Chez Serge

The Internet disruption

Archive for category business model

Facebook, the future operating system of the Internet ?

Sep 5

Posted by serge in Little thoughts, business model | 10 Comments

Facebook is, apparently, a social network. Let us point out that a social network is not a directory of persons, but a directory of links. The underlying principle is the Six Degrees Of Separation theory, which says that between any two people on earth, there are no more than five intermediaries. Take a bushman, for example:  you know someone who knows someone who, etc. who knows the bushman. All online social network are the same, in that they are not that much interested in people, but rather in relations. The proof is simple: if someone quits Facebook definitively (which is not that easy), what disappears is more than one entry in the directory, it is all links and all interactions with one’s 130 friends (the average number in mid 2010).

We shall not dwell on the amazing Facebook statistics; 500 million members, whose half connect every day, thus contradicting a sentence I hear quite often: “I am fed up with Facebook, and I quit” (sentence putting Facebook equal to the television). Actually, the celebrity on Facebook who lasted the shortest time was a charming English Lady, Ivy Bean, who joined Facebook in 2008 at the age of 102, and died in July 2010 with 5,000 friends, and also 56,000 followers on twitter.

Facebook is fundamentally different from other social networks, by at least one aspect: its open APIs. Most social networks, including professional ones, have very poor semantics in respect to the links. Linkedin, for example, allows two main definitions for the link : “we know each other“, with a few parameters (we are colleagues/one is the boss of the other/we have contracted work between one another , etc.), and “we are part of a same group“.

In Facebook, the semantic of the link is open through a set of programming interfaces, thus enriching the relation between two or more persons. Users can then calibrate their interaction, for instance they may poke each other, send flowers, play together, invite one another to an aperitif, or even share a kitty to buy a common gift. This opening of the APIs is what attracts brands, with the hope to make low cost viral marketing.

Facebook was, from the outset, designed to be an applications platform, something much more sophisticated than a simple social network. And so, little by little, Facebook overflowed out of it’s platform, and, as a cuckoo, positioned itself on sites beyond its own, each time bringing with it, a very interesting function, albeit a little intrusive.

The first one was Facebook Connect . The principle here, is quite straight forward: when someone develops a web site which requires authentication, why bother developing this feature, when Facebook offers it for free, and at low integration cost. The interesting side effect is that the end user is no longer obliged to enter one more time their own identification. It is a real win-win-win deal : the user avoids burden, the web site owner avoids development, and Facebook gains not only more users, but also more knowledge about usage. One should remark that, on Facebook, one has usually a single identity, which is contrary to the basic usage of the Internet: why should one be the same in a professional bulletin board, a forum of enthusiasts, on dating site, or as an avatar in Second Life? Facebook is always chasing for people taking another identity than his own; Facebook is anything but anonymous.

Then came the like button, which was also free to put on any other web site. If someone browses a site, and likes it, they can very easily share it, by making a single click and have it then appear on their Facebook wall. It is the same win-win logic, the web site developer sees his own viral promotion done for free. The same logic applies for other plugins : recommendations, videos, and so on. They are all offered by Facebook, under the obvious term “Social Plugin“, with the intention to be the viral marketing enablers; not only on Facebook site, but on any other site. This is a big shift from other social networks, and is in the same spirit as the Amazon set of widgets which were invented 5 years ago.

The very last one, still under test, is the button subscribe to, which allows users to follow someone else’s messages; a service directly competing with Twitter.

Facebook is forging its expansion in creative ways. In what is perhaps the strongest signal yet, it made a significant shift in its relationship to virtual goods. Those little virtual objects, which are to be found on social platforms, and hugely in 3D immersive platforms, represented, only in the US, a 3 billion US dollars market in 2009. Facebook started selling them in 2008, but the revenue was low, a few ten million dollars, almost nothing. Facebook therefore decided, earlier this year, to totally change its strategy, stop selling virtual goods, offering instead, a virtual currency, Facebook Credits, which allows 3rd party virtual goods to be brought and sold inside the platform. If the same logic applies, this virtual money could, one day, be used on other platforms. Will Facebook become the apps store of virtual goods? Will Facebook also be interested in the market of real goods, trying to do better than Google Checkout, an attempt to be a front-end, unique payment system, which never really took off?

Then, very recently, came Facebook Places…

The market of local information is the biggest battlefield of the Internet nowadays. Being straightforward: I can quite easily, owing to twitter or blogs, know what happens in the street of Tehran. But it is 5 to 8pm on Sunday in Paris, there are three boulangeries which close at 8, I have not enough time to visit all of them, and I don’t know which one has remaining bread…

Many actors are in this market of local information. Google with Google maps, Yellow Pages, Craig’s list, Tripadvisor, Aroundme, and the last one, Foursquare. Facebook is now clearly entering this arena, and wishes to position itself on geo-localized data . Mobile Facebook apps already offer this function, albeit in the US only, at the time of this writing.

The difference between Facebook and Google is striking. We must not forget that the power of the Web is in peer to peer relations. Google, counter-intuitively perhaps, has never been 2.0. Google groups, is a revamping of the Usenet hierarchy, through the acquisition of dejanews. It is the only Google service where users are interconnected. On the other hand, Facebook Places, like Foursquare or Aroundme, allows users to exchange information on a wall. Google Maps does not.

If Google is still heavily controlling search, and advertisements on the Internet, Facebook is positioning itself more and more in crucial places: authentication, virtual goods and why not one day real goods payment, exchange of information, local information; and all this not only on its own platform, but everywhere, through its plugins. For Google, the search engine is free, and makes its income by leveraging other products, such as AdSense or AdWords. Facebook is now doing what Google has done so well, giving away its core service, to allow side businesses to come in. The difference is that it is doing it where Google is not that present: peer to peer.

Facebook is clearly taking control of some crucial applicative layers of the internet, specially the viral ones, with a probable desire to be, one day, the operating system of the Internet.

Tags: ,

About the french “Graduate response” law

Mar 9

Posted by serge in business model, innovation | No Comments

We are, in France, facing a big threat : the introduction of the graduate response bill, a law which would require Internet Service Providers to disconnect subscribers involved in multiple instances of illegal file-sharing (see here or here for some comprehensive understanding). Urgency procedure have been used to avoid multiple discussion and to pass the law as quickly as possible.

I had, before this, sent an open letter to our minister of digital economy. There was a strong response from other people on this topic, and I have been asked to translate its content in English.

You will find in this post the underlying concept of file sharing. These ideas may shed light on “illegal” downloading, and highlight the limitations of the French Government’s proposal.

1) Economy of tangible versus economy of intangible

There is a fundamental difference between a tangible and an intangible economy. The first one is an economy of scarcity, the second one an economy of abundance. Sharing a tangible good means divide it. Sharing an intangible good means multiply it. Therefore, the economic laws which govern the sales of pizza are totally different from the one which govern sell of a music file.

Internet propagates the caracteristics of an intangible economy. However, the content industry is doing its best to go back to an economy of scarcity, first by using DRM (a practice increasingly abandoned), or by applying heavy penalties – even, in some circumstances, imposing jail sentences – to people who have supposedly donwloaded content illegally.

This is the opposite of were the world is heading. Analysis (like this compelling excellent article, in French, by Roberto di Cosmo), show that Internet model creates more money for the artist. We see the emergence of web sites such as sellaband whose goal is to put together artists and fans, in order to allow the latter to invest in the production of a CD, then to be later rewarded by a percentage on the sales. This is not new, Kevin Kelly’s prescient an NY Times article dated 2002, already gave some hints about where was the value shifting. Another excellent analysis, from Sloan Scholl, had shown that Long Tail (which many people confuse with Pareto law) have added another 500 MUS$ revenus to the book industry, just by letting people acess to rare books.

The world is shifting to a peer to peer model (as I have propsed here), wathever the content industry think of it. Instead of projecting themselves in the future, they complain to the government. If the French government protect them, they will die, because history shows us that over protection always has, for consequence, lack of innovation (just like the Egyptian, who refused the alphabet and prefered to increase the power of the scribes, were eventually overtaken by the Greek who decided to  built their society on top of the alphabet as a paradigm shift). How will they die ? Very simple, as new artists are digital natives, and as they understand the value of the peer to peer model as a faster mean to increase awereness of their potential audience, they avoid the old-school industry major. The catalogue of the content industry will therefore shrink, leaving us only with old stars or low quality artists. Not very sexy.

2) Network as the real issue

Internet is a peer to peer technology. It has always amazed me that very few people know the meaning of “A” from ADSL. It is very simple : Assymetrical. ADSL was invented by telecommunication operators, focused on VOD. Of course, for video viewing or downloading, more bandwidth is needed to stream the movie than sending information. For a 12+ Mb/s download, the upload is at most 1 Mb/s.

But, again, Internet is a peer to peer technology. People upload their content on flickr, Youtube, dailymotion, facebook. People write in forums, wiki, they send emails with tons of digital images. Going away from top-down vertical market, Intenet shifts the value in the peer to peer.With this vision, assymetry is a total non-sense.

We need more : as long as bandwith will remain low, usages will not take off. Back in 1997, when I was at France Telecom, Wanadoo had multiply by four teh capacity of a backbone link which was saturated. The link was saturated again 12 minutes later… Optical fiber network is the true crucial issue. No e-health, no e-environment, no e-learning can be properly done with ADSL networks. Our ancestors were able to build electricity down to everybody, the same must be done with fiber optical network.

Above this, the business model is of great importance. Any pay-per-use business model prevents the expansion of usages. In France, we have flat fee, but other countries are still on pay per use. Mobile usage is another example where not all operator undestand the value of flat fee. The roaming data tarif may become outrageous (in some cases, 10 euros per mega transfered !!!). But the telecommunication operator are reluctant to give up such models. Again, what will be the consequence of this lack of innovation in the business model ? As soon as an Internet provider, borne of the Internet culture, secures a a 3G license, s/he will install femtocell on every subscriber modem, and create a 3G network with little investment. Should a community business model (like Fon) sit atop this, then there would be real innovation, and real disruption. It would create a peer to peer based economy, which may prove more robust and more suited to the need of the citizen than any traditional vertical subscription based model.

The world is changing from a vertical flow of goods and information to a market place. Peer to peer is not to be feared, it is the future of many organisation. Rather that setting punishments, the role of government is to encourage, lubricate this dynamic into this market place, because this generates value.

Government should let the traditional economy gently, and concentrate their energies on building high bandwith symmetrical networks, both for fixed and mobile, and framing the creation of value inside this, for the benefit of citizen, and innovative companies.

Tags: , ,

The paradoxe of long tail

Dec 13

Posted by serge in business model | No Comments

I have seen many people, many post, many blogs, talking about the long tail.

It is not my purpose here to explain what the long tail is. Well, it was  not. Until I realized that many people did not understand exactly what the long tail is. To make it brief, many people confuse the long tail with Pareto distribution, or, in a broader sense, Power law. But the paradox is that long tail effect is, in fact, somehow flatening the power law.

Long tail is about something slighty different : long tail is about the size of the catalog. This is visible in the october 2004 original article by Chris Anderson, but it has been mathematically formalized in a paper published in october 2003 at MIT – Sloan School of Management, called “Consumer Surplus in the Digital Economy : Estimating the Value of Increased Product Variety at Online Booksellers“.

What is interesting in this paper ?

First, it states that the size of the Amazon catalog, when it is about books, is about 23 times the size of the Barnes and Noble catalog, and 57 times the size of an independant bookstore. This is true also for CD, DVD, etc…

Amazon has a very interesting sentence : “everyday, we sell more books that we did not sell yesterday than books that we sold yesterday”. This is the opposite of mass marketing, which tends to sell the same thing everyday.

Second very interesting idea : the pareto law still applies, but it moves from 80-20 to roughly 70-30. The distribution tail is growing. This can be seen on table 4, which shows that books after position 250.000 (which means 90% of the catalog) accounts for 29,3% of sales. The tail is flatening.

Third very interesting point: a huge catalog increases consumer spending. In this case, the paper estimates a move from 731M US$ to 1 billion US$.

This was in 2003. What about now ?

Let us make this experiment : let us type ipod in the search engine of some online and brick and mortar store.

  • bestbuy.com : 94 products
  • Amazon.com (electronic department) : 32356 products

Let us search books about ipod.

  • Barnes and Nobles : 141 items
  • Amazon.com : 13926 items.

The size of the catalog keeps on increasing. What would be the impact of this ? Huge customization, community marketing, nich marketing, but, overall, a flatening of the tail. Which would mean somehow the end of power law, the tail would no longer be a tail, the statistical distribution transformed, clusterd into category of products, of course some being more bought than others, but not that much.

How can traditional brick and mortar manage this ? This is the question.

Tags: ,

A new business indicator for enterprise 2.0

Dec 7

Posted by serge in business model | 4 Comments

Traditional business is usually a vertical one: companies buy raw material, assemble them, and create products and / or services which they deliver to customers. Even Google, one of the most recent powerful enterprise, is following this model : Google buys computers, adds an algorithm, and deliver a set of services (this is of course a simplified vision, but not a wrong one). The innovation of Google relies in its business model: the core service (the search engine) is free, while the peripheral service (advertisment) is paid. But it is not a paradigme shift.

However, there are some interesting little signals of traditional industries moving from this vertical model to another one: an enterprise who manages a market place. Here are a few exemples.

  • A traditional auctioner sells goods to people. Ebay creates a world wide community of people who trade together.
  • A traditional bank loans money coming from its suppliers to their customer. LendingClub, like many other social lending companies (Prosper, virginmoney, zopa, ppdai, dhanax, fynanz, etc..) creates a marketplace where people loan to other people.
  • A traditional major record musics from artists, and delivers it to customers. Sellaband, like many other sites (SliceThePie, spidart, indiegogo, etc…), creates a platform for people to invest into music, or film, and get revenus on the sale of the album, or the movie.
  • A traditionel telecommunication operator buys products to create a network infrastructure to sell minutes, or bandwidth, to customers. Fon creates a marketplace where people do exchange their Internet access.

The underlying business is not yet huge. The social lending market was 647M dollars in the US in 2007. Not big, but it was 269M in 2006; a very good progress. Could this move amplify ? Well, there is no reason it could not, except if traditional businesses, or the regulator (when not the two..) fights againts this; the social lending space has been recently shaked: Zopa is closed in the US, Prosper halts operations, all because of non compliance to SEC regulation. Only LendingClub resists so far.

But I still believe that this move may generalize. I recently discussed with a retail brand who sells food products in many shops over the country. The trend is there: customers want fresh products coming from less than 100 miles away. Well, what happens if, economic crisis helping, people start producing fruits, vegetables, in their own gardens. What happens if the retailers becomes an intermediate between, on one side, his customers-consumers, on the other side, his customer-producers ? He then starts a local market place…

Therefore, I would propose to work on a new economic indicator of a “Enterprise 2.0″ : the ratio of the horizontal money which flows between customers, to the revenue of the company. This indicator shows how many dollars are exchanged between customers in the marketplace for one dollar of revenue.

There is a case where this indicator can be easily computed: when a company earns a percentage of a transaction, the indicator is the reverse of the percentage. If we assume that ebay is a company which earns 2,5% as an average, the ratio is 40.

Linden Lab is another interesting company: if we assume that it generated 40 M dollars of revenu in 2007 (my guess after talking to them), and if we assume the SLifers exchanged 400 Millions dollars, the ratio is 10. Not bad.

On the other side of the scale, a traditional telecommunication operator, though delivering a service which is personal communication between people, is totally unable to generate any financial flux between his customers. I once proposed to a telco to create marketplaces where customers could exchange SMS, or even trade phone minutes. The answer was “are you crazy? We do not want to see a decrease in our revenues”. For telcos, the ratio is zero…

Interestingly, in traditional business, companies who already do trading between customers want a high percentage, therefore a low ratio.

The indicator I propose does not mean a low percentage, but rather the capacity to create a dynamic marketplace between customers.

It is a totally new approach.