Twitter.com vs The Twitter Ecosystem

John Borthwick, co-founder of Betaworks, parent company to bit.ly, twitterfeed, tweetdeck, chartbeat, and many other interesting web services, posted yesterday on "Ongoing tracking of the real time web …

Through these various Betaworks companies, John and the team have access to a tremendous amount of data and if you are interested in this subject, you really should read John's post.

I develop many of my theses based on what I see happening on this blog. And I've been seeing something on this blog that has gotten my attention.

Traffic is way up to this blog in the first half of January. This blog has seen as many visits in the first half of January as a normal month.

Sitemeter stats

So I went to Google Analytics to find out why. And I didn't see anything particularly new and different in the first half of this month.

Goog analytics
 

But that direct number bugs me so I sent John an email to see what I could learn. The first thing I learned is that he was planning a post (link above) on this exact topic. And he sent me some data on the clicks to avc.com from bit.ly in the first half of January. Here's a snapshot from John's email to me:

Email from john
 

Now, where would google analytics be capturing those 35,147 clicks? Well Twitter.com for sure. But that's only 7,567. Could the other ~28,000 clicks be in the "direct" number? I am absolutely positive that a bunch of them are.

But think about this for a second. Of the 35,000 clicks I got from bit.ly in the first half of January, only 20% of them came from Twitter.com. So exactly how big is Twitter.com vs the Twitter ecoystem?

Well, let's go back to John's post and pull my favorite chart out of it:

Twitt+eco
 

John's chart estimates that Twitter.com is about 20mm uvs a month in the US (comScore has it at 60mm uvs worldwide) and the Twitter ecosystem at about 60mm uvs in the US.

That says that across all web services, not just AVC, the Twitter ecosystem is about 3x Twitter.com. And on this blog, whose audience is certainly power users, that ratio is 5x.

Just to double check, because this is a seriously big deal, I checked all the links I bit.ly "ized" this past 30 days. Here's where they were clicked on:

Bit

So the links I put out into Twitter in the past 30 days generated almost 39,000 clicks. Nice. But only 10,000 of those clicks happened on Twitter.com. The rest happened elsewhere in the Twitter ecosystem, including Facebook which is part of the Twitter ecosystem when they showcase a post that is generated on Twitter, as all of mine are.

So that's a 4x ratio. That's a good double check. Whether its 3x (John's post), 4x (my links), or 5x (incoming traffic to AVC), it is clear that there's a big difference between the two.

My point is this. You can talk about Twitter.com and then you can talk about the Twitter ecosystem. One is a web site. The other is a fundamental part of the Internet infrastructure. And the latter is 3-5x bigger than the former and that delta is likely to grow even larger. 

Reblog this post [with Zemanta]

#VC & Technology

Comments (Archived):

  1. andrerib

    Where Facebook fails compared to Twitter (in the long run) is that the Facebook ecosystem is a closed one and will eventually suffer from that, even with their huge number of users.

    1. fredwilson

      kind of like iPhone vs Android. facebook and iPhone are not closed though. Zynga has built a very large business on Facebook and look at the hundred thousand plus apps on iPhone and the money all of those developers are making.but your point is ‘sort of open’ vs ‘fully open’. and i am with you in terms of long term sustainability of model.

      1. andrerib

        Yes, exactly. Facebook & iPhone are not closed but they are limited in ways that twitter isn’t. I can build any application on top of twitter and make it available on the Web for anyone to use without any restriction.It will be interesting to see analytics starting to include data from services like bit.ly. That should be happening asap as the percentage of people using ‘apps’ on the ecosystem is getting bigger and bigger.

      2. falicon

        ‘sort of open’ is exactly right and exactly the large term problem as I see it…though it’s probably not obvious, right now they are in the classic middle which we all know is a horrible place to be…they (Facebook) have an API that’s open enough to let some people have success, but isn’t really open enough to everything/anything people might want.As an example, Facebook offers access to the stream, but they don’t offer up very much of it (from the Stream.get documentation “you can return the last 50 posts from that user’s profile stream (Wall) for the last 180 days”)…without a way to page or get more, a client is forced to either ping the service like crazy or simply not get very much of the stream (even a casual facebook user with limited friend connections is likely to have 50+ items pass through their stream in just a couple of hours — but imagine the number for anyone with a large set of connections!?)Contrast that to Twitter’s API where you are free to pull back as much history as you want via pagination (well technically it’s as much as they have archived, but it’s a massive amount for most people — and their default is 200 per page to start which is already MUCH more useful than 50).So, in my opinion, right now one gives the facade of being open (I think primarily so they can continue to get the press and at least partially appease developers) and the other actually is open (though they don’t really bother to market/push this fact very much yet)…It is an interesting situation to see play out though…always fun to be a part of the early days…

        1. Mark Essel

          It’s the dynamic unpredictability of the system that draws our attention. Sports for geeks is techno trend spotting?

    2. awaldstein

      Facebook, unlike Linked In, is aware of its value and playing with ways to be open IMHO. The incredible power of its open demographics for the advertising and social games is too massive for them to not find ways to extend their hub and stem traffic flight before it starts. Making wall posts searchable is the first step.

      1. ShanaC

        I think they are aware of thier value, I’m not sure if they know how to leverage it. They want to be a one stop place (and they are) they know they have to evolve (and they are).Into what? It’s target of the social web. It’s not bad to be target, but there are moments I show up and I go, is this a photosharing site, a place to play a game, a place to have a chat, or what?It might be easier for them once they know what their identify is (even if their identity is to be target) and to focus in on that…right now life is confusing.

        1. awaldstein

          Hi ShanaI think they know what they are doing and understand the immense power of what they have. They are one view of the perfect intersection of global, social and mobile.

          1. ShanaC

            Doesn’t mean they are not the Target of the Web. Target of my in person life makes a lot of money last I checked. Doesn’t mean I always go to Target.

        2. Carl Rahn Griffith

          Facebook increasingly strikes me – and many others I speak to who have used it for some while – as a Swiss Army knife. Useful to have access to, but certainly not the tool for all one’s needs. Totally unscientific and a relatively small sample size (all pretty much early adopter types), all are spending less and less time on Facebook but increasingly seeing the potential and value in Twitter.Facebook seems to be predicated on it being a primary destination page, with a high level of stickiness/time spent therein. Not for me.It’s the old test time – if you could only have access to one of the two which would it be? For me it’s Twitter.

          1. awaldstein

            Good news is that we don’t have to chose;)

          2. Carl Rahn Griffith

            Spoilsport, lol – come on, play the game and give us your choice 😉

          3. awaldstein

            Honestly, they are just completely different. I have communities on both that feed and migrate to my blog and to others like this one where true conversations can take place.

          4. paramendra

            I just clicked over to your blog for the first time.

          5. awaldstein

            Thnx–I hope it is of interest to you. Just pushed a new post live a few moments ago.

          6. ShanaC

            Remember, while this may not be true of twitter, in terms of Facebook, I’m way earlier than the lot of you are, and I feel very locked into it.In fact, as a generational notice: I know a guy who quit and came back a month later… Facebook == mainstay of your 20s now!

          7. fredwilson

            i sure hope you are right

      2. fredwilson

        Yes indeed. Zuck is smart and paying attention

        1. awaldstein

          No doubt about it. I agree.

      3. paramendra

        “Making wall posts searchable is the first step.”I have been hammering this point.

        1. awaldstein

          It will happen cause it makes sense.Thnx for your thoughts on this.

  2. gardnersmitha

    These are some really interesting numbers, thanks for posting them, Fred.However, doesn’t some of this ambiguity about link referrals through twitter seem dubious for any sort of marketing or advertising campaigns that make use of twitter? Or maybe twitter has some internal data from the API that could be used to bridge this gap? Just a thought; fuzzy numbers make me nervous.

    1. fredwilson

      twitter is working on addressing this. they will need the assistance oftheir API partners which I think they will get.

      1. gardnersmitha

        I figured as much. It seems like you can piece it together pretty well but it involves a fair amount of manual labor at this point. We’re doing a fair amount of work with the API and overall it gets very high marks among our guys. Happy to support such an effort to figure out click-thrus if/when asked, I agree that most in the API network will.

    2. Mark Essel

      There’s an ecosystemneed for reliable real time end to end link tracking (heralds back to the content versus affiliate/ad monetary split). But it buts up on privacy, so I’m not sure it can happen.

  3. David Noël

    In the last 30 days, Facebook was the biggest referrer to SoundCloud, outranking direct and organic traffic. Twitter comes in fourth behind direct and organic traffic.

    1. fredwilson

      right. facebook is a juggernaut and should be the top referrer for many websites. but the question for you is “how many of those direct visits areactually from the twitter ecosystem?”

      1. David Noël

        Of course. I’d be interested to know that number and I suspect it to be high, given that all top referrers are social sites.

        1. fredwilson

          bit.ly has an offering called bit.ly pro that will give you that infoi’m not sure its commercially available yet but if not, it will be soon

          1. David Noël

            Yes, thanks I read about it and will give it a closer look. Thanks!

      2. Soren Macbeth

        This question is easy to answer via the use of google analytics url tagging. http://www.google.com/suppo…Bit.ly had a nice blog post about how to build a google spreadsheet that does exactly this here: http://blog.bit.ly/post/128…Although, I wish that it was more native to bit.ly itself. This is something that http://awe.sm/ does much better then bit.ly right now.

        1. fredwilson

          That is dopeness. Now I gotta figure out url tagging

          1. Soren Macbeth

            You need to work ‘dopeness’ more into your daily vocab. It works well 🙂

        2. David Noël

          Very cool, thanks for that links, Soren!

          1. Soren Macbeth

            Happy to help.

      3. howardlindzon

        let me know when someone figures out stocks and markets and facebook. i wish facebook could be our number 1 referrer.

        1. Mark Essel

          You consider a Facebook app that will help discover folks that are passionate about stocks?

    2. Mark Essel

      Social links are unsurprisingly coming out of the larger (and growing) Facebook ecosystem. But I would be hardpressed to find other Facebook users serendipitously that share my interests and passions. The nature of Twitter (and other default public services) allows developers to look for new opportune links between users. Much like social search where users don’t have to know precisely what they’re looking for, novel social (business/personal/hobbies) links can be incredibly valuable to the user. That’s what we’re betting on at Victus Media.

      1. David Noël

        In our case, we have a pretty popular Facebook app that our users (mainly artists and music professionals) love. The fact that Facebook is the top referrer as opposed to MySpace which comes in fifth, is a striking one.

        1. Mark Essel

          I can’t deny the power of such a connection. If your users find you on Facebook, then you have to optimize for their preferred connectivity. Totally understand (and congratulations on the traffic/usage!)

          1. David Noël

            Thanks, Mark!

  4. kidmercury

    the big question is what happens when twitter starts to aggressively monetize, can they do so in a way that enriches their ecosystem (i.e. google adsense) or is it going to be a tax on the ecosystem, or is it going to be independent of the ecosystem (though i doubt that is even possible). or who is it going to tax, who is it going to enrich.and speaking of twitter i saw a link tweeted by @chartreuseb regarding thought crime on twitter: http://www.p2pnet.net/story

    1. fredwilson

      well i’m privy to what they have in mind so i’ll keep my mouth shut. thegoal is thisit will be fascinating, non-traditional, and people will love it.

      1. Mark Essel

        Whatever form Twitter $$ takes I hope it’s a huge success that gives me plenty of opportunity to work on social tools and ultimately virtual personalized assistants. Bigger firehose helps.

      2. falicon

        That’s a good way to set the bar high…anything short of MAGNIFICENT in what they do now and I’ll be disappointed 😉

        1. fredwilson

          Those are not my words though. Its a quote from dick costolo at real time crunch.up. I agree that the bar is high. But I think its always been on this one

    2. Mark Essel

      They’ll enrich partner developers, and ride the sweet network effects if their mojo is well planned and executed 🙂

  5. LIAD

    were you surprised by the numbers? i would have expected the difference to be larger3-5X multiple isn’t that much – am impressed that .com is managing to hold its own so well

    1. fredwilson

      i still use twitter.com when i’m on a laptopbut when i am mobile, i always use third party clientsi think power users migrate to the apps and most light users and new users will be on the .com

      1. Mark Essel

        Ha, so “old school” Fred. Surprised you aren’t SMS texting your way most of the time.I’m split about 70% mobile client (tweetie), 10% desktop client (tweetdeck), 20% twitter.com.

        1. ShanaC

          I’m actually wishing I could move away from Tweetdeck. I want it to move to UTF-8. It looks like it is currently encoded to be ASCII displays, and I know people who tweet in non-Roman character languages. Their tweets show up as nothing…

        2. fredwilson

          I used to read and tweet via sms all the time. But then socialscope gave me an sms tab of all the people I followed exclusively via sms and that got me off that behavior

          1. paramendra

            “But then socialscope gave me an sms tab of all the people I followed exclusively via sms and that got me off that behavior.”LOL

  6. NicolasVDB

    “Now, where would google analytics be capturing those 35,147 clicks? Well Twitter.com for sure. But that’s only 7,567. Could the other ~28,000 clicks be in the “direct” number? I am absolutely positive that a bunch of them are”Any idea was the url referrer was picked up by g analytics for some but not for others?

    1. mydigitalself

      It depends on how the traffic is referred. A lot of it is done by redirects that don’t pass through the referring URI and GOOG sees this as direct traffic.

      1. NicolasVDB

        It seems the bit.ly passes twitter.com but not the other ones as referrers, I wonder if that’s a technical issue or a strategy (to sell the premium services)

    2. paramendra

      Google Analytics does not capture bit.ly clicks, but does it capture goo.gl clicks? It would be nice if it did.

  7. ewiesen

    One thing about this observation that perplexes me:You note that twitter.com has 20 million uniques and estimate that the ecosystem has 60 million uniques. That implies that there are 40 million people using twitter clients of various stripes, all of whom are monthly actives. But is that true? Which of these clients have 5M+ users? This may or may not be an in-bounds question, but it’s counter-intuitive to me.

    1. fredwilson

      We can’t mix up uvs with usage. My 3x to 5x is usage but not necesarily uvsThe only uvs I cited in the post are the comscore numbers and the numbers from john’s chartComscore’s numbers are for twitter.com (20mm US and 60mm worldwide)For the post I used clicks as a proxy for usage. I think its a good oneThere are many power users who don’t ever visit twitter.com and they are the ones who drive the vast majority of clicks via their usage and their retweetingHope that clarifies it eric

      1. ewiesen

        Yes, thank you. If usage is being measured (rather than number of people), the multiplier seen with third-party clients makes perfect sense. I don’t know if I qualify as a power user (~2100 tweets) but I suspect I do, and I’ve probably made no more than 1% or 2% of those tweets via the web interface. as you note (and as I’ve read and observed), the power users account for a large volume of content creation, so the observation that usage skews heavily away from twitter.com is indeed intuitive to me. Appreciate the clarification.

        1. fredwilson

          That’s why this post isn’t particularly surprising to power users. But the mainstream media doesn’t understand this

    2. fredwilson

      one more thing on this eric. i think of twitter as a system whose value is the traffic it generates out, like google. so the thing i like to focus on is traffic being generated to third party services more than unique visitors. they are linked, for sure. but i think clicks out is the most important thing to measure in the twitter ecosystem.

      1. Mark Essel

        The best tools/services will likely do the same. Portals out that folks appreciate enough to revisit.

  8. Mark Essel

    Interesting find. My concern (if I owned Twitter shares): Does the 3-5X multiplier need Twitter as a centralzed host (the Internet brain) or can they function just as well using an agreed upon standard not owned by any single corporate body? Aka is a variant of the Twitter API the “real time” protocol.I’m very interested as a twitter developer, but for me it’s all about data accessibility and an API. Ideally information systems will be source insensitive. The processing can work just as well on Facebook statuses as it can on Twitter statuses, the Huge difference is easy public access to most accounts.Why stop with user statuses? In theory the meta data that is applied to user tweets can be derived from object tweets as well, enriching network connectivity.

    1. falicon

      I think it would be VERY hard to work Twitter out of the system at this point…mainly because it would mean, in many cases, training the average user (who is the person putting data into the Twitter ecosystem) to do something different (not to mention getting all the developers to update their apps to use a different pipe — for no real reason/advantage)…So I don’t think we *need* Twitter as the centralized host, but it doesn’t really seem like there’s a big enough reason (yet) to remove them since they are already there…as long as they continue to be open and available to developers, I don’t see them losing that central position…

      1. Mark Essel

        I was considering alternative sources (not removing Twitter). Some type of super distributed real time aggregator that had many local hydrants to tap the stream from.Is a syncronous replicable real time stream possible, and who would pay for it?I’d use it as a dev, one API to rule them all where many corporate data providers feed the pipeline (and are rewarded by the network).Is it too much to ask for an omniscient real time state machine ;)?

  9. Kevin

    To my own surprise Twitter has become part of my morning routine, along with e-mail and NY Times online. I’ve noticed more and more that I click links suggested by people I follow at a pretty high frequency. Higher, I bet, than pretty much any other source (including FB in my case). I think it’s specifically b/c I know so few of the people I follow on Twitter, yet I find who they are / what they do interesting. FB is people I know personally, and while that’s cool, it’s not as interesting in many ways as the people I get to know via Twitter — such as Fred and various readers on this blog.

    1. Mark Essel

      Totally agree Kevin. It’s the novel network relationships that I value as information sources so much now. Been busy working to enhance that experience with tools I’m helping build at Victus Media.

    2. awaldstein

      Kevin, you’ve hit on a key item.It is very hard to efficiently extend your network on Facebook. It is about people you know and are connected to. It’s hard to ‘meet’ new folks and make new friends.Twitter as a market extension paradigm is remarkable as a means to meet new people, extend your network and drive meaningful new friendships and contacts. Twitter is the metaphorical ‘pebble in the pond’.

      1. Mark Essel

        Serendipitous social edge extension.Kevin Kelly predicts the edges are where the most successful mutations to species can occur in relative safety. Intriguing that the edges are the closest to chaos.

        1. awaldstein

          I paused for a dictionary check Mark ;)Serendipitous or not, the extension is by dint of understanding a key fact. FB works because people share their interests with friends and that data is the key for targeting both direct and trend wise. More traffic, more friends, more sharing = more data = more efficient advertising and new ways of connecting goods with needs.

          1. Mark Essel

            Thanks Arnold for the reminder of what’s important. I shouldn’t discount the power of any medium that can enhance user sharing and network strengthening even if I don’t absolutely love the framework. In fact I’d very much like to speak Facebook-ese when it comes to social data processing. We have the tech to do so (working external interface app), it’s just not on our short list at VM.

          2. awaldstein

            Mark–certainly no criticsm, just clarification on the value and power of the FB phenomenon as a sandbox for behavioral change.You are right thinking about open as the future and honestly, one of the folks I look to for sensing where the future lies.My DNA is about making markets and connections between people and people, and how that impacts brands and commerce. That’s always my bias.We make a healthy foil for one another cause at our core, we agree in principal just from different points of view.

          3. Mark Essel

            “We make a healthy foil for one another cause at our core, we agree in principal just from different points of view.”Certainly!Never worry about being critical of an idea I fire out. I thrive on critical feedback from folks who’s opinion’s I respect.

      2. Kevin

        The ability to leverage what other people know, through what they link to, and who they know, by conversations like this one, is very interesting and useful.

        1. awaldstein

          Absolutely.We all value info from people we meet in environments that engender trust.And the basis of advertising or resource matching is to find the most trusted and reliable connection.

      3. ShanaC

        A) Skipping pebble in the pond.B) It’s one of the reasons I’m finding the transfer over from Facebook to say LinkedIn so awkward. They seem very the same. Except Facebook is more fully functional. All the same people. I actually could get more done on Twitter in terms of “meeting people”

        1. awaldstein

          Yes, to meet new folks Twitter and spots like this are the place to be if you are comfortable in this environment.

      4. fredwilson

        like you and me arnold. Not sure when we happened upon each other. But I know where you have coffee and what wine you like

        1. awaldstein

          And I know that I need to try the Mac and Cheese at the City Bakery next time I do a meeting there 😉

          1. fredwilson

            that you do

  10. HowieG

    Very interesting data. One issue I have with the ‘ecosystem’ and twitter in general is that I have to assume that most active Twitter users following 100+ accounts only view/read 20% (I am being generous with the 20% number) of all tweets sent their way just due to the massive volume (Facebook’s live feed experiences the same or worse views). So where does the ‘vacuum’ of unviewed/read tweets get assigned in the ecosystem?

    1. Mark Essel

      It’s a form of filtering. Links/shares require multiple reshares to get in front of wide band followers eyes.On the flip side I only follow 50ish people and really enjoy catching a better view of their streams.

    2. fredwilson

      Nowhere. If nobody clicks on the link, its off into the ether

  11. MattCope

    Great forensic accounting!When you can make numbers tell a story, it’s a beautiful thing.

    1. fredwilson

      Particulalry the story you want to tell

  12. howardlindzon

    seeing the same ratios at stockwtits

    1. fredwilson

      That’s excellent. Now we’ve got a triple check. Thanks howard

  13. Aviah Laor

    Another post in the “social web will be finally distributed” area. Again, the real hero of this post is DISQUS which was built that way. They can really re-shape social web, once they move on and broaden their product category. Guess they are not the only one working toward that goal.

    1. fredwilson

      The only people who understand disqus are those who use it actively. So many don’t get it

      1. awaldstein

        You and Aviah are onto something that we’ve all experienced. Participate in blog communities and comment and Disqus is like adrenalin to the athlete. To those that don’t, they just don’t get the sport. I would venture that there is a huge win for Disqus if they can bring the spectators into the field somehow.Might be an interesting discussion at some future time although Disqus I’m sure is figuring this out in real time.

      2. Prokofy

        what is the part of the Twitter ecosystem that generates the most clicks then? Is it some mobile phone application? or is it Tweetdeck? Or what is it exactly?As for Disqus, I will get it more when they can make a widget that I can put on my own website that has a window into all the comments I make everywhere, maybe scrolling by in real time.

        1. fredwilson

          Mobile

  14. steveobd

    Fred – I came to a similar conclusion in my own analysis – although I have seen traffic outside my blog, here’s how it applies there: http://steveobd.blogspot.co…I wonder if like the feed reading ecosystem, we will see it go from rich standalone clients back to web clients. There aren’t any real standalone feed readers anymore.The big wildcard is mobile – even now I am seeing about 8% of my traffic coming from mobile twitter clients. I personally believe it will take a long while for HTML5 to catch on and we will see mobile applications rule the roost here for some time.But anyway, look for that gap to widen, especially because of mobile.

    1. fredwilson

      Thanks steve. Speaking of mobile, I’m on my blackberry in the subway but I’ll check on your post laterBtw – when did goog analytics start reporting feedburner clicks? That’s relatively new, right?

  15. Tony Karrer

    I’ve seen some similar trends on my blog, but looking across at other kinds of sites that we work with, I’m convinced that twitter clicks are not equal to other sources:http://elearningtech.blogsp

  16. Christian Brucculeri

    I’ve been noticing the difference become larger as well. From a “marketing funnel” perspective, I feel like a user:1. I introduced to Twitter.com – awareness2. Creates a profile and tweets from it -trial3. Downloads an application and stops going to Twitter.com – conversionIt’s perfectly logical for Twitter.com’s numbers to go up and down- avid users of the platform rarely go to the dot com at all.

    1. paramendra

      Disqus is my Twitter client of choice. 🙂

  17. GlobalPatriot

    The interrelationship between social media platforms is creating this multi-dimensional landscape where blogs are tweeted and tweets are posted to Facebook, which itself has profiles, pages, groups and causes. Videos (YouTube) and photos (Flickr) add to the mix too. It is quite fascinating to explore how links and referrals get passed up and down the digital food chain.

  18. Nitin Mittal

    Historically, how much direct traffic was AVC receiving before you actively started using bit.ly? Unfortunately, it is not that black and white since, I assume, you have been using it for awhile and the sight has been growing in popularity. Maybe you can back out a chunk of “organic” direct traffic (adjusted for growth) and see if that provides another sanity check.

    1. fredwilson

      I’d have to go back at least a year, maybe 18 months. Not sure that’s going to deliver a lot of insight

  19. Warren Whitlock

    Twitter is much bigger than Twitter.com

  20. ShanaC

    I’m wondering how much link forwarding is happening in different websites within the twitter ecosystem. Say link a appears in twitter on tweetdeck and then ends up on facebook and linkedin- I want to know where these links are appearing that are causing the rise up.it’s also where the links are going. I’ve seen you in various places…Is it worth it to try damming certain kinds of flows and opening others to direct links to go elsewhere, is what I am wondering…what is the best place for these links, or is there no best place, looking at these numbers?Or is that a little too far out there?

    1. fredwilson

      I feel like there is no best place which is why I want them to be everywhere

      1. ShanaC

        That may be for you- but if we are qualifying that say bit.ly links are much like any sort of information flows, then they should obey information economic laws, and I could imagine someone wanting to dam them and reroute the flows for the sake of benefits that come from having “more exclusive sources” of information. I’m not necessarily of the school of thought that every click or look is of equal quality. It might be better to have less refers if those refers are more engaged.Then again, I’m still trying to decide what a link is…

  21. LaTease Rikard

    Fascinating article; I have been using Hootsuites stats report and realized the influence of one of my posts…it received over 10K clicks in one day. When I saw that, my heart literally jumped out of my chest! Twitter, is turning into a fascinating avenue of traffic for me. Thanks for reaffirming what I already knew!

  22. Darren Herman

    The post = interestingThe comments = inspiringLots of solid conversations being discussed. I was thinking much of what you (fred) wrote in the initial post as I was writing about my Twitter Tipping Point (http://bit.ly/7OZLxw) which was really about coming to terms with the ecosystem.The things I’m keeping in mind with Twitter:1. They are central to a large ecosystem – becoming the plumbing2. Do they stay neutral or do they want to own more than the plumbingI view Twitter as MS-DOS (http://en.wikipedia.org/wik…, whereas you can get things done within DOS but a slick interface on top of it would allow you to do just so much more – quickly, efficiently, and effectively. Those UI’s are just being built now (Seesmic, Tweetdeck, Tomzy, etc) and what’s interesting to me is going from beyond *just* a UI, to an actual value-driver.

    1. falicon

      I think this is a great way to view the current world…there are lots of people tackling the problem of a ‘slick, value-added, UI’ for Twitter…none seem to have hit it perfect yet, but it’s surely coming.How and what Twitter does with their stance as the integral plumbing is the real hot topic that we are all cautiously monitoring…

    2. fredwilson

      Great comment darren. I’ll read your post when I come up out of the subway

  23. vadhri

    I liked the article. I am wondering if we can go more granular – how much of the “off-twitter” data is driven by Mobile clients vs driven by other ecosystems like facebook vs independent blogger sites …the trends indicate mobile clients drive most traffic “off-twitter” ..

    1. fredwilson

      I think you are right. We ‘snack’ on the various web interfaces. But we ‘live’ on our mobile device

  24. Jan Schultink

    Sudden traffic spikes beyond an increase in RSS readers must be driven by specific posts. Did you analyze which posts were popular in the 1st half of January? The Nexus One reviews?

    1. fredwilson

      There were several that were big traffic drivers. But even the low traffic posts did 6k to 7k. Everything just amp’d up by 2x

  25. SharelOmer

    Great post, thank you.I love to see numbers and analysis, thank you for sharing this numbers with us, its not trivial.When you talk about “Facebook which is part of the Twitter ecosystem” do you have the numbers of how many people came from FB ? my guess is that most of your Twitter ecosystem came from there, if so it really make me think again about syndicating twitter to FB. (i dropped it since people keep telling me they don’t understand the RT/#/@ language)

    1. fredwilson

      you can see FB on my goog analytics and bit.ly analytics numbers in the post. Its not a big factor for this audience

  26. Meteor Solutions

    Another interesting post Fred. Love the data and it’s consistent with what our customers are seeing as well. What the data you’ve presented do not explain however is why your site traffic is up by so much this month. Conversations and sharing that occurs on client apps like IM, Tweetdeck, etc. is kind of like the “dark matter” of the Web as there historically hasn’t been a good way to isolate the role they play in driving traffic to sites. It all gets dumped into the “direct” bucket as you pointed out – annoying. While it’s absolutely true that the “direct” traffic source is significant, it doesn’t stand to reason that this is the driver of the sudden increase in traffic to your blog unless it’s the only traffic source that’s showing a huge spike from Dec to Jan. Those client apps have been around for awhile now with adoption increasing at a predictable rate.So, I’m curious to know if a benefit of this analysis is that you’ve isolated the source(s) of the sudden increase in traffic. An immediate 2x lift in UVs doesn’t happen without a clear cause. What are you or your audience doing differently in January that’s bringing you more UVs? it looks like your PVs/UVs are essentially steady so you’re attracting a whole bunch of new users it appears. What do you believe is the reason for the increase?

    1. fredwilson

      i am still scratching my head on that one Ben. most referrers are up about the same this month. and i see the increase in multiple analytics packages.

  27. james

    I came from an article posted by TechCrunch that stated this blog as one of the top VC Blogs (the first I think?). “Every so often, venture capitalist Larry Cheng puts out a list of the top VC blogs.” You can see the article here: http://feedproxy.google.com

  28. Peter Renshaw

    “… John’s chart estimates that Twitter.com is about 20mm uvs a month in the US (comScore has it at 60mm uvs worldwide) and the Twitter ecosystem at about 60mm uvs in the US. …”I need a legend for the jargon. UVS is an acronym for ?

    1. fredwilson

      unique visitors per month to a given website

      1. Peter Renshaw

        “unique visitors per month” thanks @fredwilson knowing the metric units helps.

  29. Liz

    This is fairly consistent in the numbers I’ve seen. In Fall 2008 (I believe) Twitter’s blog said that 40% of users used Twitter.com to send out updates and lately I’ve been seeing ~30% figure quoted. I’m still a devoted website user (no API limits, instant access to profile pages) but I see an increasing number of Tweets sent in via Blackberries these days.It could also be that the users who are interested in your blog are more likely to use Twitter clients than the kind of users who access the website.

  30. Ed Freyfogle

    Offsite is the new SEO

  31. Guest

    I would love to know more about the session times, repeat visitors and bounce rates on the traffic coming in if you are willing to share sometime!

    1. fredwilson

      i’ll do that. bounce rates are high (but they are for all blogs) and session times are about 1 minute. repeat visitors are very high

      1. patricia

        That’s really good. I’m really curious on the session times coming off the social networks in general — the traffic can be valuable and convert to audience. I wonder what makes it tick.I think blogs have lower session times by the organic nature of the content but I’ve wondered if UI plays a part — most styles and templates aren’t really in sync with the standard stuff proven to keep people on pages. It’d be cool if someone did a UI for blogs based on research and stats regarding what users do when they hit the site, what keeps them there, etc

        1. paramendra

          I have a feeling a lot of people come over, read Fred’s blog post they came to read, and walk away. They skip the comments section. I think that is a mistake. But then maybe I am overestimating the rest of us here. 🙂

  32. M. Edward (Ed) Borasky

    So what’s the takeaway here? What’s the actionable insight? What’s the recommendation? Over the weekend, I consolidated six blogs into one and started looking at my traffic using Clicky. I had a huge burst of traffic a couple weeks ago when I tweeted a blog post on the Twitter Streaming API.But I had another huge burst of traffic yesterday when someone picked up a post of mine on the R programming language and posted the link on Developer Zone. That came totally out of the blue, but that’s why I use a real-time tool like Clicky – so I can craft followups based on what’s actually happening! I would never have thought to announce things on Developer Zone. Now I post everything there.My point is that this whole online marketing / social media “ecosystem” is full of surprises like that, and “conventional wisdom” will lead you astray if you let it. The truth is in the real numbers you measure yourself.

    1. fredwilson

      the simple actionable insight is don’t believe your twitter numbers in your google analytics dashboard. they are 3-5x higher than that. invest more heavily in that channel if you can.

  33. Danny Sullivan

    Yep. I covered some of this last year in “Is Twitter Sending You 500% To 1600% More Traffic Than You Might Think.” You can Google it for the link. You’ve definitely got to understand that Twitter is far more than the site and that Google Analytics sure ain’t counting it all (I had comments from Google on this, in that article).

    1. Ed Freyfogle

      Danny, I’ve heard several media owners suggest that Google Analytics inflates Google visitors. (based on comparisons with their own log analysis and other metrics tools). Any thoughts on that?

    2. fredwilson

      that was a great post Danny. i am still trying to get closer to the real numbers and suprisingly it hasn’t gotten any easier

      1. Norbert Mayer-Wittmann

        Mathematics is a language, but it isn’t as easily “grokked” the way natural languages are. For example, it would be quite difficult to form a logical expression that is the equivalent of the English word “twitter”. Linguists have tried to do such things for decades (do you know who Terry Winograd is?) but they’ve had very little success. If you’re really interested, you might look up the philosophical tradition that goes back to Gottlob Frege and Ludwig Wittgenstein. (and good luck — it’s pretty heady stuff 😉

  34. Norbert Mayer-Wittmann

    Hey, if the guys @ bit.ly charged 1 cent per click they could get rich quick!:D nmw

  35. Jimmy Gilmore

    I’ve noticed the same thing. Clicks come from apps, mobile apps, aggregators and usually not Twitter.com. Actually not seeing much from Facebook.

  36. Chris Hill

    Interesting web statistics regarding twitter driving web traffic

  37. William Mougayar

    Very telling stats. Look at this post as another validation: 123 Comments, 455 Twitter-ecosystem Reactions (3.7 X). There is a parallel to the ratios being discussed here. I wonder what the Facebook ecosystem analogy might be like.

  38. Olivier Coudert

    I am glad to see that these data correlate to what I could find and published in http://thenextweb.com/2010/….Basically, I found that between 2/3 and 4/5 of the twitter traffic is not due to the website, thus that the Twitter ecosystem is between x2 to x4 larger than the twitter.com traffic, and that the gap is widening. The more engaged the user, the less he/she uses the twitter.com website.

  39. vivek1234

    The ability to leverage what other people know, through what they link to, and who they know, by conversations like this one, is very interesting and useful.

  40. Bhushan

    The best tools/services will likely do the same. Portals out that folks appreciate enough to revisit.

  41. Rich2001

    1) This is my third time linking to this article (2 mobile + 1 laptop = 3 apps). Could this be part of the 3x?2) Is the Twitter ecosystem feeding on itself? Are the same power users talking to each other, and creating a feedback loop of links? (80-20 rule, in this case the 20-80 rule).3) There is a ‘generational reality gap’. My children (16-21) are not using Twitter. I do not know a single over-40 who is NOT in tech/media who IS using Twitter. (Allows for even more growth)4) avc.com is going viral. Amazing word of mouth. As much as I appreciate your posts (start the fire), the comments are inspirational (keep it burning). Wish there was a music community as high level.5) And yes, I finally joined Disqus. Wish it worked with wordpress.com.

    1. fredwilson

      i know kids and grandparents who use twittertwitter’s demo is much wider than you might thinkcomscore media metrix has twitter overweighted vs the internet at large inthe sub 24 and the over 45 crowd and underweighted in the 25 to 44 demo.

      1. Rich2001

        It would give me cause to reexamine those numbers when a real world sampling of a small, but very diverse group, offered such a contrarian usage view.

    2. paramendra

      Disqus and WordPress, tell me about it.

  42. Nathan Vingoe

    That is really useful information!! And congratulations on the click increase..Anyone who gets links from twitter will have seen the direct links number in their stats, I was just annoyed there was no easy way to know if that’s the emails, offline ads, or twitter.. Mine aren’t power users, so the 3x figure is a really useful figure for me.. and also validated from Johns work too..the reading between the lines figures from seeing your stats is also a nice side effect of this post. Nice to see others’ data, not something most would share.. So thank you for that too..Interesting to see that your blogs go out on feed burner to 61k and you get 16k clicks back from them.. as you can tell, I am easily amused with a few facts and figures..So the circle completer, how many google/ad clicks do you get.. revenue is irrelevant really, as that’s different to each genre of blog, but would be interested in how many convert to clicks…

  43. paramendra

    This is why I have not believed those who have been “observing” that Twitter’s growth has flattened out. A lot of the growth has been happening in the ecosystem. Twitter is utility. It has become fundamental to the web experience.