The Official Blog of Cogmap, the Org Chart Wiki


Archive for June, 2007


Transaction Costs in Web 2.0

Wednesday, June 27th, 2007

Coasean economics is something most of the Internet is intimately familiar with.  Ronald Coase won a Nobel prize for describing transaction costs.  When he described how firms work, he determined that a firm could continue to expand as long as the additional cost of the expansion was less than the cost of acquiring the same service from outside plus the transaction costs associated with integrating the outside service.

Technology has changed the nature of the firm in many ways.  When Coase first formulated his theories many of the ideas centered on pulling services into organizations, eliminating transaction costs and lowering costs by having in-house HR, IT, manufacturing and marketing.  Firms could get larger and larger by taking services internal, eliminating costs associated with vendors, lowering prices to reflect this cost, and expanding market share.  Theories like this were the basis for companies like GE.

Today, all of these approaches have been radically altered.  Transaction costs have declined so sharply due to the Internet that the nature of the firm no longer tends towards expansion, but instead tends toward reduction.  How small can I make my company?  Everything can be outsourced easily and is built for integration.

Let me offer up this idea then:  Internet Company 2.0.  Marc Andreesen recently poked the blogosphere with a sharp stick by saying there is no Web 2.0.  There is definitely some spin here, but I would say that one of the most interesting things about Web 2.0 is the capital structure that drives it.  Lots of people have talked about how capital requirements have changed to build Web 2.0 companies.  We live in an era where Internet Companies have virtually no capital requirements and this is because all of the outsource-able things that a company might need have been outsourced, their cost has gone to almost zero, and the transaction cost to take advantage of the service has gone to zero.

Technology is the most common example: Hosting is a couple of bucks a month, development environments are free, everything is open-source, and support is in free forums of like-minded samaritans.  Marketing is another area where costs are rock-bottom.

This doesn’t apply to other businesses.  I love software because there are no manufacturing costs, but it certainly offers a new model for business building that is exciting, inexpensive, and offers more people the opportunity to be entrepreneurs, which I love.

This post was supposed to be about APIs, but I guess that will be the next post.

Rendering matrix relationships in Cogmap

Thursday, June 21st, 2007

How should we render matrix relationships?  Many people ask for this feature, but I have rarely seen it well-expressed in any organization chart.  Obviously, for an application like Cogmap it has to be something that scales infinitely.  How does that work.  It is hard as the matrix gets bigger.

Some people have suggested I allow making copies of people in org charts.  That seems to lack elegance because it is not obvious that a person has a matrix reporting set of relationships.

Post links to pictures of your ideas, maybe?

Trip to San Francisco

Sunday, June 17th, 2007

God forbid someone reads the Cogblog that is not already a close personal Cog, if you are in San Francisco and want to hang out this week, drop me an email – brent at cogmap dot com.

Cogmap supports hCards

Sunday, June 17th, 2007

Another step in making Cogmap more data friendly, profiles are now hCards.

Go check out a profile: Alex Markson, for example.

If you have installed a microformat detection system like Operator (Mozilla Firefox extension) or Tails (Mozilla Firefox extension) then you have lots of fun ways to look at and download the data on the page.

Props to Kevin Lawver for convincing us that it was worth the trouble.

Sorry we don’t have a more inspiring post, but I am pretty sure that I am meeting Chris Messina tomorrow so I figured we had to get this launched so we would have at least one thing to talk about!

Organization Chart API: XOXO XML charts!

Sunday, June 17th, 2007

UPDATE: Massive changes to this:

Organization Chart Data is accessible as an XOXO document by calling a URL formatted as<CHART ID>

xoxo.php returns a nested set of unordered lists that describes the organization. Each listed item also contains several pieces of metadata associated with the list. The format is:

<li><a href=”profile.php?id=<personID>”> <FirstName> <LastName> </a> <dl> <dt>Title</dt> <dd><TitleValue></dd> <dt>CogID</dt> <dd><personID></dd> </dl>

  • personID: the ID in Cogmap’s database
  • FirstName: Person’s first name
  • LastName: Person’s last name
  • TitleValue: Person’s title
  • The title of the document is the chart’s name.

So here is an awesome example: XOXO No Inc Chart!

Hopefully this starts to satisfy the data-consumers out there like Stamen Design, who were emailing ideas for how to implement this virtually from launch.

Let the gobbling of organizational data begin! Make sure you post something here or send us email if you do something cool.

Book Review: The Spider and The Starfish

Sunday, June 17th, 2007

This was a book I saw someone blog about that seemed very topical to work, so I figured I would pick it up.

I thought this book was an absolute train wreck.  I can’t even link to it because it does not deserve what little link juice I have.  How can a book about centralization versus decentralization not discuss Coase or transaction costs or say the word “friction”?  And I should love a book like this.  Cogmap is a poster-child for the power of decentralization and open source data communities.

Before I bash this book, let me start with the Amazon reviews.  First review is from Auren Hoffman (a man regularly savaged in Valleywag!).  He says, “The book discusses the management techniques of wikipedia, craigslist, al Qaeda, the blogosphere, and more. Though these are first time authors, I found the book mimics the unique observations of someone like Malcolm Gladwell.”  Of course, he neglects to mention that he is a case study in the book and is insanely biased.

The other “Spotlight review” is a Top 50 amazon reviewer who somehow decides to use his review as a forum to break down his political theories and bash the CIA, his former employer.  The pro-decentralization theme of the book offered the perfect chance to bash the government in a review despite the fact that government was never discussed in the book.  Fortunately, I loved the comments on the spotlight review.  Here was my favorite, which I must quote in it’s entirety:

“Blah, Blah, Blah! This was the worst review I have ever read in the 9 years I have been reading them on Amazon! My favorite part is he blather on Mr. junk science (Al Gore). At his peak Hitler used to have crowds of 100,000 interupting him with applause over and over again. Like that means anything.

Read the Halo Effect by Rosenzweig, to see why books like this are such oversimplifications for the masses. Just like the Global Warming movement. Real scientists don’t make sweeping generalizations of complicated cause and effect relationships and then ask people to go out and make decisions based upon weak evidence. That’s what politicians do! This guy should be dropped from the top 50 reviewers list. I will certainly never read his reviews again!!!”

It terrifies me to think how many other suckers bought the book after reading the spotlight reviews.  I was tricked, I swear.  Anyway, on to my book bashing:

The authors were so lacking in analytical capabilities that the book had no prescriptive value.  They offer up some models, but caveat the models with comments such as “A decentralized organization stands on five legs.  As with the starfish, it can lose a leg or two and still survive.”  They then offer up strategies which their earliest examples of successful decentralization do not rely on.  Was there a pre-existing network for P2P music theft?  Was there a catalyst/champion advocating eMule?  Does stealing music involve a sense of community?  I know the music I procured using Napster did not involve building community.

Another great line: “Decentralized user ratings proved to be eBay’s biggest competitive advantage.”  Ironically, they explain network effects in relation to eBay, but they emphasize it in the context of user ratings (“But so far, no one’s been able to come up with a better technology than eBay’s user rating system.  Buyers and sellers therefore stay at eBay – it’s where the action is, and it’s where they can find a network of trusted buyers and sellers.”) 

First, a statement like this completely misinterprets network effects, which is strange because they seem to understand it in other parts of the book.  Network effects would imply that because they have user ratings, they tend to get more user ratings.  I don’t think that this is true.

Second, eBay’s user rating system is not their advantage.  Their advantage lies in the strength of their marketplace.  If they did not have the most buyers, sellers would go elsewhere.  If they did not have the most sellers, buyers would go elsewhere.  Because they have both, all new sellers and buyers go there for the strongest and most competitive pricing.  Even if Yahoo! rolled out a better user rating system tomorrow, or even one that allowed people to import eBay’s ratings, I don’t the odds are good that they would overcome eBay’s powerful network effect.

How is this for analytical: “The decentralized sweet spot is the point along the centralized-decentralized continuum that yields the best competitive position.  In a way, finding the sweet spot is like Goldilocks eating the various bowls of porridge: this one is too hot, this one is too cold, this one is just right.”  Helpful advice?

Another comment on eBay: “The company would lose market share if it moved further toward either centralization or decentralization.”  Yet how do we know this to be so?  No analysis was done or model offered that demonstrated that this is the most precise and sweetest spot.  They compound this challenge with an assertion in the context of their Toyota versus GM case study: “Toyota occupied the decentralized sweet spot in the automotive industry.”  Yet we have know way of knowing if it was the sweet spot or just a spot that was sweeter than everyone else’s spot.

If you have enough time on your hands to read this book, go read The Innovator’s Dilemma again or something.  This book is not nearly rigorous enough to satisfy anyone.

Do widgets ever generate revenue?

Wednesday, June 13th, 2007

Jeremy Liew recently posted on the wave of posts (most notably Marc Andreesen and Seth Goldstein) about API power and how that works with Facebook.  I am working on my own post in this area, but obviously the part I gravitated toward was his mentions of the need for a social network advertising standard for widget vendors.  He cites this as the key to creating an environment for API developers to create revenue streams.

I am not so sure.  Before you have common standards for social network advertising, there has to be a demonstrated success of a mechanism that actually makes money (and probably competing incompatible standards that both appear to work).  I cannot think of a single widget (and maybe this is my lack of exposure) that is generating anything more than pennies in revenue.  Widgets are marketing expense.  They are part of a service/drive traffic to a service by adding some sort of value on other pages.  There have not been nearly enough good examples of a monetization model that works across widget genres.  For example, pre-roll ads would work in video widgets or slide widgets, but not in something like mybloglog.  Until some common ideas emerge for how to monetize these things, how can we drive to a standard?

Even his post on the need for a standard starts by citing a study where MySpace basically tries to sell that “friending brands works”.  Can a widget company survive on selling skins?  I have no idea.  Jeremy needs to be telling his portfolio companies to have Dynamic Logic run a branding study or something on the impact of widget skins.

AttentionTrust and the Network Advertising Initiative

Tuesday, June 12th, 2007

I read a fair amount about AttentionTrust in the press but almost nothing about the Network Advertising Initiative and this seems strange to me.  Maybe this says more about their respective interests in seeking an audience than the audiences interest in hearing their views, but let’s assume not.  I think people that think about AttentionTrust a lot are kind of missing the boat here because I have never heard these two organizations mentioned in the same sentence before. 

AttentionTrust was the non-profit associate of the Root Market model that advocated people get on the bandwagon and manage and sell their own clickstreams.  The problem with this, as I pointed out earlier, is that aggregating all of that data one at a time from consumers is probably too inefficient.  What most advertising networks are doing instead is negotiating directly with a publisher. 

If you need one million people in a population of TV watchers, would you rather negotiate with one million consumers or negotiate once with  The approach is confirmed by the off-line corollary which is buying lists.  Company-to-company negotiations are more commoditized, so it is easier.  The Network Advertising Initiative is the non-profit that is addressing this issue.  They advocate for consumers in the context of how companies share on-line data.

AttentionTrust gets a lot of press and interviews for advocating protection of consumer clickstream data, but day-to-day the Network Advertising Initiative and the NAI principles are probably impacting more privacy related activities.  Are these groups synced up?  I have no idea.

Do I love AttentionTrust’s message?  Sure, why not.  Am I familiar with what they are doing day-to-day?  Nope.  Am I sticking my foot in my mouth?  Probably.  This is basically also true of NAI. 

But that is what I am seeing out here in the trenches.

(I must emphasize again that I have an employer, these views do not reflect the views of my employer, I look forward to using this forum to become more educated regarding the myriad topics I am ignorant of, etc.)

P.S. I must also say that Seth Goldstein, the man, the myth, the legend himself, was kind enough to post a comment to clarify Root Market’s model a bit.  Obviously, they were smart enough to do the same math I did.  I just published it for the world to see my math skillz.  He has been kind enough to arrange a get-together when I am in San Francisco next week where I am sure I will get an earful of AttentionTrust, so I will post on fun stuff at that time.

P.P.S.  I haven’t posted my “thoughts on other bloggers” post, but I wrote it a few weeks ago.  Letting it age like a fine wine.  But I have to say that my thoughts on Seth’s blog were right in line with Josh’s critique.  I was pretty sure I was just not smart or well-educated enough to understand his renaissance-man style of posting.  Looking forward to seeing more frequent and more off-the-cuff posts from Mr. Goldstein.

The Urbis Business Model

Friday, June 8th, 2007

I love Urbis.  Urbis is a fascinating web site, although I don’t enjoy spending time on it.  What I love is the mechanism.  Urbis is a site for people to post poetry and stories.  The kicker is the review system.  People review your work and give you feedback, but to read the reviews you need “points”.  To get points, you write reviews.  No surprise, there is not a 1:1 correlation.

The result is a page view generating machine.  Who can choose not to read a review someone wrote about you?  If someone writes a review of your stuff, you have to read it.  So there are always reviews being written and people unlocking reviews.  This is a great example of many of the mechanisms people are talking about today when they talk about using game mechanics to make social media more exciting.

Unfortunately, this is probably not a super-valuable audience to advertisers, although it is the creative class.  Maybe these people are the mythical “influencers”?  Also, there are probably the typical UGC concerns advertisers have with much of this content.

I also love their home page.  It does a great job of giving a sense of active community.  They have done several nice implementations of Ajax throughout the site.  Despite the page view generating nature of their site, and despite the constant reviewing, unlocking, reading, and rating that goes on, they don’t force a lot of page reloads.  Pages change dynamically and are intuitive and easy to use.  This is a marked difference from MySpace, where everything requires an entire screen refresh (with accompanying advertisements).

Regardless, I would love to find a more interesting market to apply this review/page view generation mechanism.  So far my best idea has been a “blog review site”.  I wish I could come up with better.  Post ideas in the comments if you have ‘em.

Root Market & AttentionTrust: Bad Idea?

Wednesday, June 6th, 2007

Having never had a blog before, I never had the chance to get some perspectives out in the market that I have been dying to get off my chest.  Here is one:

So when I first heard of Root Markets, I thought, “hmmm, so they are gathering behavioral data directly from consumers.  I should learn more about this because it is something unique and different in the market.”  That is true and I love people thinking out of the box, however that model is super-hard to make work.  Their proclaimed business model was that they would sell this data to advertisers for behaviorally targeted advertising campaigns, however that model requires relatively large populations to target to create efficiencies in advertising.  For example, let’s say that I wanted to spend $70,000 at $10 CPMs to target people that have visited TV-related web sites.  That’s 7 million impressions.  Assuming we run this for a week, pace it evenly and have some frequency cap like 3/24, you need 333,334 unique people that have exhibited that behavior.

And that assumes that we see all 333,334 people 3 times every day for a week.  Here, network reach comes into play, but even if your network was the entire Internet, that probably doesn’t work.  Best case, maybe you need 600,000 unique visitors that have exhibited the behavior you are looking for.  Realistically, you probably need 2 or 3 million?

Is it realistic to think that Root could collect, package, and sell that volume of behavioral data?  Especially when they imply that they will allow people to essentially opt-in to an advertisement?  So Root espoused that they would allow people to essentially be compensated for receiving messages.  Let’s use this example some more.  Maybe 30% of the $70,000 would be spent by Root to acquire inventory.  So the maximum amount left over for paying consumers would $50,000.  So if we say you need 2 million people in the targetable population, then each one could get an offer from the marketer of $0.025.  Holy micropayment!

And that assumes that Root runs the business at no margin and gets inventory dirt cheap.  If they have to rev share with publishers, that could be cut in half.

Would this change if the behavior was more direct response?  Let’s run some quick numbers there: A mortgage lead (because Root was obsessed with mortgage arbitrage) might be worth $35.  If someone exhibits a behavior “correlated” with mortgage leads, maybe you have a CTR of 2% for ads you show them and then a conversion rate of 5%.  Once again, you have to buy the inventory, so lets carve out 30% for that.  Ok, now we do the math and see what we have: Each person that exhibited that behavior could get a big fat check for $0.0245.  Not as lucrative a business.  And frankly, there are probably much smaller populations of people “loving mortgages” then people “loving TV”, so the direct response opportunity has lower absolute dollar value.

So if you are an individual participating in Roots scheme and it takes off and millions of people sign up and advertisers are shooting them offers left and right, are these opportunities lucrative enough for people to take the time to review and accept them?  If you had 20 advertisers similar to the two examples offering you the chance to see their ads as you surfed the Internet (is it realistic for someone to opt into all of these offers or is there a saturation point where they can view no further ads?) and you reviewed and opted into each offer then you would have made a little less than $0.50.

Root, I love you for being different, but I am not smart enough to see how it works!  Of course, I think they subsequently switched directions and are no longer doing this.  But they got a TON of press for a few months.  Why didn’t anyone else do this math?  Why didn’t the business plan writers?

This post was insanely long.  I will do another one to talk about AttentionTrust and the more macro business challenge in a bit.