Search vs Social

I was at a meeting yesterday regarding the ongoing online piracy discussion and the conversation turned to search as a source of traffic to sites with pirated content. I stopped the conversation and noted that search isn't what it used to be and pointed out that many websites get more traffic from social than search.

Here at AVC, it is no contest. Here's the top ten traffic sources to AVC in the past thirty days:

AVC traffic sources

Google/organic is search. Direct and feedburner are regular visitors. Everything else (Stumbleupon, Twitter (t.co), Hacker News (news.ycombinator), Techmeme, google/referral, Facebook, and Linkedin are social. So if we break the top ten into three categories, direct is about half of the top ten's traffic, social is 40%, and search is 10%.

This blog isn't normal in a few ways. The fact that Twitter generates 13x Facebook in traffic is one example of that. And the very high level of direct/regular readers is probably a bit unusual too.

I'm curious if anyone is aware of a broader study of traffic sources on the Internet and how search and social compare these days. I suspect that they are neck and neck across the entire Internet or possibly that social has surpassed search. But I have not seen that data and I'd love to.

#Web/Tech

Comments (Archived):

  1. Rohan

    Here are mine for the past month.Google search is still significant. Close to 50%. No idea what the data is overall. But, my guesses are as follows – – Social is likely to be higher if you are famous. This means a correspondingly high circle of influence. An easy way to measure that is the number of twitter followers. (in the difference between my baby blog and yours that’s 150 vs 100,000+) – Famous/authority = more people want to quote you/say they’ve read your opinion => More shares on facebook, more up votes on Hacker news, Stumble Upon and the like. – If you are starting small and building your audience, your growth is largely (in my case, 50%) thanks to my own circle of influence. The social aspects is from places I engage in – HN, HBR and the like. – And as the audience grows, I expect social will become much bigger over time. For now, I’m still very thankful to, and dependent on Google search..

    1. Dave Pinsen

      “Social is likely to be higher if you are famous.”In general, social media seems to be a lot more effective at enabling a famous person to engage with his fans than in enabling a non-famous person to become famous. Neil Gaiman, for example does a great job of leveraging Twitter to drum up fans at his appearances, but he built his fan base via old school media (graphic novels, novels, TV, etc.).

      1. jason wright

        Yes, and as discussed in the Louis CK post some weeks ago. However, this may be a transitional phase. As this new channel matures purely social celebrity will evolve. It takes density.

        1. Dave Pinsen

          There have already been a small handful of examples of that, but I suspect purely social celebrities will be as rare as celebrities of the same rank in RL.

          1. jason wright

            I think the definition and character of ‘celebrity’ will change with the rise of social. Conventional celebrity is almost ‘imposed’ as a one to many process, hence the phenomenon of the celebrity famous for being…a celebrity (a mass media one-way channel manufactured icon).

      2. Rohan

        Very true, Dave. That’s a great point! 🙂

    2. William Mougayar

      “Social is likely to be higher if you are famous.” Good one Rohan. 

      1. Rohan

        Thank you sir. Nice to know I spout wisdom every once in a blue moon..haha

        1. William Mougayar

          There is a quote that I couldn’t find…paraphrased as “You will be famous before the age of 40″…or something like that. Can you find it? – posted via http://engag.io

          1. Rohan

            You spiked my curiosity, William.. couldn’t find anything though?!

  2. jason wright

    When Google adds its Plus social signal to its organic search results there’s no further need to suspect that social has surpassed conventional search.

  3. Rayhan Rafiq Omar

    My blog doesn’t get much traffic, but I find Twitter completely outstrips facebook as a referrer of traffic (and every tweet is automatically/annoyingly posted to facebook).I also find engaged communities like this and Hacker News are great referral sources.http://news.ycombinator.com… 

  4. LIAD

    Google scrambling to integrate all manner of social signals into search is proof enough that this phenomena is widespread and growing.Rather than increasing their own relevancy and longevity by giving social increasing exposure in their search results I think it may backfire and end up hastening algorithmic search’s demise.Virgil, but more importantly the fountain of wisdom that is John Mason, (aka Sean Connery) in The Rock said it best – ‘”I fear the Greeks even when they bring gifts.”

    1. jason wright

      I think Google is worried.When Microsoft offered $45bn to buy Yahoo it was the right price,…but the wrong company.Will Android be the foundation of Google’s future social search?

      1. JamesHRH

        What is hard for goog is that their current business is so lucrative.FB will not generate revenue as efficiently as Adwords. My guess is Larry Page will not be able to stop himself from a desire to make everything be search,When did he get the JLM style seasoning ( multiple businesses w different core characteristics) that gives him the perspective to build a portfolio of online businesses that are effective but different?

        1. jason wright

          Yes, you’re right, Google is addicted to the sauce.Could Larry become the next Jerry?

    2. fredwilson

      or possibly allowing other search engines to gain share, like firefox, safari, and chrome have done against IE in the browser spacein the past month, direct searches per day have doubled on Duck Duck Go. is that a coincidence?http://duckduckgo.com/traff

      1. LIAD

        would be pretty incredible if the growth was a direct backlash against Google SPYW. Feasible I suppose as early adopters and techies seem to be the only ones to know what Google are up to and DDG is becoming a contrarian search engine.I havent read anything in tech world about google defections though. You’d have thought if techies were moving to DDG en masse it would have surfaced on the blogs.-if DDG manage the make a dent in search, it will definitely be a story worth telling.

        1. fredwilson

          approaching 1mm direct searches per day. if they can take that up 10x this year, they’d be doing 300mm searches per month. that would be a story worth telling. i think it is absolutely possible

          1. Elie Seidman

            It’s possible. Take a look at the above the fold results in Google for a search in a high value vertical. Measure how much of that space is occupied by Google’s own properties and ads. For a long time the debate among people who watched, or cared about, search was “will I rank on the first page and where on the first page at that”. Now the question is, how far down the first page (or onto the second) will you be pushed as a result of Google’s own use of the space. Google is a business – I understand that. And paid links pay and organic links don’t. If they want to maintain their 2x higher PE (vs Microsoft) there needs to be revenue and profit growth. But while they do that, they’re decreasing the quality of their search results and opening up the playing field to competitors. (simultaneous to the change in consumer behavior within search)http://searchengineland.com… 

          2. kidmercury

            IMO google’s search is still way better than everyone i’ve tried. more importantly, though, general search is more about distribution than about product. i also think that google thinks their strategy of ranking their own stuff higher is in many ways better for the user — perhaps not immediately but long-term i think it will be.  

          3. Elie Seidman

            I do think they believe that. I think that it will be very hard for them to provide the best solution in each vertical. Each vertical has folks who do that and only that. Those focused companies – startups or mature – will, almost always, be better at that vertical than a group within Google will. Especially as google matures organizationally, they simply won’t have a competitive advantage in better, hungrier, people and they’re teams will make worse products. The precedent of Yahoo, aol, and MSN will be very hard to avoid copying. There volume comes from having good search. If, within search, they favor inferior Google products over the competition… I think what we’re currently seeing is the future of organic Google. The first page in any valuable category is half google and the other half massive well known brands who are major mktg spenders – both on google and in regular advertising (TV). Google has to show those brands somewhere near the top because if they don’t the customer will start to doubt google. That doubt won’t be conscious – but it will start to happen. If a customer types in home power tools both Lowes and Home Depot basically have to show. Google does not the direction here – it merely follows the reality of the world. If google results don’t emulate reality close enough… But this means that for companies that want to be big on Google organic – what remains of it – the solution will be ever simpler: be big off of Google so Google is forced to follow. I’m short all SEO tactics that don’t adhere to this basic logic. No SEO, good or bad, will keep Food Network and Conde Nast from the top of most food and recipe related queries. Google basically “has to” rank them well because the real world ranks them well. The action for 2nd tier brands will be on pages 2 through N in organic. There is some volume there – perhaps more as the 1st page gets more commercial – but it’s obviously small in comparison. As an aside, the days of a company being made by Google as TripAdvisor almost certainly was and Yelp arguably was, are over.

          4. Aaron Klein

            Very impressed with what Gabriel is doing. Switching my browser to DuckDuckGo for the next week and I’ll see if I can make it stick.

  5. sigmaalgebra

    If social can beat search, really, without even trying, then search has a way to go! E.g., social ‘gets at’ the ‘meaning’ of the content, and so far mostly search does not. How to do something about that? Hmm …!So, having search be more ‘social’ is how search can catch up with social? No, not really! Social wasn’t even trying, but if actually try then can do better, still.Search done well can knock the socks off social for search and also much of old search.How to know such things? Hmm …!

    1. Dave Pinsen

      That must be your shortest comment here ever.

      1. sigmaalgebra

        Not now!

        1. Donna Brewington White

          Good humor for early morning.

      2. fredwilson

        ^10

      3. kidmercury

        i used to really dislike sigma’s long comments, but then i started making fun of them and now i enjoy for them that reason. i’m torn as to how i feel about his new short comment image.

        1. fredwilson

          i am so happy. he’s smart as shit. but i could not get through the long ones as hard as i tried.

      4. Otis Funkmeyer

        Dave, that was the lulz.

      5. kidmercury

        lol most liked comment in this thread…..hahhahaha

    2. JamesHRH

      Google has realized that people spent moretime visiting than at the library.Awesomely effective search will be more valuable than social, even though it will be a much smaller percentage of people’s lives.

  6. Pramod Dikshith

    I think the dynamics change when it comes to Commerce. People want the cheapest deal to buy anything unlike content which is more or less free nowadays due to influx of blogging platforms and eminent bloggers like you who give quality content free. So the decision now is to choose the best content of the lot and that is where social brings into picture. So far social has not been as successful in the commerce element. I think for commerce search will be a bigger share.

    1. awaldstein

      I think this is not as black and white as you portray. The growth of marketplaces is changing this paradigm. 

  7. awaldstein

    Depends, I think, on the type of business.My focus is engagement with my blogs. Traffic will come, engagement matters and is the core of my businesses pipeline. It is different for my clients depending on the business segment.For me, the average breakdown over the last year is 25% search, 50% referral/social and 25% direct.Two things I would add:-My two blogs are somewhat different. For wine, Facebook and Twitter are huge for me. For marketing and community, Facebook less so but overall social is somewhat higher. -It is not possible to have this discussion, about natural search, for those of us dedicated to Disqus to again make a plea for them to focus on improving their natural search capabilities within the comments. I would love to have Daniel have an open discussion on this. It’s important.

    1. fredwilson

      i’ve been told they are working on this

      1. awaldstein

        Good.They are a huge community platform and their community is the bloggers. There’s a lot of strong ties here to build on.I’m not a kvetcher. Not making this an open discussion is a missed opportunity IMO.

        1. Aaron Klein

          I agree. “Where Disqus should go” would be a great discussion for Fred to host here on AVC. You’ve got a huge community of power users here, most of whom then went and adopted Disqus for their own blogs.

          1. awaldstein

            I didn’t mean to pile on the Disqus folks.I’m their booster for certain. In fact, just convinced some buddy’s  blogs in Austria, UK and Italy to move over.I had a wine post last week on natural language for the mass market to share wine. It went sideways in a very active comment string with a bunch of notable ‘blind tasters’ jumping in in enormous detail with great info.I thought…will this info ever get found by search? Then Fred’s post…

          2. Aaron Klein

            I don’t know Daniel well, but I’ve been to their office and know one of their community guys. I think they’d welcome that discussion!

      2. FAKE GRIMLOCK

        THEM ARE.NOT SURE WHY NOT DONE YET.

    2. Avi Deitcher

      +1 on disqus. I don’t use them on any commenting platform as a content creator, but come across them time and again as a commenter. I love the idea, but have been frustrated by: poor functionality (being charitable here) on mobile; poor search; poor cut and paste to use elsewhere.They have some serious potential here if they leverage it correctly.

      1. RichardF

        spot on avi.I am becoming increasingly frustrated by the mobile/tablet experience on Disqus and given that tablet use is exploding it needs to be addressed now.

      2. awaldstein

        I’m a long-term supporter. Big believer.I’m less concerned actually about things Disqus doesn’t do and moreso that I don’t know where they are going.Comments and engagement are the currency of community. This commitment is two way with your chosen platforms.

        1. Avi Deitcher

          I am not sure I understand you. What does it mean, “where they are going”? I am not 100% convinced that they know, other than building a community and figuring it out.I think that the mobile platform is crucial to build a community. Far more interaction happens now by mobile (iPhone/Android/iPad) than desktop. If you cannot interact there, you hurt your community-building.

          1. awaldstein

            True…What I meant was that certainly they will build mobile and pad end points. Can’t be a commenting platform without that. It’s just a schedule and priority factor I presume.But in today’s world, every company and product is a work in progress at all points. Whether you are figuring out how to surface and display data or where you will go to meet market demand.More and more, I want to understand a company’s promise when I commit to them so I can decide whether to jump on for the long ride.Understanding that directional promise is what I was trying to articulate.

          2. Avi Deitcher

            Ah, I see. You feel you are making some commitment by using them, you want to know where that will go, not that they will screw it up (viz. Google, new privacy rules, no opt-out, Max Wessel’s post http://blogs.hbr.org/cs/201… )I would think that would be more of an issue for the community creator ( avc.com/Fred in this case), as opposed to the community user/contributor (i.e. you/me/charlie/grimlock/everyone)?

  8. Bavan Rajan

    I suspect that for many websites that get more “mainstream” traffic, the importance of search will be much higher – most people I know that I would classify as non techie still use google as their primary way to navigate around the web – sure they click on links on Facebook and twitter, but they really want to do / see something, it is all about google.

  9. William Mougayar

    Why are you considering Hacker News and Techmeme as Social?There is a blurring of lines between social and search distinctions especially as Google integrates social results (eg your friend liked or visited this one), and if/when Twitter makes their search a bit more friendly & open, then is that search or social?Add the fact that Social is really about People. The social media is just a proxy for people’s activity.

    1. fredwilson

      good point about techmeme. hacker news is like digg or reddit. people post links to that site. so it is people powered, and therefore social

      1. JamesHRH

        I had the same question. If business Insider was in the top 10, would that be social?You are describing a link directory. That makes early yahoo social.Maybe it’s a different category. News?Social, it seems, should have a more personal aspect.

    2. falicon

      I think because he only had two groups…so if it’s not search (which HN and Techmeme clearly aren’t) they have to be grouped as social…

  10. jstylman

    There’s no question social has emerged as a primary traffic source, as search has waned a bit.  Just a few years ago, 90% of all traffic started at a search engine – more often than not, Google.  I do believe we’re still in a transition, however, as this all evolves.  A few points to consider:- Evergreen vs. Time sensitive content: it’s not surprising to me that your site leans heavily towards social, given that you publish every day, usually with timely information. People are incentivized to share because it’s “breaking” and is part of a dialog taking place on and off avc.com.  I imagine reference sites that are comprised of content that’s a bit more static are still as reliant on search as ever.  In other words, the trend towards social media being THE key traffic source may not be a full on migration for all.- Search –> Social –> Destination: for virtually any query, the search results are stacked w social media sites (Facebook, Twitter, blogs, etc).  Combine this with the fact that an overwhelming majority of users still use search engines as a navigational tool – a trend that has become even more pronounced with the ubiquity of toolbars and browser plugins. The net result is that social media sites are driving more direct traffic, but search is still nestled into a user’s workflow, though it’s not a given that they’re the “last click” any longer.  This may just be a moment in time, but it’s worth noting.- Personalized Search: search engines are evolving from a one size fits all to a far more personalized experience.  While some think this is wonderful (some users, advertisers) and others are afraid of the consequences (Filter Bubble, etc), an obvious takeaway is that the emergence of the read/write web has spawned an entirely new, more personal way to discover the internet that is somewhat of a hybrid of search + social.  If anyone doesn’t think this is happening, just look at Google Search Plus Your World and think about the experience there.  That product is also telling about where Google thinks this is headed. We’ll see how this all plays out as users get more sophisticated and the social platforms continue to become a meaningful part of people’s every day lives.  Very interesting consequences, regardless.

    1. fredwilson

      thanks for sharing your thoughts on this josh. i’d love to sit down with you and talk at length about this stuff. i’ve been thinking a lot about it.

      1. jstylman

        For sure, give a shout any time.

    2. Elie Seidman

      very much agree with your drawing the distinction between topical/news and evergreen. Very different behaviors. 

    3. ShanaC

      personalized search has all the same drama as ad targeting.  Both need to get a lot lot better, and rules have to be made for exceptions so we don’t live any further into an empire like world (bread and circuses anyone?) (unless we want a real empire)And this is someone who likes personalization.  I’m not exactly like my friends, though I have certain similarities to most of them.

    4. sigmaalgebra

      Personalization can make search much more powerful because (1) it lets the search engine have more data it can use when doing the search and (2) that data can be especially relevant.  In particular, the content on the Internet is now so large and varied that search is struggling due to a lack of relevant data:  Keywords and key phrases in text with results presented in some order from a heuristic ranking algorithm work well for less than half the content based on text and still less well for video clips, still images, music recordings, and more.  Broadly one weakness of keywords is that they make hash out of meaning; getting closer to meaning, i.e., doing better getting at or handling meaning, could help.  The usual computer science approaches to meaning are based on natural language semantics, but that needs content based on text (or with good text metadata), doesn’t work very well even there yet, and is useless otherwise.Personalized search and ad targeting are related but not nearly the same thing.  Data from the social Web can be helpful for ad targeting but is from not very useful down to harmful for search.  Generally, for personalized search need to be careful about what data is used for the personalization.  With some but not all such data, there can be some serious issues about privacy.Net, search has a long way to go.

    5. laurie kalmanson

      news and information are perfect examples of me-media vs. finding out the unexpectedif all i read is the things that i have already tagged as things i want to read, i will only be served things within those tracksbut the delight of so much evergreen content is exactly the unexpectedness of itlongform journalism used to do this very well. until i read a 20-page story by john mcphee about the mississippi river, i didn’t know the history of the atchafalaya or how to spell it (i still don’t know how to say it). and that story is as good today as it was then.http://www.newyorker.com/ar

    6. Seenator

      Josh & Fred,I have a “macro” question:If overall search volume is increasing: http://searchengineland.com…AND if Social is growing at a faster rate than search, then is there a broad category of “referrers” whose absolute volume is declining?

      1. Alex Murphy

        While there are certainly some referrers on the decline, the size of the overall pie is increasing.

  11. Paul Edelman

    In our marketplace, TeachersPayTeachers, Google organic search used to be the top referrer, until their May Day algorithm changes in 2010 penalized long tail user-generated ecommerce sites (like ours) for God knows what reason.Now, direct is number one, Google organic is second but more than half of those searches are for our website name so those can be considered direct. Coming on like a tsunami over the past two months to take third place is Pinterest which now drives more traffic to us than Facebook, Bing and Yahoo combined!Equal to all of the above, when taken together, is traffic referred by the individual blogs of the sellers in our marketplace which is great to see!

    1. fredwilson

      i gave a talk last night at social media week about social. today the big four are facebook, youtube, twitter, and tumblr. but coming on strong are foursquare, instagram, pinterest, and possibly path. social is fragmenting, not consolidating. that was the main point of my talk.

      1. William Mougayar

        I’m still looking for a link replay to that talk yesterday. 

        1. fredwilson

          http://smwknd.posterous.com… you are going to have to watch some other stuff or fast forward to get to it it’s not microchunked enough to post on AVCat least that is my take

          1. William Mougayar

            Thanks. I will go through it…Your most quoted remark on Twitter was “Twitter is becoming the newspaper of the future”. The medium is the message & the message is the medium. Mark McLuhan is rolling in his grave :)- posted via http://engag.io

          2. jason wright

            Err, Albert, so is Marshall 😉

      2. Paul Edelman

        social fragmentation. i wish picasso were around to paint that.but back to your original point, it’s certainly true in our niche that social has quickly overtaken search when it comes to teachers seeking teaching resources and materials. google is pretty terrible at it, so teachers have had tap in to their networks. i’m sure your investment in edmodo will help them to enable more of this as well.

        1. jason wright

          He did – Guernica.

      3. abhic

        Key takeaway here folks – “Social is fragmenting, not consolidating.” Couldn’t agree more!

    2. sigmaalgebra

      “for God knows what reason.”No, I’m not God!  The reason is that their ranking techniques are merely intuitive heuristics, like throwing this and that into a diet without actually adding up nutritional content.  Net, they have no very clear concept or quantitative measures of what are the best search results.  That is, they are guessing.  Some of their guesses work better for this than that, and then they make some changes.”Long tail” is a major area of failure for search based on keywords and key phrases because what is crucial about an instance of long tail content is the particular meaning, and keywords and key phrases make hash out of meaning.  E.g., in principle an unabridged English dictionary matches any English keyword search!  In more detail, given any keyword/phrase search, typically the usual ranking algorithms will present the more generally popular results instead of the desired, focused long tail results.  I.e., much of the point of long tail content is that it is not generally popular.  E.g., the usual ranking algorithms applied to music would always want to report the songs on the Top 40 or the ones that went platinum and have a super tough time getting around to, say,     http://www.youtube.com/watc…That music is just what it was like the first day in the seventh grade when I saw Terry, Ann, Sandra, and Donna!  Later when I saw Gogi and Ginger!  I didn’t try a keyword search based on those names!A closer example is the keyword search     “home decorating”which can give     “About 17,500,000 results”or     “home decorating” blogswhich can give     “About 2,540,000 results”There the results are essentially the most popular instead of conforming to the artistic tastes (meaning) of Mary, Sandra, or Virgina who, of course, can’t describe their taste in keywords anyway!”more than half of those searches are for our website name so those can be considered direct.”Yes:  That’s an example of where keyword search works well and an illustration of where and why and how often it doesn’t.”Pinterest which now drives more traffic to us than Facebook, Bing and Yahoo combined!”Yes, makes good sense, but it’s already better than I would have predicted for just Pinterest so soon:  Rip off the P and get just interest, and now are starting to get somewhere!Yes, people mentioned the interest graph and then dropped it!Yes, the Internet started with, say, at Yahoo, large interest taxonomic hierarchies, but keyword search replaced those.  Still, interests are important! 

  12. Josh Gordon

    I learned to take sources of traffic with a grain of salt any time $ is even tangentially involved.Two examples: was impressed with our incoming traffic from Citysearch (where I was a paying PPC client), until I noticed I was getting click spikes (99 one day, 1 click the next etc) and their traffic was about 100 times more than would have been expected — a classic case of click fraud (they blame “bots” that are out of their control, but have no incentives to stop them). Another case is google’s adwords: advertisers have two options for their paid keywords: limit to organic google searches or allow advertising on google’s “network.” Google encourages users to advertise on the netwrok and let their algorithm determine what the network is, and many beginners do that. And they are impressed with their traffic numbers. Until they realize that the many of the network sites are completely irrelevant to their paid PPC keywords (more advanced users either disable the network or create a separate campaign for it). Google is incentivized to get advertisers to pay for these irrelevant clicks, so they keep the definition of their network vague. Point being, sources of traffic are many times real. But sometimes they are not, at least from my experience, and I keep reminding myself to question where they are coming from and why. (Though Fred’s traffic sources seem real to me!) 

    1. LE

      “Google is incentivized to get advertisers to pay for these irrelevant clicks, so they keep the definition of their network vague”Right.They also have something called “google custom search”. This is a search box that can appear anywhere on any page. And the person putting up the search box can specify the key words if they want. And then a search page appears with ads.Attached is a screen grab of what happens on one of our sites where the search string is “Josh Gordon” as an example of clicks that would not be relevant (if clicked on). 

  13. Alan Minor

    I believe I stumbled upon this blog by clicking on a link in a TNW article. (It might have been Mashable though.)Oddly enough, I heard Fred speak a few weeks prior to that at Northside Festival’s “Ideas Festival.” It wasn’t until I looked at the “About” page on this blog that I realized this.On a related music note, the same sort of revelation happened when a friend sent me a link of Real Estate’s “It’s Real.” It was only after I clicked on the link that I realized I had seen the band perform twice this past summer.

  14. Elie Seidman

    I think that it depends a lot on the category. For topical news type stuff or things – like AVC – that you reengage with every day, the behaviors are different from something like travel that you do infrequently. But I think there is something bigger and more interesting here – specifically, the role of search as the internet evolves and matures. As the Internet matures the role of search will decrease- particularly in high value verticals. I’d be curious the percentage ofsearches on Google for books that include the term Barnes and Noble or Amazon.Further, even when those brands are not included in the search term, I’mguessing a huge percentage of the click throughs go to those two sites. Googleis not making decisions for people in that category. It has to show thosebrands in organic and the paid links probably sell cheaply and perform poorly.Now multiply that behavior by every monetizable category on the web that has well known brands. This is not social per se but it’s another way in which therole of search is evolving. More broadly, this is also a commentary on Google’s power. Their power – and how they’re using their own search results to favor Google properties – has gotten a lot of attention of late. But what I think is often missed in those discussions is that Google’s search business is under assault – not from other search engines or even from social but from the change in consumer behavior. At the mall, you don’t use a search engine – the mall is dominated by a set of well known brands. The internet will have a longer tail than the mall but I expect that, particularly in high value verticals, the internet will look a lot more like the mall than it did the internet of 2002 (where search was absolutely essential). For a lot of searches, Google is actually more navigational than it is decision making. You use Google to find something on Amazon but at no point in that process are you really going to click on a paid link – you know you’re going to Amazon even if you use Google to get you there. 

    1. ShanaC

      Malls ended up closing because of the internet.  I wonder what will happen to the top of the internet tail when it comes to purchases.

      1. Elie Seidman

        Is there proof that’s true? Anecdotally, I don’t know of a single mall that I went to 10 years ago (granted, all in middle class to wealthy areas) that has closed because of the internet.

        1. LE

          “Is there proof that’s true? Anecdotally”Definitely losing critical mass. The malls used to be quite the ball busters. It was tough to get a lease in a mall. A older mentor-type friend had about 25 stores in malls years ago. All he did was complain about them and how tough they were to deal with in a dictatorial way. He had to open up a store to compete with himself or the mall would allow in another competitor (so he would have a different name selling similar merchandise to his existing store). He was able to sell the chain just for the leases he held. That was years ago. Now you can tell about the decline just by a) the vacancies (covered storefronts) b) the types of stores they allow in that would have never been in a mall 20 years ago. Same thing with many strip shopping centers.

          1. kidmercury

            the primary cause of any decline in physical malls is not the internet but rather great depression 2.0. the internet doesn’t replace physical stores, it simply demands they use the internet as a new component to their game (i.e. augmented reality, check-ins and game mechanics, real-time social networking, etc)

        2. Dave Pinsen

          Same here. I’ve even seen one revived in recent years (albeit, after hundreds of millions of dollars of renovations and the addition of a Whole Foods as an anchor tenant).

    2. Zack Kaplan

      Think this is excellent perspective…

  15. Casey Winters

    You have to consider the type of business you have. When it’s frequently updated content e.g. blog, news site, social is going to win hands down because what it does well is update real-time. For sites that offer services, ecommerce, or static content, search still dominates traffic. Also, you have to consider new vs. old sites. The organic search space has been optimized over the last ten years by big sites for all of the top queries. So, new sites that come into a VC’s portfolio are going to have a hard time breaking in by developing organic links and buildiing long tail content. So, for their marketing initiatives, they’ve rightly gone in the direction of emerging media to attempt to drive traffic, and as huge as social is, the big brands haven’t blocked out all of the ways to drive major amounts of traffic.

    1. pixiedust8

      I think this is an excellent summation. It does depend on the type of business, and also the quality of their social media content (and their plan overall). Many retailers do not have more than a token presence in social media (or they do too hard of a “sell” in social media). 

  16. Tom Labus

    Evgeny Morozov reviews “Social networks and the Death of Privacy” in this week’s NY Times book review section.http://www.nytimes.com/2012… 

  17. William Mougayar

    You asked for other studies/data. Here’s a trio from emarketer: UK Social Network Traffic Overtakes Search Engine Visitshttp://www.emarketer.com/bl… SMBs Turn to Social Before Searchhttp://www.emarketer.com/(S… Search and Social Together Aid Online Shoppershttp://www.emarketer.com/(S…

    1. fredwilson

      this is great. thanks Albert.

      1. William Mougayar

        My name is William :)- posted via http://engag.io

        1. ShanaC

          long week maybe?

          1. William Mougayar

            Yoga time for Fred…then he’ll be fine. – posted via http://engag.io

          2. Mark Essel

            Too late, your new nick is Albert.

          3. William Mougayar

            It’s not even close to my middle name which is Antoine. – posted via http://engag.io

          4. fredwilson

            yes

        2. Rohan

          Hello William, pleased to meet you .. haha

          1. William Mougayar

            Is this Silly Saturday? – posted via http://engag.io

          2. Rohan

            Saturday is what we make of it….. 😀

        3. fredwilson

          hmm. senior moment.

    2. ShanaC

      thank you!

  18. leigh

    Big media clients are all saying the same thing but with Facebook being the biggest traffic driver in social…. (vs. Twitter)We are starting to look at Facebook optimization similar to the way we used to look at search engine optimization ……

  19. Alon Maimoni

    How would you explain the higher quality in search traffic vs. social traffic from the data you showed ?Cheers,Alon

    1. fredwilson

      how do you measure quality?bounce rate ins’t a great measure on blogs like AVC because i put seven whole posts on my front pageyou can spend a ton of time on AVC without leaving the “landing page”

      1. Alon Maimoni

        Well, seems like it’s not just the bounce rate but also the time site & pageviews.I work for an online agency, we tend to see the same shift towards the social space, however the attention span in usually shorter from those channels … 

        1. fredwilson

          i wonder if quality is the right word to usei understand your point though

      2. allon bloch

        Fred: one way to look at quality is to look at e-commerce transactions.  From my experience search is still far superior to social when it comes to conversion to sales.  So e-commerce sites are still much more reliant on google.  I’m willing to bet Amazon spends much more advertising on Google than on Facebook

  20. abhic

    This is the most important topic for any consumer web startup. As much as 25% of my daily work is committed to experimenting with different sources of traffic. I have attached the 4Q2011 sources summary for an experiment I ran in stealth (not in valleyesque ‘stealth mode’, but needed anonymous and organic traffic data to validate hypothesis) : 1. 91k uniques2. Referral – 97.7%3. Direct – 1.88%4. Search – 0.34% !Bear in mind, this was from a cold start (0 visitors, fresh URL). The primary interactions in the product are crafted to be ‘viral’ (apologies again, but need a better word), so I was expecting about 85%+ referral traffic, but at ~98% even I was surprised. 2 Takeaways1. All “Social” isn’t the same – Academic discussions require us to lump stuff together. Product creators need to be able to differentiate between different experiences vis a vis how to participate in the various networks2. Don’t ignore Tumblr – This just doesn’t get said enough. Whether you like what it’s audience does or not, whether you are sick of their ‘downtime’ or not, ignore Tumblr at your own peril.

    1. fredwilson

      yeah, last night at social media week, i listed the big four as facebook, youtube, twitter, tumblr. you have to be on them.the up and comers are foursquare, instagram, path, and pinterest

      1. abhic

        Agree on 4SQ + Insta. If you consider Path vs. Pinterest (which isn’t fair in any case), then you see that Pinterest is ‘old school’ like that. The ‘Pin It’ buttons, the aggregate on a destination site bit, only public boards etc all chase the long-tail. This ‘works’ and the rest of the web is starting to get it. Path isn’t there yet. 2.0 made it easier for us to grasp the product – Our Digital Journal. Tons of activity inside. Very high engagement. But I doubt that Dave & team will chase the traditional hooks. I am very curious about what they do to compete for our API attention.

        1. Cynthia Schames

          I’m not sure how (or whether) Path could be considered a channel per se, simply because it’s intentionally so tightly curated.  It’s meant “for the ones you love”, not for a wider audience. It’s gorgeous, but it’s really not a community building or engagement medium at all.  Which is okay.

          1. abhic

            True!

      2. jason wright

        Then I’m completely f%@*#d,…or maybe not.

    2. Alon Maimoni

      Nice experiment, however, this doesn’t mean that in the long run this will stay the same. Getting traffic from organic search requires a lot of SEO work, for how long did this test run ?

      1. abhic

        Hey Alon,Thanks for commenting. Yes, I agree absolutely about the SEO legwork. And of course you would know 🙂 (thank you linkedin).So, the idea was not to bias it against search. The idea was what if we designed a dead simple application that served just 1 purpose and was inherently viral, would it get organic adoption? We realized that it would only if it could piggy back off of a top 10 platform. Tumblr was our drug of choice and it worked.The test continues to run and the data keeps growing. But our intent was to only look at 90 days worth of data and since then we have been busy with analyzing that data. Now are about to launch the 3rd experiment soon. Might be more public about this one.Btw, I can’t wait to visit Tel Aviv. Its unreal – the pace of innovation happening there.

  21. ShanaC

    Point 1) I’m a google analytics dork, and love that you post your statistics.  One day I would love to dive deep and talk about who comes using the joy of filters! </dorkiness>2) Back to search vs social.  If social is growing as rapidly as you posit as a referer, then there are serious problems when it comes to piracy.  We want people to share music so that they somehow monetize it.  If no one shares it, then how will they know XYZ Band (dibs on that name for my nonexistent band) is good? (beyond sampling it from like pandora/radio, which last I checked is causing problems of a different sort for music people)DRM has never been the answer.  I think the answer involves relying on more of the nature of social to monetize….but beyond getting everyone to pay for spotify/rdio/rhapsody/what have you, I’m not sure what that future is going to look like.

  22. perfy

    Fred,Why would you expect your site to have similar metrics as one that supplies pirated content?  I think search is still the 900 pound gorilla when it comes to those types of sites and you may have stopped that conversation prematurely.

    1. fredwilson

      this blog post suggests otherwisehttp://www.techdirt.com/art…

      1. perfy

        Fred,That article does little to change my opinion.  Like the other commenters were clever enough to state, the pirate bay is effectively a celebrity web site.  They are highly publicized and covered frequently in the media.  The analogy of a celebrity fits them well.  Hence, they over index towards social sources.Let me rather present two examples. These are the real piracy sources.  Interestingly enough, most are hosted on blogspot.I searched google for “black metal blog” and the top hit is this:http://blackmetalownsyou.bl…  (feel free to delete it after you look at this.)  Now I would consider that search term fairly benign for someone who is just interested in that genre of music.  The site is nothing more than blog posts with the title and name of a band, a brief description, album artwork and a link to download it (pirate) from media fire.  Perhaps media fire will be the next mega upload.Second example is just like the first, just a different music genre:  Searching for “punk rock blogspot”  not only the top 1, but about the top 10 are all sites like this.http://punknotprofit.blogsp… (also feel free to delete)I think torrents and pirate bay are not the problem with piracy.  Torrents are already yesterday’s methodology.  The media caught on to the buzzword, and now articles like the one you linked get written and makes things murkier than clearer.

  23. Mark Essel

    I think each community case is separate, there is no such thing as “normal”. Web browsing is not a tv show I’m tuning into with 20 million other folks, it’s a medium I share with a billion other humans, all seeking information relevant to their interests and needs. There may be super clusters with like distributions of visitors, that’s something that a tool like Hunch or other datamining methods could help reveal.

  24. Douglas Crets

    These are not links to broader studies, but they are pretty good articles, written by me, about how the search game is going to have to change to finding people and finding experts among those people you don’t know. I believe the most interesting move, for me, will be when search goes away from the searching for a platform where one person writes content, to just finding a person, wherever they are, so that you can ask them a question or get an opportunity started.I think education, as a field of operation, has to go this way. Link kids up to the world of doing, rather than the world of studying.Hope you enjoy the articles below. http://www.fastcompany.com/… And another involving our good friend William Mougayar. http://www.fastcompany.com/… 

    1. fredwilson

      thanks, i will give them a read

  25. kidmercury

    i’m not sure if the dichtomy of social vs search is fair. personally i get most of my traffic from google with youtube as #2. #googlesharecropperanyway, the true measure is who can spawn an economy. and google wins that game by a wide margin. facebook has spawned zynga, maybe some others but not much. google has spawned android, youtube, and the whole SEO industry. social disrupts search, but that will be clearest when a player with social roots launches their own search engine that takes the throne from google. right now, i don’t see any of the players being able to do that. jdawg had the right idea with mahalo, although i think you need to start niche and connect the niches rather than start with the big thing. everything is getting fragmented which is why it’s all about the niche game. tough to get dollars directed properly here though because who wants to invest in a small pie? soon enough, though, that will be the only pie in town.  

    1. kidmercury

      also, regarding piracy, i bet the most common strategy is for people to 1. find their little pirate community that obeys the density principle and concentrates lots of piracy in one spot2. search on google, click through to some other site(s), eventually they get directed to the torrent or whatever they’re looking for. 

  26. Jed Sundwall

    Facebook sent more traffic to Perez Hilton than Google all the way back in 2009. Makes sense.

  27. RichardF

    i’d like to see broad statistics because I think your blog is atypical in terms of recurring visitors and your average reader as opposed to main stream internet users.

    1. fredwilson

      what would you like to see?

  28. Peter G

    One problem with sources reported as search is that user behaviour has changed so that many (if not most) people search for a domain instead of typing it directly, so I would assume a reasonable amount of search traffic should actually be considered direct. This is especially the case with the popularity of the Chrome browser which uses google search to autocomplete a url and means that even when someone is typing a direct URL they are likely to click the search result. And a lot of people open google to type in the URL instead of typing the same into the address bar..

    1. Cynthia Schames

      @accme:disqus  I was thinking about the same thing.  It would be interesting to see a study contrasting user behaviors on longer domain names versus something to elegant and spare as “avc.com”. I wonder if the results Fred shared are at all skewed because it’s just as easy to type in avc.com every time as it would be to search?  In my case, I don’t even use a bookmark for this site, since it’s easier for me to type three letters than it is to dig through bookmarks.

    2. fredwilson

      top four search queries are avc, fred wilson, a vc, avc.com

  29. David Petersen

    The percentage of first time visitors statistic is interesting.  Stumble provides this site almost entirely first time visitors (86% have never visited before).  As expected, feedburner provides the lowest (29%).  Harder to understand is why 60% of direct visitors are new to the site when only 50% of google traffic and 40% of techmeme/ycombinator traffic is first time.StumbleUpon does a great job of introducing Fred to new people, while the other sites bring back the repeat visitors.

    1. fredwilson

      people clearing their cookies?

  30. Carl J. Mistlebauer

    I remember back in the days when we first began building our first ecommerce website, I kept asking myself, “…and how will they find us?”Through the spending of a bunch of money on experimentation, hiring companies, and hiring consultants, here is what I have learned (again, I am one of those “unreasonable people”):1.  “How will they find us?”         A.  Define who “us” is.         B.  Define who “they” is.         C.  Define what you mean by “find.”         D.  What exactly do you mean by “how.”Organic search was an easy one for us because we were an established brand name in a niche market, so of course that is going to be the number one way existing customers would find us.But, paid search was an absolute waste of money, what I now refer to as the “Black Hole of Google.”Back in the days when you had MySpace and Facebook, neither of which is worth a damn, MySpace was better than Facebook, because I could target our advertising to those who clicked the box on their profile “more to love.”  Now, while social can increase traffic to ones site, in ecommerce it is a waste of money in regards to turning traffic into sales.After spending tens of thousands of dollars on experts my nephew, who is majoring in entrepreneurship, came up with the brainy idea to search MySpace for “Bears” since we had some new printed shirts featuring bears, deer, and such.  So, I hired him for $8 hr and said knock your socks off.  Well, he brought us tons of traffic, the trouble was, neither of us knew that the term “Bears” has a sexual context to it, and now we have established a beach head in the gay market with guys who like hairy guys (didn’t sell any printed shirts but sold a lot of blank shirts! Oh, and got a whole bunch of requests wanting to know if we were “available.”)So, what do I know now?  I know that by buying a database and using direct mail I get an awesome return on my dollars.  By adding a landing page with a contest I get an awesome return on my dollars when advertising on blogs and social media sites.Through the evolution of defining “us,” “they,” “find,” and “how” I have determined that my market, my consumer demographics, just do not fit the popular ideas, or the trends of the internet. I mean lets be honest, to look at the sources of traffic to a blog, a blog that likes to refer to its self as a “Cheers” type of bar, should see social as the primary driver of traffic; social should support social.  Search is becoming the “mass media” of this generation, and thus it should be no surprise that it is the newspaper or television ad of the 21st century and will be dominated by the bigger players.So, being the “unreasonable one” I have decided based upon our history, our data, and our demographics, to create my own social media site, to give my consumers more reasons to return to our site and promote our site to their friends.  In turn I will then have the means to better determine what my consumers want and tailor products on a consumer centric basis rather than me using a crystal ball.  

    1. Douglas Crets

      I would like to know more about your experiences early on with e-commerce. And what are you struggling with now? 

      1. Carl J. Mistlebauer

        We are a manufacturer that has been selling to retailers for 20 years; we were always getting phone calls from consumers wanting to know where to buy our goods and we would always direct them to retailers in their area that carried our product.While I have always saw the internet as the perfect vehicle for niche products such as ours to grow their future the reality was when we decided to build our first ecommerce site we did so believing that B2C would only be the “icing” on the cake, not the cake its self (that would always be retailers).The reality is that even with going into it as an after thought, and even with all the mistakes we made, we still ended up having to shut down our B2C after 6 months because we could not handle the business from an inventory standpoint (our internet efforts not only created a dramatic surge in B2C sales but it also saw a dramatic increase in our B2B sales both from new retailers interested in our product and increased sales from our existing retailers).Now, our traditional customer has always been a middle aged white male who is middle to working class; we also only sold pocket tees.So, then the second go at B2C we had added tee shirts and long sleeve tees. All of the sudden we started getting sales from women, younger males, and a much broader racial demographic.At this point I realized that B2C was definitely in our future and would eventually be our future (I still was not able to accept that our retailers would not account for less than 75% of our business.)The basic problem is that a 350 lb or 500 lb person shops differently than the mass market does. Thus the trouble with dealing with consultants and companies that provide services is that they think “t shirts” and then rely on their own experiences to come up with solutions; I would give them a whole 15 page document of information about our market, our consumer, and the psychology of the big and tall consumer and I would end up being presented what I call “a mass market plain vanilla solution.”Then a couple of years ago I found out that we were selling more big and tall tee shirts than JC Penney’s did in a year, that all I had to figure out was how to offer the big and tall consumer the same options that the mass market enjoys because first of all, none of the major players in big and tall can provide these options (Threadless, Custom Ink, Cafe Press, for example) due to their size and off shoring all production, and secondly, all I had to do was figure out how to connect with college and high school age big and tall kids (male and female) then I would be locking in their loyalty for years to come (JC Penney is attempting this with their new big and tall retail stores – but again, its brick and mortar and not consumer centric.)Then again, over the last 6 months I have visited over 15 college campuses in states with a high percentage of obese population and I realize that I need to really focus on women; that’s a whole other world for me.Right now my real struggle is with the fact that while I have a vision of what I want and where I need to head and I have coders plugging away attempting to turn my vision into reality its obvious that there is a person missing between me and them; I just cannot seem to use the right terminology or something but it sure feels that we are speaking two different languages….

        1. Douglas Crets

          This is brilliant. Consumers hide what they really want from retailers, but consultants often only rely on their own experience, rather than the consumers’ demands, feelings, experiences. Have you heard of The Re-Wired Group? I’d totally love to take some of your comments here and put them on their blog. I am a writer / blogger for them. http://www.therewiredgroup…. 

          1. Carl J. Mistlebauer

            Douglas, feel free to use what you need!

  31. Youssef Rahoui

    I recently attended an e-marketing conference where the speaker was always refering to Google as a criteria to optimize a website. It was clear that he was still living in the pre-Facebook era. Today, if you target the consumer market, Open Graph optimization is as important as search engine optimization.

  32. FAKE GRIMLOCK

    GOOGLE WAS ALWAYS DUCT TAPE.NET FINALLY GET AROUND TO NOT NEED ANY MORE.

    1. jason wright

      It had great product/ market fit. The market has moved, and it will move again, and again, and again….which is great news for entrepreneurs and their investors.

  33. jfccohen

    Pew did a study and found that only 1% of consumers who are interested in local, small businesses get their information through social: http://searchengineland.com…I think that we are just about the hit an inflection point of that number changing dramatically.  More and more consumers are connecting with businesses online as there becomes a reason to do so (offers, specials, exclusive access, etc.).

  34. laurie kalmanson

    awesome. here’s a post i just did on the foolishness of music rights/geographic restrictions; i tried to be a paying customer, and first tunes said i had to go to the ireland store to buy the dylan song i wanted — when i got there, ituned said, nyah nyah nyah, us shopppers can’t buy from the ireland store.if you’re not going to let me buy something, at least make it simplehttp://lauriekalmanson.blog…i’ll just have to listen to it for free …

    1. fredwilson

      I explained that in the meeting on friday afternoon

  35. Jared McKiernan

    Yeah, but how did people find your blog for the first time? That’s really what matters, since they then can subscribe to the feed, follow you on twitter, or bookmark the blog.

    1. fredwilson

      great question. 

      1. Douglas Crets

        For me it was someone talking about it in person. Which brings me back to what search does not do well. It only helps you find things you already know a little bit about.But conversations are so meaningful because they produce ideas and people’s names that you never knew existed. When will search do that in a more specific way? It’s my obsession. 

  36. Adam Feldman

    I’d like to point out that my greatest source of referral traffic for my personal website is Twitter, followed by blogs where I’ve left comments via Disqus. In particular, I get a lot of traffic from ones with very active communities and discussions. It’d be cool to see Disqus offer some kind of analytics suite for analyzing the interaction with your posts, including clicks through to sites.

    1. fredwilson

      that is most definitely coming Adam

  37. Luke Hristou

    I agree that it’s all in the nature in which you’re searching.Twiiter, tumblr, facebook, techmeme, hackernews, they have all allowed serendipitous searching, which dominates probably 90% of my search behavior, because rarely do I ever search with something exactly in mind that I want to find. Another thing that factors into it is timeliness. A search engine is far from the first place anybody starts their browsing experience anymore. People are spending their time on the sites  I just mentioned, so even if you had something specific in mind to look up later, odds are you will see a link to it on one of these site anyway and then, “oh yea” – you check it out.   

  38. Aaron Klein

    Sorry couldn’t resist 😉

  39. Tim O'Neil

    There are two very different modes in which I use the net… When looking for something SPECIFIC: – I go to Google or directly to a commerce hub like AmazonWhen looking for something INTERESTING:- I go through a short list of sites that I know to contain interesting things…  which is typically a few Social sites that surface interesting things – through authoring, curation or purely social means.Search has always been terrible for surfacing interesting, new content because it is fundamentally designed for something else – indexing content based on its face value meaning.  Social on the other hand fundamentally sucks for finding anything specific, but is fantastic at surfacing random ‘things of interest’.There is a natural overlap, but it is small in proportion to the full set of content published across the net.  The upswing of social to me does not really signify any kind of devaluation of Search, so much as a more appropriate tool taking the place of a poor one for an increasingly common use case.  

  40. jonathanmendez

    Recently saw a large site in the teenage demographic that had the majority of referrals from the #2 Search Engine in the World – YouTube.  It’s also worth noting that mobile Search is growing at the same pace that desktop search did  in the early days.The important point in my mind as a marketer (or if I were a publisher) is not one vs. another but the interplay between them that Josh alluded to. For example, how can/does higher funnel awareness through social increase performance lower down in the funnel through Search (both organic and paid)? On the other side, how can branded queries loop back into social and amplify brand awareness?We are in a cross-channel world now with new channels (tablets anyone?) and sources within channels (Pinterest anyone?) growing at a dizzying pace that far outstrips our ability as marketers and publishers to understand the value they create. That tilts the value scale in favor of the channels and sources. However vibrant media markets need two sides to succeed as we’ve witnessed with the rise of paid search (and television). The ad dollars formula = path of least resistance divided by results + understanding is not one that plays well in this dynamic. The rise of social is great – the people have spoken – but the digital world can not survive on earned media alone.

    1. fredwilson

      yesssssss

  41. rahuliffic

    Twitter and facebook referrals are from people following/subscribing to avc. So in a way they could have a big chunk of direct visitors too. I either click through to avc through email or twitter where I follow you.

  42. Jz

    This explains why Google is in deep trouble. Just look at its share price in last 5 years, stuck in the mud and seems unable to break out….

  43. Cassieboorn

    This is a really interesting question. I am not surprised that your biggest traffic comes from social because you are someone with a great reputation who’s posts get shared often. However sites like She knows and About rely mostly on search and that is their focus. I think it depends on the type of quality you are creating but you have to choose search or social. Posts that are highly searchable are not usually interesting enough to be shared on social media. Unless you are Penelope Trunk. 

  44. Brad

    You should invest in http://www.avc.com. Seems like it creates great traffic.

    1. fredwilson

      i do. every day.

      1. John Revay

        If you and the GothamGal every retire or slow down – you should assemble and publish a “Best of AVC and the GothamGal” book – On line and open a course. :)You can raise some money for a favorite charity/cause.

  45. Muskie

    I still get most of my referrals from search, specifically Google. Sometimes something will get tweeted or liked, but generally I’m way out in the long tail and get 100s of different keyword referrals that I’ll get one of on any given day. A few pages or posts are consistently popular and will get the majority of the page views, different stats packages measure and classify data slightly differently, most everyone uses Google Analytics so this argument is less relevant than it was in the past.What social buttons and social networks you personally use may also be a factor. Stumple Upon is not a website I use and there are no Stumble Upon buttons on my blog. Twitter, Facebook and sometimes LinkedIn are the social networks that send me the most traffic. I also get traffic from obscure often foreign language forums that will discover something I posted online. Google Translate sends me a fair amount of traffic.

    1. fredwilson

      there are no stumbleupon buttons anywhere on AVC and yet it is #2 after directi wrote this a while back and i’m still curious why i get so much traffic from SUhttp://www.avc.com/a_vc/201…

      1. Muskie

        It probably is an overlap of niche or specialty interest. Pinterest is now one of the top ten social networks in North America. This came as shock to a lot of geeky guys when it made the rounds online. It didn’t suprise me though my sister, my cousins, and their friends are always talking about it and the stuff they see on Pinterest.I’ll wager that Pinterest doesn’t send AVC.com a lot of traffic. Niche social networks may become more important as they are more directly tied to passion. Very few of my friends and family have joined Google+, I especially think grandmothers and babyboomers will be slow to sign up. Facebook is may be too entrenched as the friends and family social network, it’s where you go to see pictures of your grandkids.Google+ unlike Pinterest doesn’t seem to have a natural demographic of people who just have to sign up. They went hard after Twitter’s audience of technology keeners, early adopters, and media junkies, but for the most part Twitter seems to have retained that audience. Every newscast even sports lists people’s Twitter handles. I don’t see the G+ logo in TV or print ads whereas you regularly see calls to action aimed at gaining more Facebook Fans and Likes.Social may be generating a lot more traffic, but it can be fickle, whereas if you do well for certain keywords in Google and other search engines you will regularly see visits to certain older content. This is especially apparent in less mainstream topics such as hobbies.I’ve given this a lot of thought and written down many of my observations and insights regarding online popularity and influence.http://blog.muschamp.ca/201

        1. UPM

          “Social may be generating a lot more traffic, but it can be fickle, whereas if you do well for certain keywords in Google and other search engines you will regularly see visits to certain older content. This is especially apparent in less mainstream topics such as hobbies.”Like your quote, Muskie. Here’s a non-tech’s view of Search vs. Social.Social is moving away from big farms (Facebook, Twitter, etc.), for  anumber of reasons: a) Too much noise from too many “speakers,” b) upfrontand arrogant invasion of privacy (read Facebook), c) high school-like cliquishbehavior (read Facebook), d) too broad of topics and audiences, e) crappyinterfaces that never evolve (hello again, Facebook). Mature sites likeTwitter tend to benefit entities with existing fame over those seeking toestablish a presence. From a business standpoint, these macro socialsites are generally only good for broadcasting announcements. For consumerbonding/two-way conversations, vertical social sites are more productive. Siteslike Pinterest will end up surprising businesses, as they discover the research/profilingas well as marketing benefits (which in turn, of course, will mutate/mutilatethe original thrill users have for them until they are forced to move on to other,more personable sites). Blogs are, in this writer’s opinion, very relevant, asthey provide fodder for sites like Pinterest. As for search, when isn’t a search mechanism viable?The question lies in what it looks like, how it works. Personally andprofessionally, I miss the more democratic days of search engines. I don’t carefor the current truncated set of autocratic options. Therefore, I’ll lookforward to interlopers who will offer a more altruistic vision of what searchtools should be.   In the meantime, I work with what I have,professionally and personally; I enjoy the smaller social sites that offerimmediate value and experiential reward. 

  46. Ben Elowitz

    Fred, I publish a monthly social leaderboard for media that looks at social and search traffic for the top 50 media properties online. I will be publishing the December numbers this coming week on my blog (Digital Quarters) and will send you a copy. 

    1. Douglas Crets

      I’d like a link to this too, please.

    2. fredwilson

      that would be fantastic Ben. i’d love to see trended data on that

    3. William Mougayar

      Will it be on your blog, Ben?

    4. John Revay

      Hi Ben,I enjoyed reading your Friday Blog post – Google Search Plus Your World: An Offer You Can’t RefuseI will plan to starting following your blog – thanks 

  47. John Revay

    can any one comment on the tools/app they use to track blog posting re: notice of new blog posting by author.I use William’s http://www.engag.io to track comments and discussions.

  48. LukeG

    Some significant-scale but anecdotal ecommerce data: looking at several months and tens of millions of visitors, “social” (almost entirely Facebook in our case) traffic converted (completed a purchase) at a rate more 15% worse than traffic from search engines.I suspect other ecommerce sites see directionally-similar patterns. That said, it’s a trap to think of social traffic as “lower quality.” Yes, intent isn’t there as often. Social visitors are different rather than worse. The fact that they’re coming from social sources is itself a valuable piece of data that should inform the a visitor’s experience on your site.SEO optimization has been a part of the game for 15 years, right? We actually build sites differently to capture more search traffic. Social optimization matters now, too, and will matter *more* soon enough. And I’m not talking about experimenting with 8 different versions of your Share and Like buttons.

  49. Richard

    If a data set becomes available with a good group of covariates, I’d could put together an analysis. Feel free to contact me.

  50. pixiedust8

    I think it depends on your business and audience. I worked for a cosmetics company, and most of our traffic was direct, then search, with social lagging behind. But cosmetics appeal to a very broad population, some of whom are not savvy about social media. Your audience would be very different.

  51. That Marc Fellow @ StumbleUpon

    Thanks for sharing this data, Fred.  WRT the referral traffic you’re seeing from StumbleUpon:  To turn the cliche dating comment around, “It’s not us, it’s you.” 🙂  We’ve been iterating on our recommendation algorithms faster than ever, with the goal of surfacing great content (both timeless and trending) to the appropriate audiences (we also recently introduced Channels — of which you’re one (http://www.stumbleupon.com/… — that allow StumbleUpon members to self-select specific sources from which they’d like to see more relevant content).  Turns out you consistently produce great content for individuals interested in areas like Entrepreneurship, Investing, Technology and Business, all of which are popular on StumbleUpon.  We’re particularly excited to see how many first-time readers we’re introducing to your blog, as this is the raison d’etre for StumbleUpon:  To help content find its audience as much as the reverse.  So thanks again for contributing to the content ecosystem. 🙂

    1. That Marc Fellow @ StumbleUpon

      Sorry – looks like the close-paren was appended to the Channel url; should be http://www.stumbleupon.com/… 

  52. Janet Aronica

    Hey Fred! Inspired by your post, at Shareaholic we did a larger study on this for our 200k publishers. Thought you may be interested in the findings – http://blog.shareaholic.com…Our breakdown of the top 10 traffic sources:Search – 64%Social – 11%Direct – 24%Other – 1%Interesting stuff! As a content marketer, my takeaway is that while search is the biggest source of traffic, social is the glue that pulls everything together as it builds the community that contributes your regular direct traffic and increases your rankings to improve organic traffic.Hope you enjoy our study 🙂 

    1. fredwilson

      thanks!

  53. Dan Whaite

    Just want to pose the question…why is it framed as search vs. social?  First, I’ll attempt my definition of both.  To me, search is a filtering function, a utility.  Search takes data and filters/sorts based on some criteria.  The point is that it’s a function.  Social (at least the new part that is most relevant here) is the consumption of user-generated data, specifically by people that you know (a status, a like of a page, a log-entry on a server indicating that a friend read an article, a tweet, a blog post, etc).  The point is that you have a relationship with them, therefore you trust them.  And this influences your behavior.  I’m not considering a tweet of a celebrity to be social (I know, debatable, but I’m focusing on the new part – celebrities have had megaphones for decades). Yes, there are lots of examples in the gray area.So these are not orthogonal concepts. They just happen to have evolved separately. Social data is normally consumed by browsing, when you’re not searching for known-items.  But they will most certainly converge.  There is too much utility in the weaving of social data in search results for it not to happen.  The biggest challenge is knowing that, for a given query, when social data is more important than other content.  People often point out the filter bubble issue – but imagine saying to someone who is searching for a great sushi restaurant in midtown that they shouldn’t pay attention to the opinions of their friends because they might not be exposed to new experiences.  Won’t happen.So this question, I think, could perhaps be posed as – was my blog found through social data or non-social data.  Or, were my readers influenced by people they know to read my blog?  Of course, I have no idea what the answer is…but the more quality social data that’s generated, the more it will influence what we discover.

  54. Aaron Klein

    If this comment was about politics, Charlie, I agreed with EVERY character. 😉

  55. fredwilson

    damn.