The AVC Funnel

I did some digging on Monday as I was thinking about ten years of AVC. Here are some of the more interesting numbers I came across:

– The AVC community has donated $203,939 to DonorsChoose campaigns over the years. 1,322 folks from the AVC community have supported campaigns reaching 53,486 students.

– There were 6,232 posts on AVC in the first ten years. You will notice that is well north of 1 post per day. In the early days, AVC was like a tumblog, with short posts from me multiple times a day. That changed when Tumblr came along and I moved those kinds of posts to

– There have been 154,490 comments on AVC. 22,490 of them were on the TypePad comment system in the first four years (~5,500 per year) and 132,000 of them have been on disqus in the following six years (~22,000 per year).

– 8,916 people have left a comment on AVC via Disqus

– AVC has been visited 22,392,487 times in its first ten years according to TypePad.

– I put Google Analytics on AVC on or about Nov 10, 2005. In almost eight years that GA has been on AVC, there have been 16,965,651 visits by 9,481,697 unique visitors.

So here is the AVC funnel:

10mm people have stopped by, 10,000 leave a comment, and 1,300 contribute to DonorsChoose

That works for me. Thanks for stopping by everyone.


Comments (Archived):

  1. LIAD

    …I’d need to see Y/Y growth metrics, meet the rest of your team and get a clearer picture of your monetization plans before making a final decision, but I’m interested in joining your seed round.Where should we send the exploding term sheet….

    1. Rohan

      Love your comments, Liad. Always make me laugh! πŸ˜€

      1. fredwilson

        Me too

        1. tyronerubin

          Hence why I follow LIAD on twitter

          1. bsoist

            me too πŸ™‚

          2. Andrew Kennedy

            now following him. thanks.

      2. David Semeria

        Europeans (sometimes) do it better….

    2. William Mougayar

      I would do a convertible round. Close him quick before other VCs jump on it.

    3. jason wright

      needs to get his ip sorted out;- i see no trademark application- who owns the domain name?

    4. fredwilson

      I am doing the round

      1. William Mougayar

        I’d like to commit $250K with an AL Syndicate. That will be my first investment.

        1. Matt A. Myers

          I am William’s volunteer henchman, so you better keep up with these efforts Mr. Wilson …

        2. LIAD

          paying people like Calacanis or Betaworks a 20% carry to piggyback on their network/vetting process and investment management expertise seems a cheap price to pay.

          1. falicon

            One of those ideas that seems so logical and worthwhile in theory…but will turn out in practice not to produce the results one would expect.I think the syndicates that will crush it in this new format are not the ones that existed in some other form prior to this being an option (though those are the ones that will pull in a lot of easy money and attention).

        3. Cynthia Schames

          Are you serious?

          1. William Mougayar

            Half serious…if you only knew which half πŸ™‚

      2. Abdallah Al-Hakim

        All jokes aside (and they are very good πŸ™‚ – this made me think that the AVC community could be a terrific place to form some angel syndicates for investing via – It could serve as a community consultancy platform for those looking to do investments via

        1. Matt A. Myers

          I think that is likely the idea. Tap into existing networks, communities – facilitate them.I would like to volunteer to be the first one to benefit from this syndicate …… seriously though?

        2. kidmercury

          fredland should def be a pioneer in crowdfunding…..fred should be the curator and we invest

          1. kidmercury

            this could also lead to way more intense beefing in the comments as money enters the discussion. watching and/or instigating the beefs could be a lot of fun

          2. Kasi Viswanathan Agilandam

            money is a two-sided sword …when not done as charity … so watch what you wish for.

          3. kidmercury

            you can say that about anything. water can have toxins that can kill you. air can have poisonous fumes. land can result in earthquakes. natural food can be poisonous. it’s all two-sided and we should live in permanent fear of what we wish for!

          4. Kasi Viswanathan Agilandam

            Wish is always with fear … we never wish for what we can easily achieve or get ….when there is a fear of “May be not possible” …then we wish.btw,Water/air/food we(human) have to add all those toxins to them … but money will do it on humans and that is the difference.

          5. fredwilson

            i fear blind followers. i am nervous about the “social proof” aspect of crowdfunding.

          6. kidmercury

            i agree your concern is valid, though i think that is a feature of human psychology and as such is already present in current investment markets. from my perspective, i’d rather have people blindly follow you than someone who is stupid or a scam artist (though i agree people need to ultimately think for themselves)

        3. ShanaC

          I wish I could say I could organize, but I can’t (not qualified, too busy)But I think this would be an excellent idea

          1. Aaron Klein

            You should really blog about it on a sidewalk.

          2. andyidsinga

            Hey, the AVC syndicate could be added to your service somehow πŸ™‚

      3. Emily Merkle

        I have a proposal on – started it a year ago or so. Life happened, but I plan on working it up. Advice from community appreciated.

    5. JimHirshfield

      Just say “no” politely and move on.

    6. Guest

      I’m concerned about the conversion ratio…doesn’t seem to be scaling with the visitors. But hey if you’re in I’m in.

      1. Matt A. Myers

        P.S. Like the profile pic.. P.P.S. Still not changing mine. No one’s taken a good shot of my short-hair ‘fitness’ / summer look.. so for now, I stay incognito.

      2. ShanaC

        its being reported incorrectly.

    7. Kasi Viswanathan Agilandam

      Wait …. don’t you think 10mm in 10years is too small a number?…in this FB age ?

    8. baba12

      This is not the Hockey Stick growth that Fred would be looking to see as a place to invest. I think he would decline even without the detailed y/y growth.What would be his words for rejecting the seed round? Would he be diplomatic i.e. “Diplomacy is when you tell a person to go to hell and they look forward to the visit” or will he be more like

    9. ShanaC


    10. laurie kalmanson

      lolrelated: optimizing google for google…

  2. Mark Birch

    Awesomely impressive!

  3. awaldstein

    Community is a force to change to the world.Blown away by these stats.

    1. andyidsinga

      hey man – I meant to say hi to you on monday night ..then I lost track and time ran out! bummer πŸ™

      1. awaldstein

        Absolutely…Next time for certain Andy!

    2. laurie kalmanson

      content < = > communityawesome content: foundation for awesome community

  4. Richard Carlow

    Late to the party, but appreciate the archives as well. Plenty to think about and learn.

  5. jason wright

    “10mm people have stopped by, 10,000 leave a comment…that’s such a low percentage. it surprises me.

    1. Fernando Gutierrez

      I guess that those 10k commenters are counted many times in the 10mm.

      1. Cam MacRae

        Many, many, many times over 10 years.

      2. jason wright

        is this quote saying that over the ten years 10mm individual people have come to the website but only ten thousand of them have made a comment, or have a i things wrapped around my neck?i always need to do a double take with stats before their meaning sinks in.

        1. Matt A. Myers

          You might be suffering from self-experience bias. Not everyone likes to comment or engage – mostly I find people aren’t comfortable or don’t because they don’t feel they can add any value to the conversation (from what people have told me).You’re more alone here on AVC as a commenter than you thought – but we’re alone here together. πŸ˜‰

          1. jason wright

            maybe…..the species is naturally chatty, gossipy, but there was a time when i kept my cards close to my chest, foolishly believing that my knowledge was valuable to me and that i shouldn’t give it away ‘cheaply’ for others to exploit. i now take the entirely opposite view, the avcish view i could call (wo)man is an island πŸ™‚

          2. Matt A. Myers

            Agreed. I am working my way towards being as open, though still a bit stuck in being fully open. I need to let myself realize the gains that come from sharing, and how that sharing will help others indirectly will come back towards me. Much like how Fred has success with and using AVC.

          3. bsoist

            Today’s post, and Thursday’s before it, are clear indications of the gains to be had from sharing.

          4. jason wright

            yep, and sharing in a prudent way with more sensitive info, and building a business is a sharing process.

      3. fredwilson

        No. That 10mm is unique visitors

    2. fredwilson

      There is a lot of drive by traffic

    3. Ro Gupta

      See my other comment in this thread — I think the 10MM is inflated by 10x or more because of the limitations of analytics tools like GA.

  6. Dan Berger

    When I read this headline, I was expecting the post to talk about the deal flow the blog has brought you. E.g. x entrepreneurs touched >> y deals reviewed >> z investments.I guess I’m happy to be wrong! Well done πŸ™‚

    1. bsoist

      That’s funny because I sensed right away it would not be that and my curiosity was peaked. Didn’t know what to expect.

  7. Cam MacRae

    In the words of Keating, that’s a “beautiful set of numbers”.

  8. Dave W Baldwin

    Amazing! Add in what so many members of the AVC family do and have done helping others and causes, that’s quite an impact!

  9. William Mougayar

    And the most important stat is when you said on Monday that you wouldn’t be where you are without this blog. That outcome is priceless & no analytics can measure that.

    1. Matt A. Myers

      This is where qualitative values come in.I wonder though if the benefit Fred gained from this will dilute as more mimic this same behaviour – in a sense him being a first mover. There are many niches and cross-niches to be filled mind you.

      1. bsoist

        I understand what you mean, but what Fred does can’t easily be mimicked. It’s not easy to stick with something like this if you’re doing it *only* for the impact it has on the bottom line. And I don’t think it reaches people the same way when you do it that way.One of the things that struck me early on about the comments here was that it doesn’t seem to be about trying to get noticed by Fred. People here seem to be genuinely interested in what Fred shares. I don’t think that would happen if Fred was simply “phoning it in.”I will say that recently this place has been a little introspective, but I guess that’s to be expected after 10 years and 22M visits.

        1. Matt A. Myers

          By mimic, I don’t mean copy or clone in an inauthentic way. You definitely have to be passionate and fully invested in the topic, and have reasons for doing so – if only for the reason of direction topics and conversation, putting the bias on it that will (hopefully) provide value to the readers – or is what will dictate who those readers become.

          1. bsoist

            Understood, those that get it done should have little impact I think. Even the numbers shared today indicate that the pie is large enough for others to share.Though I have to agree with your original point. It will be interesting to see if I’m wrong about that.

          2. Matt A. Myers

            To me it hints at many people haven’t found or are looking for community to join and learn from – or perhaps they have found it, gain pieces of information (or not) from AVC – and then bring it back to their own communities.

          3. Emily Merkle

            That’s how it happens. Many participate in more than one community. Nothing wrong or aberrant about that…

          4. Emily Merkle

            You don’t add value with spin and bias. I find Fred’s content at times strictly informational, at times unapologetically opinionated. But never “biased”.

          5. Matt A. Myers

            His bias is he’s posting about tech or startups or etc.. There’s nothing inherently wrong with bias, we all have it. It’s a matter of if it’s influencing your understanding of something is when it can become an issue.

          6. ShanaC

            everybody has their biases

          7. laurie kalmanson

            that’s what you think;)

          8. ShanaC

            i’m a bayesian, that is what my posteriors seem to tell me πŸ˜€

          9. laurie kalmanson

            unless you aren’t πŸ˜‰

          10. Emily Merkle

            Bias: prejudice in favor or against one person, thing, or group compared with another, usually in a way that is considered to be unfair.Yes, Shana – everyone has their own biases. But Fred’s are not reflected in AVC. He posts about startups and tech because that is what he lives and breathes.

          11. Matt A. Myers

            In favor of tech, startups, etc.. “usually” doesn’t say “always” and says “usually” because of the negative connotation that most people have with the word bias, so they use it in language in such a way – even though inherently it is neutral.

          12. Emily Merkle

            I stand by my position. “Bias” is not the right word. Preference is.

          13. Cam MacRae

            Bias: a concentration on or interest in one particular area or subject.It’s the exact word I’d choose, and I speak the Queen’s.

          14. Emily Merkle

            Upon further digging, both definitions apply. Apologies, Matt. I sprechen sie Deutch.

          15. Donna Brewington White

            Long live.

          16. Emily Merkle

            Fred adds value to readers through education and motivating readers to think. The value Fred receives, IMHO, is through the fact that his readership is essentially an enormous focus group; he keeps a finger on the pulse.

          17. Matt A. Myers

            He learns from us, and passively promote and talk about AVC, – share our own comments at times, etc..Everyone gains exposure and everyone learns.

          18. Emily Merkle

            …..which is what I just said πŸ˜‰

      2. William Mougayar

        Anyone can copy an idea, but you can’t copy someone’s brain or experience.

        1. Matt A. Myers

          I wasn’t meaning to imply an identical copy or version of Fred. It was more than as more people start following the model, which seems to work, then people in a community who may feel more at home in a slightly different environment would have more places to interact regularly – and so people could spread out more thinly. There still be thought leaders and those ideas will still trickle through communities.

      3. JamesHRH

        I think the point is that Fred gained key understanding from being an active blogger.That experience allowed him to move with confidence when he saw opportunities like Twitter & Tumblr.As he said, early on, AVC was essentially a Tumblr account. You can see, now, how that would have made Tumblr far more compelling for Fred, than for non-blogging VCs.I would estimate that he gains 10x from the experience of blogging, than he does form the community providing specific insight.

        1. Matt A. Myers

          I fully agree. Though there are only so many large core function platforms that can or need to exist – so as his insight can evolve, specifically from blogging – that boat has somewhat sailed, but he can gain insight by using writing on AVC as a tool to a) learn (from organizing thought to reading comments on his thinking), to b) teach and build relationship, to c) gain exposure for himself and USV.It is likely the community itself who analyze and read the insights of the community – it is this trickle effect that is beneficial for the readers … and for the innovators. A lot of what I have learned, and helped direct my understand – or reaffirm my own thinking – through the comments is being put into practice and will help me launch and navigate the road map I have created.I have evolved my own theory and continue to do so, and the nuances I have are there – most of which I haven’t shared publicly, at least yet. Fred and USV could possibly benefit though, if it makes sense at some point in the future for them to invest – I would be happy, even honoured if that happened.I made a comment in reply to something Fred said the other day, might have been yesterday – and turned it into a bit longer of a blog post –

  10. Mac

    ….and one clear voice.

  11. Anne Libby

    AVC, serving like a lightweight United Way for the NY tech community. (Minus the large admin/development functions.)Thank you.

  12. WA

    Are the unique visitors measured by unique addresses they are visiting from? In other words if I visit from the office or from the college or a public wifi hotspot or even from two or three different hotels in 3 different countries and am not logged into disqus for example-do these logins reflect I am a different visitor each time? Apologies for a limited tech knowledge on such things. I ask this because it would actually make the comment ratio stronger in response to an earlier post.

    1. fredwilson

      Probably double counting some people

  13. William Mougayar

    Do you have a year to year breakout of the comments growth by total number of comments & # of commenters? That would be interesting to see. I’m not sure if Disqus Analytics provides that.

    1. fredwilson

      Not sure

    2. bsoist

      comments are timestamped, so I’m sure we can get that data

      1. William Mougayar

        Falicon will have it done in 3 API calls before the morning’s end πŸ™‚

        1. bsoist

          Was that a challenge to see if I could do it in less than three? πŸ™‚

          1. Kasi Viswanathan Agilandam

            Is there another falicon in AVC community πŸ™‚

          2. bsoist

            I don’t know about that. I seriously thought William’s comment was a way to get me to do it. :)@falicon has told me he likes the way I think. Same here – and I like the way he gets work done. From what I’ve seen he lives the hacker’s motto – shut up and show me the code!

          3. falicon

            YES! Without action, there is no evolution…only questions and opinions…Here is the breakdown according to data (ie. not nec. accurate and def. not ‘official’):total comments: 290,969unique commentors: 15,0842003: 622004: 6822005: 1,0532006: 1,0872007: 5,4872008: 19,1872009: 32,6622010: 48,1322011: 58,9892012: 79,0222013: 44,562

          4. bsoist

            Nice, thanks.Never used the Disqus API, but I did take a quick peek at the docs a few minutes ago. So far I don’t see a way to do it in three calls πŸ™‚ but I will probably take a stab at it later today.

          5. falicon

            Nice! The console they have is pretty solid for playing around with data…but there are limits to how much data you can pull out on a given call (so it’s impossible to grab these numbers in any small amount of calls).I cheated because I already have all the comments indexed…and I already had code that does/can search through all the comments…so all I had to do was write a *very* tiny bit of code to populate an associative array with the years/counts (and then execute it against a ‘search all avc’ command).As an aside, this is why my numbers are *way* higher than what Fred reported…because it includes all comments in the ‘avc’ disqus forum…he uses the same Disqus short code on the tumblr blog as well (since Disqus uses the url as a key, it works fine in this way and I’ve seen others share a disqus install across domains as well)…I didn’t bother to do the extra filter of ‘just avc’ for my quick counts…

          6. bsoist

            Yeah, it looks like separating the threads will cost me even more requests. I’m looking at it now.I met a guy from Disqus the other night. Maybe he can raise my request limit if necessary. πŸ™‚

          7. bsoist

            Tried to time it so I could get 2K consecutive calls, but I didn’t time it quite right. I was able to get 1,619 in before getting cut off. 3098 threads found, so that gives me some data to play with. πŸ™‚

          8. William Mougayar

            Interesting the run rate for 2013 will be around 60K, which is lower than 2012 & flat from 2011. Hmm, need more Occupy WS or Elections / gun control posts & they will spike in no time πŸ˜‰

          9. bsoist

            The numbers I get from Disqus are very similar. I thought the charts link I posted earlier might show a preview inline, but the numbers look like this. My 2007 is far enough off of yours to tell me that I probably did not quite get all the data before I was banished.2007 35532008 187392009 311642010 468662011 580492012 759882013 42543

          10. falicon

            Awesome stuff…you do this all via the console or your own code? If your own code, what language did you use?Regardless, you officially earn an extra +100 geek points in my book. πŸ™‚

          11. bsoist

            Python.…I set API_KEY to my key and FORUM = “avc” and DOMAIN = “”It’s a little rough, but it worked ( Well, the version I used got stuck in an infinite loop so I killed it. I think I just fixed that, but didn’t test. )

          12. William Mougayar

            Yup. Bill can hold his own in that dept.

          13. William Mougayar

            lol…do it.

        2. kidmercury

          lol #truth

          1. kidmercury

            i like hashtags as a social convention. i enjoy them greatly and plan to continue using them in the manner i have. #hashtagpride

          2. Emily Merkle

            Carry on, my wayward son…

          3. Emily Merkle

            Carry on, my wayward son…

          4. pointsnfigures


    3. ShanaC

      it doesn’t. which is a total pain when you want to id who to talk to in your own community

    4. fredwilson

      check out the comment near the top of this thread. it’s right there.

  14. Kasi Viswanathan Agilandam

    There is one stat missing :-)how many ‘upvote’ or ‘likes’ you have given (approx)… I rarely see your name on the upvotes….luckily yesterday i saw one.

    1. fredwilson

      i am starting to do that more. but i like to “like” via a reply

      1. Kasi Viswanathan Agilandam

        thanx for liking that comment of mine πŸ™‚

  15. Andrew Kennedy

    This post is really interesting to me on a couple levels. Love that you are looking at the AVC funnel (in this case) in terms of philanthropic participation rates. $203,939 / 1,322 = $154.27 avg donation. Not sure what I was expecting for avg donation, but this feels a bit smaller than I would have thought. Do you have any stats (or are you willing to share stats) on the distribution of gifts in terms of size? i.e. min/max/mean

    1. Andrew Kennedy

      typo — meant median not mean…

      1. bsoist

        you know, since you already figured out the mean πŸ™‚

        1. Andrew Kennedy

          I could go back and edit the comment, but happy to own up to my mistakes πŸ™‚

          1. Kasi Viswanathan Agilandam


    2. bsoist

      I’d like to think that indicates a “grass roots” effect – a lot of small donations with the average skewed higher by a few large ones.

  16. bsoist

    but it’s just a blog, right?

  17. mikenolan99

    What a great success… this community inspires me everyday… I was sad that I couldn’t join the party on the 23rd. Hope all had a good time!

    1. pointsnfigures

      was fun. Good to see people in the flesh.

  18. Tom Labus

    Do you think that AVC is still evolving?

    1. fredwilson

      i sure hope so

  19. Kasi Viswanathan Agilandam

    The 1322 looks bigger than the 10mm

  20. Matt A. Myers

    The 100/10/1 rule.AVC community fits it quite nicely. Well, AVC’s more niche, so fits more like 100/1/.1 which makes sense as niche means the field is narrow, so therefore leads to a more narrow point.

    1. Emily Merkle

      Forgive me, Matt, but what exactly are your saying here?

      1. Matt A. Myers

        The 100/10/1 rule is…100% of people consume.10% of people curate.1% of people create.I realize this doesn’t fit fully with the DonorsChoose aspect being included, though the same ‘algorithm’ extends well into how caring you are.There will be some – perhaps many, perhaps even the majority – of people who are just viewers in AVC, though don’t engage as much – so I am applying an assumption that those who are more engaged with AVC would also be more likely to donate to something promoted by AVC; It would have to be something they’d want to promote anyways on their own of course.Linking the stats of who donated to the engagement numbers on AVC would be really cool – though not sure if possible, and not sure if Fred would want to share – though I’d be really interested to see.

        1. Emily Merkle

          Where did you get that model from, and how does it in any way reflect how”caring” one is?

          1. Matt A. Myers

            I never said I was trying to predict how caring someone is, nor does donating money to something even come close to being analogous to caring.It’s a modification of Fred’s 90/10/1 rule -“Out of 100 people, 1% will create the content, 10% will curate the content, and the other 90% will simply consume it.”- via…Those percentages are what Fred has seen across their investments -“That plays out on this blog, that plays out in Twitter, and that plays out in most of the services we are invested in.Anyway, my modification is to 100/10/1 is that technically 100% of people consume, since curators and creators must be consuming.”And so if this rings true across ecosystems then extension is my own theory based on how the universe and things work..

  21. pointsnfigures

    On the other hand, have you invested in a deal that came exclusively from your blog? Out of the deals you see from your blog, how many do you actually diligence, and then invest in? That would be an interesting stat as well.

    1. LE

      That is one of those intangibles like advertising or even some “little things” that are hard to measure.For example JLM mentioned on this blog at one point the bathroom fixtures that he uses and that he thinks are the best. Even though I’ve forgotten the name if I am in a fixture showroom it will come back to me. And I will most likely put a higher value on those fixtures when deciding to buy than a brand that I know nothing about. How do you measure something like that? You can’t.One of the examples I like to use is this. Try to quantify the effect that Saturday Night Live, you know “Live from New York It’s Saturday Night” has had on NYC. Do you think for a second that that little blurb, around since 1974 hasn’t somehow benefited New York? How do you measure something like that?

      1. pointsnfigures

        There are people that go through the blog to get to Fred, send him a deal. As a VC, he sees hundreds of pitches every year. They invest in a few. Agree, there are positive externalities to blogging that are hard to quantify, but I bet USV and AVC have a gut feel about the deals they see that come to them because of the blogs they have. Mark Suster at Upfront says blogging has helped his deal flow.

        1. LE

          Agree. The gut feel part is interesting. But of course it doesn’t take into effect the time spent vs. if that time was spent on something else.I actually made some money this past year as a result of making comments here. A nice amount that on the surface seemed respectable. Then I decided to do the math and see how many comments I had posted and estimate the amount of total time spent commenting.Suffice to say when I dug deeper the results surprised me quite a bit vs. what my gut was.Of course I never commented to make money obviously but I could easily justify (to a third party w/o giving them the opportunity to do the math) how great I had done. And I could easily play with numbers to have whatever spin I wanted.Taking this one step further of course things haven’t even played out in terms of “the thing that leads to the thing”. Something bigger that ends up with a big payback. That started with comments.In any event this is all the case for working hard every day as hard as you can. Because you end up having to put so much stuff out there in order for some lucky event to happen. At least that is what I have found (and why I did so well with online dating – boy did I work hard at that one and boy did it pay off..)

  22. Emily Merkle

    Fred, with those metrics, it is patently obvious that you’ve been buying advertising from my firm. Let’s get that next IO signed…

  23. sachmo

    The informal gathering on Monday was great. Thanks for hosting. I finally got to meet FakeGrimlock – an interesting character that one : )

  24. Jorge M. Torres

    Inspired. A blog can be about so much more than the blogger. It’s about impact. And measuring it.

  25. Kirsten Lambertsen

    Those are some wow numbers. I wonder how many VC bloggers started blogging as a direct result of your example.

  26. Pramod Dikshith

    That is phenomenal has been a great learning experience. Keep it going Fred. Kudos to you for having sustained this writing day after day for 10 years. That is quite a feat.

  27. ShanaC

    Your conversion rate is higher. Think about it, Donor’s Choose campaigns are only up for a month. Actually conversions should only be included for that month.So even better news.

  28. Steven Kane


  29. DaniΓ«l W. Crompton (webhat)

    never noticed the @voicebunny version before, it’s awesome.

  30. Ro Gupta

    Fred, I think the top of your funnel measurement is way off. GA unique visitors measured over anything longer than a month or so tends to be highly unreliable and inflated due presumably to cookie deletion.I’d suspect the 10 million UV number is actually more in the ~200K – 1M range. While that may seem like a bummer, it actually makes the conversion rates to commenters and donors make more sense to me.To confirm my suspicion, I just looked at GA for Disqus, and it says we’ve had 20 billions UVs over the past 5 years.Cool way to think about things here, though.

    1. Matt A. Myers

      GA uses more than just cookies, even IP address is more reliable as a leading metric than cookies.Obviously it’s not possible to share the Disqus statistics in full, however are things like Event Tracking uniques (which might be using cookies) differentiated out? The reach Disqus has with the widgets, etc. makes it a unique case, though so takes a more nuanced analysis than probably what GA would display for what is good enough for the general use case.

      1. Emily Merkle

        What are you talking about when you say “event tracking uniques”?

        1. Matt A. Myers

          It is a way to track ajax-triggered actions to be associated with a specific user or “unique visitor.”In many analytics systems it is easy to have a discrepancy, when widgets are involved (content loaded on a third-party site, even if through an iframe) because that behaviour is tracked through cookies only.Fred’s “unique visitors” from AVC likely aren’t primarily dependant on cookies for tracking, however actions on widgets – if third-party widget analytics included – are going to influence how accurate the “unique visitor” / user count really is.This is a big difference and influencing factor between AVC and Disqus.If Disqus has this solved to be accurate then I’d like to know how.

          1. Emily Merkle

            Why do you say AVC is not cookie-dependent for tracking uniques?

          2. Matt A. Myers

            Not in the same way as Disqus might be relating to how they track users who would be accessing Disqus through third-party sites (because of the widgets).I am assuming the “20 billion” unique visitor count mentioned above isn’t directly to …

          3. Matt A. Myers

            A downvote? Why not explain why you disagree, if I got something wrong or misunderstand something?Downvoting is lazy and doesn’t bring any value to conversation. And saying it’s lazy that to piss you off so you respond.Still wish Disqus would kill it.And wish I could pull people like @danielha into the conversation … though that could be annoying, so needs to be done right.

          4. Emily Merkle

            I think all of this “voting” is stoopid. Not a popularity contest.

          5. Matt A. Myers

            Upvoting to move things to the “top” has value in the sense of bringing engaging content in front of users faster, which will lead to higher engagement or stickyness overall. A lot is missed or lost though by having it.

          6. Emily Merkle

            Up voting has no internal logic. Just because 20 people vote for a single post doesn’t mean they are up voting for the same reason – and doesn’t mean subsequent readers will find the content “engaging”. Might be valid with clones.

          7. bsoist

            Just because 20 people vote for a single postThis is why I was never a fan of Google’s PageRank, and now why I have to question my use of upvoting. πŸ™‚

          8. Emily Merkle

            Super. Now upvote me 27 times. Momma needs some street cred.

          9. LE

            Agree but at least page rank doesn’t have a time dimension which comments do. A comment posted early on, especially with a certain degree of “profoundity”, ubiquity or something that makes people laugh or that they can relate to will gain more upvotes if done right at the start than the same posted later on. And all that depends on the community makeup.Obviously something being popular doesn’t mean it is correct.

          10. Emily Merkle

            So the key to win the upvote contest is to wake up early, speed-read Fred’s pearls, and start frantically posting profound, funny, ubiquitous missives?I am not raging against the machine because I am envious of the upvote elite. I am more of a thought challenger than complentative muser or comic.I just find it as useless as the people on FB who indiscriminately “like” every reply they get to their posts.

          11. LE

            I just find it as useless as the people on FB who indiscriminately “like” every reply they get to their posts.Hmm. I’ll excuse you for not being familiar with my concept of “party in your brain” [1]. Those people don’t know that you, Emily, feel that way. They think it impresses. And I’m sure it does impress some people.Anyone who can get pleasure out of something that doesn’t cost any money and which doesn’t become to addictive is ahead of the game in life and happiness.[1] In short, it’s not what others think it’s what you think others think (when applied in this particular situation).

          12. Emily Merkle

            You are spot on. Who am I to point fingers at others and shout ” you are too stupid to live!” It’s my opinion, not fact, not the rule of law. Who am I to judge.Thanks for setting me straight – in a kind, sensible manner.

          13. Matt A. Myers

            Agreed, it’s not fully effective – though is the quickest and easiest way to make gains in engagement vs. random.

          14. Drew Meyers

            “Downvoting is lazy and doesn’t bring any value to conversation.”Agreed. Unfortunately, many people in the world are lazy πŸ™

      2. Ro Gupta

        They do use more than just cookies, but UVs over a long time horizon has always been very difficult to measure accurately. Going the panel or hybrid route like comScore may be somewhat more reliable for that kind of measurement, but we know that has its own limitations.Re: Disqus, we hit 1 billion monthly UVs earlier this year, so if useful you can compare to the 20 billion number over 5 years and come to your own conclusions.Re: the distributed widget vs. destination site issue, I checked UVs for, and there was an even bigger disparity between the most recent month and the past 5 years (>50x). The latter number didn’t seem realistic to me, while the former does, so I think it confirms my skepticism here.Re: event tracking, I believe we use Mixpanel mostly for that and it’s not really meant for the sort of audience measurement we’re talking about here.

        1. Matt A. Myers

          So, I’ve used Mixpanel too.Mixpanel creates a new unique user for anyone unidentifiable, even if that’s the same user – but on a different domain, unless you can pass the user’s unique identifier through beforehand – which I’m not sure is possible to do properly.But I assume you have Google Analytics tied into widgets too – or rather, they are loaded in the iframe when comments load. I just don’t know how GA tracks third-party widget activity / traffic, or determines if a user is unique, etc.. Cookies don’t transfer between domains, though if cookies are used at all for tracking unique visitors – then each time a person visits a different blog / site with a Disqus comment system, then they would be seen as a new unique visitor? Save if there is that layer of IP address — but many people share the same IP address, so that on its own isn’t fully reliable either.But I’d say it’s more reliable when the statistics being gathered aren’t influenced by widget loads on third-party sites, etc. as AVC doesn’t do or have. You can read my reply to Emily too, maybe can help understand me better what I mean with that –

          1. Matt A. Myers

            Sample size of 3,430,081 isn’t very large, though I understand that it helps with tracking uniqueness – it will dilute as you get into the masses. The people going to that URL test will is a group of people with certain curiosity, who likely come with certain interests or from certain groups of thought, etc..

          2. Cam MacRae

            You’re right. It’s not very large; it’s enormous.You’re also right that it’s non-random which tempers my enthusiasm.

          3. Timothy Meade

            That number ties into something I’ve been thinking a lot about recently, and maybe Shana can comment on this, but it gets to the point where gathering opinion or sentiment based on raw numbers rather than random polling might make sense. 3MM is about one-hundredth/1% of the US population, and if that many are interested in something like Panopticlick what does that say about the number interested or motivated around more general questions of personal autonomy or privacy?(Corrected stupid math).

          4. Matt A. Myers

            I’m sure you could spend 3-4 years on a PhD trying to start to open up the question and only just start to get a glimpse of the answers..

          5. Cam MacRae

            Simplifying, a sampling plan is a function of desired accuracy and cost.It’s expensive to take a census, but costs are (rapidly) decreasing for some applications (which I take to be your point.)

          6. Ro Gupta

            It may be more reliable for (although fyi/afaik, GA just sees our widget as a page and the websites using the widget are considered referrers)…but like I said with the Disqus blog sanity check, not to mention from when I worked with analytics for a large news site, UV metrics for more than say a few months has always seemed out of whack regardless.

          7. Matt A. Myers

            Just a curiosity. πŸ™‚ Thanks for the dialogue.

        2. ShanaC

          meh, google can do that too (re: event tracking.)I don’t know of any tool for that kind of audience measurement

      3. Emily Merkle

        From what I understand, IP is not more reliable, due to NAT and Tor.

        1. ShanaC

          forget that. Dynamic IP ad well

    2. Emily Merkle

      Nice thinking, tho, cookie deletion is far from widespread / and depends on type of cookie.

    3. ShanaC

      Hard to say. I agree the panel approach is right. I also think top of the funnel to bottom should generally be expressed as a rate change of some sort. And it is atypical to do a year over year reporting.

    4. fredwilson

      great point Ro

  31. petenixey

    I love the stats that you share through this blog, Fred, they’re always super-interesting. These ones got me thinking about the 100/10/1 rule of social media you once mentioned.At first glance your numbers doesn’t play out however I think they’re distorted by people changing machines. Over that time, the unique visitor count could easily have inflated by much as 10x due to clearing cookies / swapping machines, using phones etc (I’ve been reading you since 2004 and I’ve easily visited from that many machines).That gives you a 100:1 visitor:contributor ratio so I wonder whether you’ve also got a 10% “engagement” figure playing out too. I suspect it shows in the number of people who’ve tweeted a link to or shared it on Facebook. I’d hazard a guess that you’ve probably had tweets from about 100k unique individuals but it would be interesting to know.

    1. falicon

      It’s also nearly impossible to tell how much of that ‘unique visitor’ traffic is bot or not…over the weekend my own tiny blog traffic jumped from an average of 10-20 to 400-500…google seems to think it’s real traffic, but it’s almost certainly bot-related (and not sticky)…but it makes my ‘uniques’ seem *way* *way* higher than they actual are.In the end, the numbers don’t matter…the impact does.

      1. Matt A. Myers

        And the impact you see is how action trickles up into different, higher level layers, into the more holistic / – the more whole pieces.

      2. andyidsinga

        true that re the impact.Tell me why you think the traffic is bot even when google thinks its unique? You must have some interesting intuitions here …

        1. falicon

          The various reports let you dig into referrals and sources. Usually for ‘real’ traffic you can dig down to a specific link/mention somewhere that is bringing people in (with a related small uptick in ‘direct’ traffic).In this case, that actually did happen with site/page that is entirely in Chinese…so at least some of that big spike is/was real people checking something out that someone pointed out (though I still can’t actually determine what that is/was)…but there is also a larger volume of a specific referral (twitter / buffer is what it’s being labeled as) that does not appear to actually track back to anything physical (like a link that would be the case if twitter was really sending the traffic).Outside of that sort of digging….it’s also a lot to do with the ‘time on site’…in this case it was about 23 seconds on average….the only reason that makes me think it’s bot is that, a real human following a link out of interest is likely going to spend more than 23 seconds on the content (even if it is as bad as what I write).So it might be ‘real’ people, but it’s not ‘real’ traffic…meaning they aren’t making the intentional decision/choice to visit…and so I count that as ‘bot’ (which btw, I have no problems with since I write bots to do a lot f grunt work for my own selfish needs as well) πŸ™‚

          1. Emily Merkle

            I come from that world. Bots have their place, purpose, value. In addition to time on site, as you mention, folks value pageviews per visit. Bot traffic and uniques are tricky and GA should be augmented with – as fail on sez – referral data.These matters are primarily key to ad buys. If you are working with any network with a conscience or soul, you can get this info, optimize, block sources, etc.

          2. andyidsinga

            BTW, we never did talk about fredbot πŸ™‚

          3. falicon

            Yeah – we ran out of time chatting about everything else…we’ll have to do it via email until I can make it out your way! πŸ˜‰

        2. LE

          Wanted to add to what I said before that the fact that is three letter .com also results in a base amount of background radiation. People type it in thinking it is something else.I own many 3 letter .com’s. I just checked and I have one that gets 36,000 unique visitors per month. That’s 432,000 per year. And 59,000 page views per month. Etc.

          1. falicon

            I’ll give you $50 bucks for it…regardless of what the three letters actually are… πŸ™‚

          2. Emily Merkle

            Right. That will happen. When pigs fly.The 3 letter domain rates are going for your first-born child.

          3. falicon

            OK ok…I’ll go as high as $55…that’s $18+ per letter! How could anyone pass a deal like that up!?! πŸ™‚

          4. Emily Merkle

            I have a few. Or know someone who knows someone.7 for $345.23. That’s my best offer.

          5. LE

            I’ll tell you a story that just happened.I was selling a domain (not 3l) and the person wanted a lower price. They always do. (If they don’t you probably went to cheap!) [1] Obviously I had baked that into the original ask for the domain to begin with.So instead of lowering the price to much I did the following.I lowered it by only $1000 and also told the buyer that I would throw in for no extra charge 2 other domains. The only thing was I registered the two other domains right before I sent the email. So they cost me just the reg fee. I just thought them up on the spot. He went for the deal. The extra names actually were pretty good and related to the main name he wanted. I might have even been able to go back later and sell them to him. But it’s easy to think up names for me so no loss.The bone: I learned the psychology of this as a kid when I went along when my dad shopped for cars. I recognized that a buyer wants to be thrown a bone in order to feel that they are not being taken advantage of. Sometimes they want what you have but can’t admit they’ve been beat. So the bone allows them to walk away with their pride. It’s important.In the case of car buying iirc it was “ok I can’t lower the price but I will throw in the car mats for free”. With a large purchase that works better than “ok I’ll take another $50 off the price”. (Similar to employee benefits and perks vs. salary.)There is more to it than this but that’s the short of it.Oh one more thing this buyer is on TV and has a show where he invests in businesses.[1] Why it’s always important to have some kickback in a deal. Even if the kickback isn’t price but something else (like delivery or another non monetary condition). If there is no kickback sometimes the buyer/seller feels they did a bad deal. Kickback allows someone to feel as if they just slipped by.

          6. Matt A. Myers

            Sometimes it’s not being thrown a bone but price/money actually matters.I wonder if I’ve ever bought a domain from you unknowingly..

          7. LE

            Oh yeah price definitely matters.

          8. falicon

            Great story and insight!

          9. Matt A. Myers

            You traffic and domain whore…

          10. Matt A. Myers


          11. LE

            Sometimes having really good merchandise can be a curse.For example if you own something that is grade b or c you sell it and make what you make. But if you own grade a+ you hold on waiting for the big kill.Statistically the number of domains that sell for big dollars is small compared to the number of domains out there that sell at all or are just being held. You hear about the big sales you don’t hear about the ones that don’t get any decent offers.The base value of physical real estate is directly related to what end users will pay. So if you need to unload physical real estate you can you just have to discount it and make it attractive to someone who will sit on it until a end user comes along and pays market price. [1]domains on the other hand are a much longer waiting game. Can take 10 years until the right buyer comes along. So a wholesaler will factor that in and only buy really really cheap.[1] For example if you buy a condo unit in a 150 unit building there is a certain turnover of units in any given time period from people looking in that particular building. It’s merchandise that people are looking for and that they want. Doesn’t work that way with domains.

          12. andyidsinga


      3. LE

        I used to have a case a few years ago where every Monday a bot would come along and click very reliably on a particular “parked” domain page that had links on it. This happened for quite some time. I was amazed that the google algorithims didn’t catch it. I don’t remember the exact particulars but let’s say a parked page that normally got .75 on a given day might get $25 to $45.00 on the mondays. So any human could have easily seen something weird was going on. But it went on iirc for maybe a year maybe more.This still happens on other pages from time to time but not an the same regular reliable basis as this one page in particular. I gave all the unearned money back to google of course but they never cashed the check.

        1. Emily Merkle

          GA has not quite mastered that territory yet. It is nuanced, tricky, complicated, sometimes shady.

        2. falicon

          It’s to Google’s advantage to have some bot action…the loser is really the advertiser (who is the one ultimately paying for the false impressions/clicks).The trouble is that no one is *really* incentivized to fix/address the problem…the site owner gets the vanity metrics, google gets their cut, the advertiser appears to make some quota on ‘reach’ (even though it’s not bottom-line impactful — and hence the complaint that “ads don’t work”).

          1. Emily Merkle

            If you are working with an ad slinger that has a brain and a soul, they can and will work with you to decrease the probability of being sent bad traffic. Plenty of motivation – happy client = more budget over more time.Been doing this 12 years.

      4. laurie kalmanson

        smaller, more engaged core > bigger less engaged audienceengagement > volumecontent drives engagementawesome content, big engagement

    2. fredwilson

      yup, Ro made the same point. you guys are right.

  32. kirklove

    #Pimp and #MetricsWhore – impressive, Buster

    1. LE

      I’d also like stats on the wordcount of the posts and the word count of the comments and replies to comments.

      1. Andrew Kennedy

        I’d love to see your average word count per post….. Top of the league table on that stat I am sure.

        1. LE

          Now that sigma algebra is writing shorter comments.

          1. Matt A. Myers

            Made me laugh. Thanks. πŸ™‚

      2. falicon

        If/when I have time and remember, I will get those for you…

        1. LE

          One of the ways Disqus could monetize is by figuring out a way to offer special deals to people who are more engaged than others. Not lame credit card perks (which are just repackaged stuff that is offered in other places) but something that recognizes the distinct group of people who for one reason or another are more engaged in a particular community.

          1. Matt A. Myers

            I don’t think the internet and advertisers are organized enough yet to make this an viable, affordable task.

          2. Emily Merkle

            Yes, they are. Depends on how you define engagement.

      3. Emily Merkle

        I want to see net worth.

        1. LE

          One of the interesting things about net worth is that it is always stated as “so and so is worth $y”. But as anyone who has gone through divorce knows, the spouse many times will get 60% of $y so net worth could be better expressed as “x is worth .4y”.

  33. LE

    “1,322 folks from the AVC community have supported”Unique folks? – I wonder if that takes into account 1 person donating multiple times.

  34. LE

    – There have been 154,490 comments on AVC. 22,490 of them were on the TypePad comment system in the first four years (~5,500 per year) and 132,000 of them have been on disqus in the following six years (~22,000 per year).Does the 154k take into account your comments?Excluding your comments, the top 20 commentors by my quick calculation account for about 82,000 of those 154k comments.Or, if the top commentor calcs are just for disqus then 82k of the 132k or about 62%.

  35. bengfai

    I thought it would be interesting to see how AVC has evolved over time … So I wrote a basic script to scrape (I know, so inelegant) this data from Fred’s archives over lunch.I’ve attached a screen shot of the early results …The most commented post was “Occupying My Mind” –…The longest post (by words) was “Friday’s Chat With Howard” –… (which I realise is a transcript)Unsurprisingly, the most used tag is “Venture Capital and Technology” with 3,385 uses.

    1. bsoist

      Excellent! I am working on pulling some data from Disqus now.

    2. bsoist

      So I wrote a basic script to scrape … this data from Fred’s archives over lunch.Favorite line of the day!Number of comments are very close to what we found from Disqus and

      1. bengfai

        Thanks @bsoist:disqus !Happy to share the raw data if you want it? Let me know.With @falicon:disqus’s stats, your’s and what @fredwilson:disqus has already mentioned we could put together a pretty funky, dare I say it, infographic πŸ™‚

    3. fredwilson

      wow, wow, wow, can you send me the data via email? this is so awesome to see.

  36. John Revay

    AVC by the numbers

  37. awaldstein

    Fred–the stat that always stops me is time on site.When I see hours of time on a post across a cross section of uniques, something is working.You have person years on yours, if not a few lifetimes!

  38. Prokofy

    The social media power curve — a tiny percentage of the readers provide the content for others to view.