
When it comes to shortening links, the brevity of the URL is not the only factor you should consider. Speed and reliability are also important. After all, the main purpose of URL shortening services such as bit.ly and TinyURL are to redirect you back to the original (long) link someone shortened in the first place. You want the redirect to be fast and reliable.
The rise of Twitter and its character constraints has made URL shortening services popular. But which one is the best? Royal Pingdom decided to conduct a test over the past 30 days to put nine of URL shortening services through the paces. Overall, Ow.ly and Bit.ly came out on top, while Tr.im came in dead last (no wonder nobody would buy it).
Pingdom tested two things: the average time it took each service to load a page (overhead) and its uptime. The fastest service was actually Is.gd (with an average load time of 163 milliseconds), followed by bit.ly (261 milliseconds) and ow.ly (289 milliseconds).
Tr.im was the third slowest after Cli.gs and Snip.url. But when it comes to reliability, Tr.im had nearly twice the estimated yearly downtime as the next worse service (Twurl). Even with nearly 80 hours of estimated annual downtime, Tr.im still had 99.10 percent reliability. Ow.ly, however, had 100% uptime, and bit.ly came in second with 99.98 percent. (Is.gd came in fifth).

When both the reliability and speed rankings are combined, ow.ly and bit.ly tie for first place overall, followed by Is.gd and Su.pr. Here is the complete ranking (the scores are the rankings for each test added together):
Overall Ranking
- Ow.ly: 2
- Bit.ly: 2
- Is.gd: 3
- Su.pr: 3.5
- TinyURL: 4.5
- Twurl: 7
- Snipurl: 7.5
- Cli.gs: 7.5
- Tr.im: 8.5









Trunc.it, the URL shortener built into TwitterGadget has not failed me yet. It’s as much a media upload as it is a URL truncation service. It’s rather new, I think, so likely didn’t make the list?
Everyone knows http://bacn.me/ is the best url shortener
Everyone knows all the spammers use bit.ly… isn’t it blacklisted like everywhere?
Actually, bit.ly has the strongest spam filters in the business.
It reminds me of bacon, which is a plus, I guess.
Since Tr.im is out of the look and in headed to the deadpool anyway – why do we care about their stats now?
Because Tr.im suggested that it was unfairly driven out of business by bit.ly and Twitter. This data tells a different story.
Seems like another consideration – functionality – is one that is left off.
I want reliability and speed, yes. Other things that might be important:
- Traffic Stats (How many clicks have there been?)
- Custom URL’s (can I specify my own short name?)
- Preview (Can viewers see the URL to make sure it’s safe?)
- Comments (Can I append a comment onto the tinyURL?)
- User Accounts (Can I sign into the service and manage all my tiny URL’s)
- Extended Character sets
There are a lot of these services, and some have some very nice features. I’ve seen each of these features in various services, but no service with all of these features.
Am I missing any features?
su.pr has most of the features listed (no preview and I don’t know about extended characters) and did well in the test
Why is no one asking the reason for twitter choosing bit.ly rather than is.gd or ow.ly which saves one character and equally better performing?
That’s a weird comment given the data analysis Erick.
Nothing is “fair,” so I think tr.im made the mistake of assuming Twitter would encourage competition in addition to being open; however, it should be clear to all developers that Twitter has definitive favorite services (this was made clear by the notes Techcrunch published), so if you are not hearing from Twitter you’d better assume you could be chum for a competitive investment by a V.C. associated with Twitter, or Twitter itself.
Next, the service analysis doesn’t really address twitter’s selection of bit.ly — that had to do with sharing v.c.s with betaworks. No one disputes this, and as a result, tr.im’s suggestion is still valid.
Beyond that, if tr.im made the decision to shut down Aug. 9th or thereabouts, why would you consider a test of the last 30 days “July 16, 2009 – August 16, 2009″ to be a valid representation of tr.im’s data.
” The rise of Twitter and its character constraints has made URL shortening services popular. But which one is the best? Royal Pingdom decided to conduct a test over the past 30 days to put nine of URL shortening services through the paces. ”
Past 30 days — “Tr.im entered The Deadpool on August 9, 2009″ Crunchbase — why would you consider the data to be a meaningful test if tr.im was already unwinding/not expanding hosting services to keep up with growth etc.?
sorry for typos and redunancy.
@blah blah
Spot on. Seriously, Eric, it didn’t occur to you that the sample period was particularly unfavorable to trim ? And why all this fuss about them ? If you really don’t care, then stop mentionning them!
The truth is, the whole twitter/betaworks/morning 3rd-party-app-CEO jogging with the api team stuff is disturbing.
Well actually not, it’s just proof (if needed) that networking and relationships are key, and often more important, than product features etc. when it comes to business, specially with vaporware…
If Techcrunch had balls (like they used to), they’d do a decent investigation work on why certain apps are endorsed by twitter, how apps get featured in the sponsored defs, etc. etc.
I’m not judging, cuz in the end business is driven by relationships, but for god’s sake please stop the hypocrisy. We all know why bitly was chosen by twitter (it’s not for the product features), we all know that their servers have been moved to twitter’s datacenters, and we all know that it will pan out for betaworks as it did for summize. But please, this all benchmark thing just to put one more nail in trim’s coffin is ridiculous.
Schonfeld: Are you suggesting Twitter put these services through their paces and made their choice based on performance?
Erick:
The numbers in this “study” are flawed b/c Tr.im announced they were shutting down for three days during the test period. Take out those three days and tr.im goes to the top of the list instead of “dead last” as you wrote.
ooooops
It’s interesting to see those numbers. No real huge surprises, but validation is nice.
@myrridin
Would have loved to see u.nu in the list. Nice comparison though.
ow.ly is the slowest if you want to get the actual URL in your location bar, because it uses frames, forcing you to click the [X] to get rid of the frame. Booo!
There is an option somewhere to turn this off im sure – cant remember how off the top of my head
Just to the right of the close bar, you can choose to permanently opt-out.
The other factor is if the service can handle really long URLs such as
http://run.imac...2ZWFzLmh0bSA%3D
Here, TinyUrl is the best!
Oh, and I see that the TechCrunch blog software handles this link also well – congrats
This is a great post, and just goes to show that there is good reason why bit.ly is so popular, and that there is more which can be evolved from a very simple idea. The competition in this space is healthy.
I’ve been observing what’s been happening with Tr.im over the last weeks and I’m not impressed at all. They certainly shouldn’t be putting blame on Twitter or others and should be coming up with strategies to offer something different or find other USPs.
I once heard the phrase “Competition doesn’t kill a startup, they commit suicide.”
Angry at Tr.im? Seems like you are injecting bias into your stories.
Yeah, facts are bias.
I think he is talking about “while Tr.im came in dead last (no wonder nobody would buy it).”
But.ly…. God I love typo’s.
I’m still dumbfounded that someone, somewhere, took the time to compile this data.
Interesting…but this stuff is easy. It isn’t surprising that bit.ly (flush with recent capital) and ow.ly (a product of an operating concern) are able to solve problems that are generally easy to solve if you have some money to spend.
The real trick will be how these guys earn a dollar. Seems like bit.ly and ow.ly are the only ones getting close so far…
More and more sites are supporting the rel=shortlink standard which obviates the need for 3rd-party shorteners. It does this by allowing a site to advertise their own short links (e.g. http://tcrn.ch/) which can be auto-detected by browsers & other clients (e.g. Twitter). With last week’s WordPress.com announcement there are now over 115 million pages advertising shortlinks in the HTTP headers and/or HTML!
I didn’t realise the overhead would be such an issue (upto 1s!?!?) but it could well be reduced to negligible levels where the canonical and shortened domains are the same or where the content is served (ideally with a rel=canonical link) without a redirect.
Sam
I can’t do a test to see how fast TechCrunch’s own URL shortener (tcrn.ch) is but I can check it’s popularity. Just did a search.twitter.com on tcrn.ch and found 6 uses in the last 11 days.
Hope you all didn’t pay too much for it.
It’s exclusive for techcrunch, they don’t have it publicly available, only the blog posts can be shortened with it. They’re using http://totally.awe.sm/
I love using http://pie.im it is fast and reliable as well
I thought http://lnk.ly was a reliable site too!!
Hopefully the test system had the same/comparable lag to reach each of these servers, or the numbers might not make sense. E.g., from my laptop:
$ ping bit.ly
[snip]
—-bit.ly PING Statistics—-
6 packets transmitted, 6 packets received, 0.0% packet loss
round-trip (ms) min/avg/max/med = 12/16/33/13
$ ping tr.im
[snip]
—-tr.im PING Statistics—-
7 packets transmitted, 7 packets received, 0.0% packet loss
round-trip (ms) min/avg/max/med = 56/57/60/56
tr.im vs bit.ly: 57ms compared to 16ms
Of course, this doesn’t offset a huge overhead that tr.im is measured at, but a technical detail missed for the sake of accuracy.
Learn how to use traceroute. Ping times without knowing where the latency is doesn’t prove shit. Just sayin’.
But seriously… why all of the bias against Tr.im, TechCrunch? Did we forget that Tr.im is one of the most popular shorteners? And who really cares about 250ms in redirect time… obviously Tr.im users don’t. As well as having a better API than the rest. TC, don’t be even more biased when a company calls you out on it.
Are you new around here?
is.gd is good for me!
I almost never click a shortened URL. They can’t be trusted. Am I being overly paranoid or is this the worst trend ever?
you’re being ovr.ly pa.ra.no.id
who gives a crap about url shorteners? Any backend coder can write one in 20 minutes. The twitter related echo chamber is starting to melt my brain…
melt my brain…
my brain….
brain…
I understand the usefulness of these URL shorteners, but overall I find the entire category to be of little interest.
I prefer to us is.gd as well. It’s the shortest, it’s the fastest and I’ve never had reliability issues with it. Also, the ccTLD is from Greneda rather than Libya, which I feel is a better team to root for.
What, no RoR cracks?!?!
I wonder if Libya changes their domain name policy, what will happen to these cute domain names.
This is the reason AAfter search box creates a bit.ly and another aafter.us url to reduce the chances of failure.
Siteo.us is the best
is.gd is perceptibly faster than bit.ly
And in the recent spate of #WeLoveTheNHS tweets, bit.ly was the only one to falsely block legitimate websites.
I don’t agree with tr.im’s “we are/are not shutting down” antics, but wow, you really do have a vendetta against them now.
TechCrunch has gotten arrogant and mean. Time to get my tech news from better journalists (and there are PLENTY of them).
TC tested the reliability. That’s it. Why you get angry? You work for tr.im or what?
I don’t care if MA, MGS, ES, etc. get coffee for free for writing “Google stinks”… I have my own opinion. It’s good to see a personal view mixed with journalism… but finally the customer (reader) has its own brain to decide. Well, some don’t…
TC didn’t test anything; they reported on another site’s testing.
Another one for short address: http://qmc.tw
Watch the status bar…
http://pie.im/cb25
Every millisecond counts when building an empire
!
Or not
What about l.pr?
hate clicking on the ones that don’t take me directly to the link .. but to there lame tool bar .. ugh
and so often they don’t open anything
Was BudURL used in the comparison?
I know l.pr is pretty new, but I would like to see some info on it as well. I have been using it like crazy and it seems pretty quick to me.
there are so many url shorteners out there its difficult to choose, but i have been using http://www.Aafter.com, its performance its quite satisfactory.