Umair Haque / Bubblegeneration
umair haque  


Design principles for 21st century companies, markets, and economies. Foreword by Gary Hamel. Coming January 4th. Pre-order at Amazon.

Tuesday, September 13, 2005

Web 2.0 Doesn't Does Exist, eBay + Skype, and Network Scale Economies

OK. What is Web 2.0? Is there such a thing? Is it just FUD that non-geeks fall for? All questions that have been asked lately. Good ones, too - I think it's important to define this.

My answer is (you guessed it) about economics, not technology. I mention it below, in my Skype posting. Here it is: Web 2.0 is a shift to from tight, hierarchical architectures which realize exponential network FX, to loosely structured architecture which realize combinatorial network FX.

Hopefully, you can pretty simply see from the graphs what I'm talking about - that combinatorial gains (seriously) dominate exponential gains (seriously) dominate linear gains. If I can get a bit geeky for a sec, from a world of x^n, to a world of n^x, where n is network size, and x is utility/value/etc. This is Metcalfe vs Reed, if you like, although I think several economists were first to these ideas.

More simply, Web 2.0 is about the shift from network search economies, which realize mild exponential gains - your utility is bounded by the number of things (people, etc) you can find on the network - to network coordination economies, which realize combinatorial gains: your utility is bounded by the number of things (transactions, etc) you can do on the network.

The point is that this shift is combinatorial - each person can do X activities in a combinatorial network, and it's combinations of these activities that make value explode. Contrast with a exponential network, where it's the number of people on a network that create value. That is, a relationship between any two people is 1:1 in an exponential network, but many to many in a combinatorial network. It should be intuitive to you that the former kind of network has more potential for value creation.

The graphs make this pretty clear. They're meant to show what the sources of value for these three players are. Pretty clearly, eBay's main value driver is it's user base - the regression between revenues and users is almost perfect. It's not a combinatorial network; eBay is not creating value by letting activities scale faster than users. That means eBay is realizing almost no network scale economies.

Check out what Reed notes as a constraint to network FX:

"...each user typically has a fixed, small set of friends and family that they call all of the time. Since the value of these services to a particular user does not depend on the number of other users of the network, the total value of these services grows more slowly�proportional to N, not N�."

Does this sound familiar? It should - it's the eBay problem, in a nutshell.

Now, for Yahoo, and especially for Google, things get more interesting. They're deriving more and more revenue relative to users; there's not a simple linear relationship between users and value like there is for eBay. They are realizing network scale economies.

Clearly, users are not the primary value driver here - it's the transactions they engage in; that is, Google is creating (and capturing) more and more value from users as the user base grows, and Yahoo is trying, but only marginally succeeding.

The implication is that Google and Yahoo are mediating many to many relationships between users - transactions, if you like. Yahoo's deriving mild exponential gains from weak network economies, and Google is (almost) deriving combinatorial gains from strong network economies.

Why is Google realizing stronger economies? Well, pretty intuitively, it's done a much better job at both search and coordination. At least as far as online advertising goes, it's made it easy to find buyers and sellers, and to then close the deal relatively costlessly. Yahoo's still trying to find it's feet as a real maker of network markets, versus a simple aggregator/distributor, which is pretty clearly demonstrated by the graph (and should be intuitive).

This is an old theme of mine - from search to coordination. I hope this makes things a bit more concrete for those of you who I haven't chatted with.

I also hope it helps you see why so many people are getting excited about the value creation potential of Web 2.0. If the theory's right, it's going to be a very valuable market. And since the theory for Web 1.0 was pretty much on the money (despite the crash, evil ibankers, silly analysts, etc), I'd say fasten your seatbelts - note that I've used revenues in the graphs; in fact, I should've used a cap-based measure, which would have exaggerated scale effects (but didn't have time). In short, this is very real value creation we're talking about.

Technical notes: I should've used a cap-based measure; didn't have time. My user numbers are probably off, since they're based on Googled public data, but hopefully close to get the idea across. I'm also comparing global Yahoo to US Google. The scale effects should stay similar regardless. Oh yeah, and this post is really badly written, but again, no time to edit...sorry.

-- umair // 11:05 AM // 20 comments


If you were to plot average CPC (cost per click) * (number of searches) over time, I suspect you would get a plot like the Google one. Both are growing. Maybe just CPC alone is growing like the Google plot.

Rising CPC is in large part determined by scarcity on search engine response pages. Can you give a more concrete explanation of how that is explained by your mumbo-jumbo?
// Anonymous Anonymous // 12:32 AM

hi umair,
A good post !! Here is a post on web2.0 that I read recently that you might find interesting too.
"Why Web2.0 Matters: Preparing for Glocalization" -
// Blogger Rajan // 3:01 AM

Why do you think Google can build communities (2^n) while Ebay can only make connections (n^2)?
// Anonymous Nivi // 10:55 PM

Good Post Umair, but ...

I think you're missing the real reason for exponential vs. cumulative value in the commons - I think it simply lies in "re-use", not "1:1" - and that you originally had it right in the atomizing hand.

My theory is that if re-use and "modularity" are the prerequisites for CBPP, then they probably occur in a range which differs by marketplace - and understanding any particular marketplace's place in and levaerage of this range is what determines returns.

IMO, the (only) reason eBay's marketplace value is not cumulative is because the stuff its participants "do" is not re-usable -> you can only sell your "microchunk" once on eBay; so the 1:1 relationship is no mystery. I don't think "1:1" is a barrier to cumulative value though - it has more to do with the commercial nature of the microchunk and the license /contract under which it is sold.

If a community fascilitates re-use, its value will be cumulative regardless of whether or not the transactions are 1:1. iStockPhoto ( is a classic case in point - the transactions in its marketplace are all 1:1 but the value created is cumulative because each microchunk can be re-sold /re-used. Here is a "1:1" marketplace that creates cumulative value.

Back to eBay. With the purchase of Skype, eBay will create a marketplace for professional services; listed on eBay at a $/min basis, fulfilled via skype.

When this happens: A "service listing" on eBay, unlike today's "product listing" becomes a re-usable microchunk from the perspective of the seller! For this reason, eBay + Skype has the potential to start creating cumulative value once eBay adds professional services to their Marketplace."
// Blogger David Gibbons // 6:05 PM

Folks, when you say "exponential", you mean polynomial, and when you say "combinatorial", you seem to mean exponential.
// Anonymous Anonymous // 7:48 PM

not a mathematician but ... I think exponential is right - it describes the network effect in the community's growth in participants - in a 1:1 community, the production is linearly related to membership as in ebay's case, where you have an average listings releated to an average transactions per member - but membership growth, the multiplier, is still exponential - until any market is capped that is.

in communities with a high re-use or many-to-many transactions, each members production, and membership growth, are both exponential - the product, output, is "combinatorial", cumulative (for the participant) - maybe the formula's a polynomial - whatever, it works like the bomb
// Blogger David Gibbons // 10:59 PM

GFN-wise, the big money opportunity lies in orchestrating a process network for customized education and career services. Such a network (a.k.a. workflow market) will maximize the aggregate productivity of all GFNs, by minimizing the lag between 1) the emergence of an opportunity for intra-/inter-GFN wealth creation, and 2) the "emergence" of sufficient numbers of GFNers with the requisite skill sets to fully "mine" the opportunity ASAP.
// Blogger Frank Ruscica // 5:09 AM

Thank you for the interesting post! Realy nice.
// Anonymous jonky // 10:10 PM

Thanks for good post, Umair.
// Anonymous Hotmald // 8:11 PM

hi umair,
// Anonymous Anonymous // 6:54 PM

hi umair,
I think Web 2.0 is our future. The Web applications will become more stable and more accessible. It's probably one of the biggest revolutions of the internet.
// Anonymous Lisa Santoro // 6:56 PM

A good post. Realy good!
// Anonymous franky // 6:29 PM

Folks, you're right!
// Anonymous boltur // 6:31 PM

That is our future. I think so!
// Anonymous thailand // 6:34 PM

I agree wiht you It is really good!
// Blogger CroatiaClub // 7:47 PM
// Anonymous Anonymous // 7:50 PM
// Anonymous CroatiaClub // 7:52 PM

Thank you!
[url=]My homepage[/url] | [url=]Cool site[/url]
// Anonymous Anonymous // 11:31 PM

Thank you!
My homepage | Please visit
// Anonymous Anonymous // 11:31 PM

Good design! |
// Anonymous Anonymous // 11:31 PM
Post a Comment

Recent Tweets


    due diligence
    a vc
    tj's weblog
    venture chronicles
    the big picture
    bill burnham
    babak nivi
    n-c thoughts
    london gsb

    chicago fed
    dallas fed
    ny fed
    world bank
    nouriel roubini


    uhaque (dot) mba2003 (at) london (dot) edu


    atom feed

    technorati profile

    blog archives