Rex, thanks for adding your comment.
I think what everyone is trying to say is that it doesn’t matter to us what causes the additional visible clicks on our bit.ly links, if they aren’t visitors, we don’t want to see it.
It honestly does not matter to me how many times a link was decoded (or even that it’s called that) … if a link is decoded 1,000,000 times, but only 2 people have clicked on it and seen the resulting content, the important number to me is 2, not 1,000,000. So that’s the number I want to see.
This other number is very misleading, and it’s unfortunate that so many people are default misled by it. People are sending e-mails to and posting complaint comments about services saying that their numbers are wrong because bit.ly shows x and they show y.
Other url shorteners are able to show the difference between bots and clicks. budurl.com, for instance, makes it very clear what comes from what. The numbers there are far more in line, and they explain on the chart what is a bot or program, what is a clickthrough, etc.
The ONLY reason I stopped using bit.ly two weeks ago is because the analytics are so unrealistic, they’re utterly useless to even look at. Twitter.com decodes when you are on their front end and hover over the url, so if I get 529 clicks (it’s even called “clicks” in your interface) why does the article only have 37 views?
Then you give us charts where we can see how often it was clicked … er, I mean hovered over or decoded or bot-visited or anything else – the charts are not useful.
So every time I am searching on google.com for link shorteners with analytics and I see people touting bit.ly, it’s because when they use bit.ly, it appears on the analytics like they’re getting more clicks.
Please differentiate the decodes, bots, programs, views, and actual clicks by humans on the analytics.
@Sean Arenas – Appreciate the additional feedback. That wasn’t the best way to have worded that.
Decodes to actual click through numbers are different. Unless you have an actual javascript on the page like you would with Chartbeat or Google Analytics, you really can’t get accurate human click numbers.
Does this make our service any less valuable? We strive to let people know that these are decodes and do not at all try to be dishonest about the issue. Sorry that you are choosing to leave our bit.ly community and go elsewhere.
We wish you the best in your url shortening experience!
Rex Dixon
Community Manager, Bit.ly
Rex – is bit.ly doing something different recently? The massive click counts I used to see aren’t there anymore. Click counts seem to be much lower on average now. Some changes already in the counting mechanism?
Thanks for the reply Rex.
There are other URL shorteners that are capable of telling the difference between a human and a decode and a bot.
I’ve been trying budurl.com, and even with the free account, I can look at the clicks and see what browser someone used (or if there’s no browser, so it’s not a human). It shows the ip address of the visitor, the IP Host, info on the referrer, a timestamp, etc. This is realistic information that can help me.
When I made my comments about the bit.ly service giving information that is not useful, I did not mean to imply that you are trying to mislead users. I know you are doing your best to make sure that the users know they are the same as actual human views.
I’m just saying that the information provided by bit.ly’s analytics is not useful information. It doesn’t matter to me if 50, or 500, or 500,000 decodes or bots see my link or anything else. If only 10 people clicked it, the number I want is 10.
So, while I am not saying that you are misleading me by telling me that the # of clicks displayed by bit.ly is actually the number of humans who clicked on it …
but you are providing me information that does not actually DO anything for me.
How can I make use of the information that 900 “clicks” happened on this link, when I know for a fact that the real number of visitors who clicked it is closer to 50? What does that information do for me?
Also, you do say that you try to not mislead people, but your own interface calls them “clicks.” Is it a “click” when a decode occurs or a bot verifies the link? Why is it called a click?
(I’m not trying to be snarky, it’s an actual question)
I was wondering. Within a few minutes we had 37 hits to links to our website, according to bit.ly – Interestingly, our own stats (which we take from 3 monitoring companies) for our website they all showed no hits during that time period. It was 3:00 AM when we set it up – so not surprising. Do we lose the hits to our sites when we go through them? Are the recorded hits to them not us? Just curious.
Just did a little test – for you can gage the actual count. Put a new bit.ly in and clicked on it 3 times and go – 9 clicks showing on their stats.
But if you put the link where it can be scanned by a bunch of bots, the “click” count will go way up on it, while the humans who arrive at the destination will not. So 3-to-1 may work for actual clicks, but expose the link to bots and decoders and it will be way more than 3-to-1.
Pingback: URL Shorteners: Choose ur wpn « Steve Guengerich Ventures
I ran a Twitter marketing campaign for a client and thanks to weblog I’ve seen bots from 84.17.129.60 and 84.17.129.61 (a DNS registration company in Belgium) registered at bit.ly 2,000 ++ times as “clicks”!! In one of my experiments (over the last three days) I had up to 90% bots coming. Even if we assume that about 50% of clicks are not registered due to the non-existent JS on half of Twitter applications that drive traffic, the final result is horrible. It is incredible how many garbage is there.
We’d look at that bit.ly’s integration with Google Analytics in order to avoid charging the customer for clicks that are not legitimate. This feels like a nightmare. We’re developing our own URL shortener with only one idea — how to Verify-A-Click as accurate as possible.
Pingback: Demise of tr.im makes me realize I’d pay for bit.ly « I’m Not Actually a Geek
I experienced the same thing! I just noticed today. Here’s my post about it http://blog.jobtitled.com/2009/08/bitly-vs-google-analytics/
bit.ly cost me hours of work! I had received 2500 “Clicks” registered at bit.ly with not one optin. The only thing that can be wrong after that number of Clicks has to be the capture page. So I stopped my campaign and rewrote the capture page and started it again.
bit.ly now shows 14,024 Clicks but Aweber says 825 page views!!
(I have an Aweber optin box on the page)
bit.ly is certainly not something to use when making business decisions.
I am now using ViralURL
Steve
We’re finally out with our Case Study: TWITTER: The DARK SIDE (Does Bit.ly Enable a Massive Click Fraud?). It is here:
http://www.seo-artworks.com/Twitter/twitter-study.htm
So guys, here we are, presenting you with:
– 68 pages of research and staggering discoveries,
– BOTS vs. HUMANS Clicks Ratio on Twitter discovered (devastating for humans),
– IPA analysis of all the clicks on Bit.ly and
– Over a MILLION phony clicks on Twitter, on ONE Bit.ly link discovered. This cutie is here:
http://www.seo-artworks.com/Twitter/twitter-study-millionclicks.htm
And then much more, even funny:
– Are we going stupider because of Twitter – a funny account full of hard data including,
– Obama Girl the President follows,
– Celebrity Polluters,
– Insight into Followers,
– Twitter’s Future? Does it have any?
We even proposed a solution for the whole unholly mess that is Bit.ly. Good resource guys, trust me — we worked like animals to get that thing done.
Just did a little test – for you can gage the actual count. Put a new bit.ly in and clicked on it 3 times and go – 9 clicks showing on their stats.
โปรโมทเว็บ
โปรโมทเว็บ
social bookmark
Pingback: How Should Tweets Be Ranked in Search Engine Results? « I’m Not Actually a Geek
Pingback: How Should Tweets Be Ranked in Search Engine Results? | CloudAve
Pingback: Twitter, Authority and Search Engine Ranking: An Exploration « The SiliconANGLE
It’s so simple to figure out.You just get a tracker that shows IP addresses for each hit.Then google the IP addresses.You will plainly see which ones are bots and web crawlers vesus real humans.Read my blog post here on it with full details of date,IP address.—-> http://tristanginnett.blogspot.com/2010/02/i-often-take-break-from-all-my-social.html
I hope this helps.
Tristan Ginnett
FOLLOW ME ON TWITTER
Interesting article and comments. I consistently see low bit.ly view counts and yet my Google Analytics, Twitpic or any other stats are invariably higher – often 10 times Bit.ly’s numbers.
I thought everyone would be saying this, but they’re saying the opposite. Wow!
It’s really fascinating to me, but at the same time if the results on any given platform aren’t actionable then they’re almost meaningless. You shouldn’t expect to rely on any one number to provide you this. Still… nice to know we’re all out, one way or the other.
This is good info but I have noticed the exact opposite. I have a bit.ly link right now that shows 0 clicks, it is attached to a picture embedded in my Ads, but I know for a fact from analytics and other click trackers on my blog that the link has received something like 200 clicks. I am probably going to wholesale switch away from using bit.ly at some point. It just so convenient when tweeting.
Thanks for a good post to help me on the way, I am grateful for all your work in researching and writing this blog
Pingback: The Emperor Wears No Clothes: Is Bit.ly the Enron of Social Media? | Matt Culbertson
Pingback: Twitter vs Bitly: Whose URL count is more accurate? - Quora
I don’t know about bit.ly giving you extra clicks or not but I use to use goo.gl to shorten my URLs and I had so many clicks and now bit.ly is saying I have no clicks through. BS? It’s like it isn’t tracking my urls
I tried to use 2 different url shorten services today to check if my wondering about the very wrong numbers of Bit.ly clicks counts.
in my Bit.ly stats: it shows 1449
and my Google shorten service shows: 6
and my Google Analytics shows: 9
my goodness ……sorry
no more Bit.ly i think….
do you think this might also happens with Interspire statistics ?
we send some email marketing campaigns and usually we have a lot more clicks than our client.
they tell us that the interspire stats may be inflated by bots
this make sense to you ?
i tryed bit.ly … but i discover they also “faked” by bots…