Learning Center
Plans & pricing Sign in
Sign Out



Michel Gensollen

Where will the value come from on the Internet? How can money be made on a network where almost
everything is free? This article puts forward some ideas. In the short term, e-commerce will be able to
finance commercial web sites, but these sites provide only a small proportion of the information
accessible on the web. Free sites, in so far as they provide a place where consumers and producers
can exchange the information required to design new products and services, are a key component of
the Internet. The Internet will also help firms to change their structure in order to accommodate a
faster pace of innovation. IP applications should help to update individual skills and bring them
together into a coherent system of knowledge and behaviour. Finally, when in the information
economy the distinction between work and leisure is blurred, the Internet will create value either by
organizing control over intellectual property rights on information or by facilitating convergence of
information producers and consumers.

Is the Internet simply a data network? Is it a new medium or will it be tomorrow's basic infrastructure?
There are two opposing views on the subject. For some, the network has been subsidized to a large
extent and the services are all so much in the red that it is impossible to forecast any time when they
could break even. Thus, many financial analysts see current evaluation of Net businesses as a
speculative bubble rather than a reasonable bet on future profits. In contrast, the US government has
for years been arguing that the information industries are a leading sector and the mainspring of
current growth. The report published in June 1999 by the Department of Trade (The Emerging Digital
Economy II ) is consistent with this view, pointing out that information industries account for roughly
one third of all growth although they represent only 8 percent of the economy . In France many
authors similarly see the Internet as the key to economic growth; that is the philosophy of the
Programme d'action gouvernemental pour la société de l'information (Lorentz, S.D.) . Some even think
that Internet is the foundation of a new industrial revolution and set to be the driving force behind
economic growth in the next decade .

This article attempts to provide some elements of an answer as to why the Internet could be
considered as a basic facility, even though, judging by today's situation, it is no more than a medium in

  See, for example, the proceedings of a conference held in May 1999 in California (
insider/1999/0518/inv-zeromargins.html) of which the following extract sets the tone: "A panel of experts agrees: zero margins is
a dog of a business model, and a company like will be a loser in the long run. Presenting at Red Herring's own Venture
99 conference, being held this week in Lake Tahoe, California, a panel discussion entitled 'Money for nothing (and your chips for
free)' – led by noted Morgan Stanley Dean Witter (NYSE: MWD) Internet analyst Mary Meeker – considered whether free PCs,
zero-margin businesses, and other Web gimmicks hold any long-term value as a new way of doing business. With one
exception, the panelists – who included Kate Delhagen, director with Forrester Research; Eric Greenberg, chairman and
founder of Scient (Nasdaq: SCNT); Bill Lohse, CEO of SmartAge; and Josh Goldman, CEO of – quashed the
notion that any business should be operated without even a hint of profit margin, clearly putting to rest the perception that easy
access to capital within the VC community or IPO markets gives entrepreneurs the license to lose money indefinitely. Mr.
Goldman, the lone holdout for the zero-margin model, defended's tactics as a valuable brand-building exercise. 'It's
going to be more and more difficult to give away free products or services and, like Hotmail, have some sugar daddy come
along and buy you for several hundred million dollars', said Mr. Greenberg. 'I mean, c'mon, common sense still works'."
  This report can be found on the site:
  See Chapter II of the report, where it is states, for example, that: "While the share of the economy attributable to IT-producing
industries grew from 6 percent in 1993 to 8 percent in 1998 in current dollar terms, this increase understates the importance of
these industries because their prices are falling. A better way to gauge the importance of IT-producing industries is to look at
their contribution to real growth. Over the last four years, IT industries' output has contributed more than one-third to the growth
of real output for the overall economy."
  The reader is also referred to the following site:
for all documents on the subject. More generally, on major changes caused by Internet in the economics of networks, see
Curien, 1999.
  See, for example, the interview with Jean-Marie Messier on Radio-Classique on 3 July 1999: "I am convinced that the Internet
is the real industrial revolution. There is one per century, and we're in it, and it changes the conditions completely, not only of
the Internet industry itself but also of modes of organization in the business world and of our own individual modes of
organization and consumption. What does that mean? It means that maybe some Internet values are currently over-valued. But
even if Internet values were halved or divided by three, they would still be valued on a very different basis to classical

its infancy, struggling to survive on advertising (like television) and mail-order selling. What will future
source of value on the Internet be?

We focus only on the economy of sites here, not on that of networks (neither local distribution through
central offices and local loops nor transport by backbones). This choice is justified by the following
considerations: first, the share of network costs in overall costs is decreasing ; secondly, these are two
separate economies in so far as there are practically no cross-subsidies between sites and networks;
thirdly, the economic role of the Internet is based primarily on the dynamics of the sites, even if
infrastructure expenses incurred by the development of the networks are fairly substantial in
macroeconomic terms; and, lastly, at current rates the network sector in both Europe and the U.S. is
globally balanced , even if there are cross-subsidies between the various network operators (in
particular, between IAPs and operators of local loops).

In the first part of the article we attempt to analyse the current service economy and to show that,
although the situation is not as unstable as some make it out to be, the 'media' and 'e-commerce'
models are not enough to fully explain the recent craze for this means of communication. We
furthermore show that the non-commercial side of the web plays a crucial role in the system's

The second part addresses the restructuring of the commerce value chain triggered by Internet, and
defines the functioning of tools introduced by the network: constitution of the web of information sites
that Net surfers visit, and organization of surfers into communities based on electronic mail and other
tools for message exchange. Value is created at the articulation between the commercial and the
non-commercial web, based primarily on the externality existing between the two. Just as viewers do
not watch television to get information on the products advertised, so web surfers do not surf to buy.
Their decisions to purchase will be shaped by their participation in communities of which the main
object is not commercial.
The third part tries to pinpoint where IP networks could create value in firms . Since information
processed by the productive sector is of a very different nature to that consumed by households,
professional services will not be limited to the transposition of applications currently on the Internet. IP
services will therefore create value by allowing both the development of individual expertise through
the public network and the practise of collective occupations on private networks.

Finally, in the fourth part, we consider a more long-term model where the boundary could fade
between consumption and production. In many sectors and particularly in the area of information
goods, there is a willingness to pay to produce, or at least a willingness to produce for non-financial
rewards, while consumption is confronted with multiple constraints, notably those of time, attention,
etc. Value created on the Internet will be based partly on the broadcast model (diffusion of information
by large media firms which own the information-content and have a monopoly on the extraction of
value from it), and partly on a more liberal model of cooperative production of information (such as

   Costs of backbones are roughly equal to those of commercial sites. By 2003 they will account for only about one quarter.
   The equilibrium is very different in the two cases. In Europe local rates are based on connection time, which limits
consumption of the Internet but allows for economic equilibrium. In the US, because of the flat rate, consumption per surfer is
higher but local operators (RBOC in particular) are in deficit since IAPs (Internet Access Providers) do not have to pay them
access charges. If traffic towards the Internet were subjected to the same rules as long-distance traffic, IAPs would have to
double their rates. We can therefore consider that in the US local operators subsidize traffic on the Internet. In Europe the
development of freenets is a move in that direction, since IAPs aim at breaking even by the retrocession of a share of local
operators' margins on local traffic towards the Internet.
   Internet Access Providers are operators who provide access to the Internet. They basically provide a (dynamic) address and
carry traffic towards the Internet backbones. Some IAP also provide services (email, hosting personal pages, etc.). In such
cases they are called Internet Service Providers (ISP). In the rest of this article IAP is used to refer only to the provision of
access to Internet.
   For this part, the reader is referred to Benghozi and Cohendet, 1999, chapter II. For parts 1 and 2 of the present article, the
reader is referred to Dang Nguyen, 1999, chapter III.
    The information ('fourth') sector consists of information-producing activities (texts, images, software, etc.) used by firms and
consumed by households, and of activities producing the means to process, store, diffuse, transport, etc. that information. We
shall see further on in this article that information is neither an entirely ordinary good (it is a public good of which the
reproduction and diffusion costs become almost nil) nor entirely a service – even though the sale of information is similar to the
provision of the service of informing someone without that person being able to transform the information to suit his/her needs.
That is why information is classified neither in the secondary goods sector nor in the tertiary service sector.


It is difficult to understand the Internet from an economic point of view in so far as a large number of
accessible sites are not commercial. That does not mean that they are not economically important, if
only because of the very strong externality that exists between these 'free' sites and the rest of the
network. This mass, which is not included in measurements of commercial flows, explains both the
dynamics of the Internet and the model that it may well spread throughout the economy.

In this section we attempt first to evaluate the size of the non-commercial part of the Internet and to
ascertain what the role of this new medium really is in the communication sector. We then analyse
whether the Internet, as a medium (that is, apart from costs and income of network facilities) is
economically viable, and attempt to identify the source of value creation on the web.

The economic role of non-profit sites on the Internet

We define the commercial web as all those content sites for which information has been produced
commercially and which are financed either by subscriptions (a tiny minority of sites, except for
pornography) or by advertising (the main mode of financing today, including company's sites and
advertising banners) or by e-commerce (still in its early stages). The other sites, which make up the
non-profit part of the web, have been created on a voluntary basis and the surfers who visit them do
                                                 11                                                      12
not help to finance them in any substantial way . They consist mainly of the personal homepages of
web surfers and of university sites, public sites (public administrations, local authorities, etc; in the US
usually identifiable by names such as '.gov') and so on.

If we consider the quantity of information available on the web or the number of accessible pages, the
commercial web (content sites) represents only a tiny part of the current total: 14 percent of all pages.
On the other hand, non-profit web sites are, on average, visited less often than content sites which
often serve as entry points (portals) and hubs (search engines). While a web page is viewed over one
thousand times a year on average, content site pages are viewed almost six times more than
non-profit site pages. This means that, in terms of audience share (measured on the web in number of
pages viewed per year), content and non-profit web sites have roughly the same audience size (see
Table 1).
We have also tried to use these statistics to estimate the costs of the non-profit web if the authors of
these pages had to pay the price of producing the content used (that is, to pay themselves a salary for
the time spent creating their pages). Taking into account these fictive costs, costs of the non-profit
web account for only slightly more than half the entire web, since free pages are less polished and
expensive than commercial pages. In 2003 the non-profit web will account for no more than a quarter
of the costs of the web but will remain dominant as regards the quantity of available information.

Table 1: Free web and commercial web in 1998
Europe + North Number             of Page                 Number       of Audience/     Audience                    Cost
America        pages              in distribution         visits/year per year          distribution                distribution
               millions                                   page            Giga    pages
'Non-profit'       374                86%                 682             255           51%                         55%

   In most cases these voluntary sites have no source of income. Some may earn small sums by accepting a few advertising
banners. In the case of host services such as Geocities, the banners are compulsory and serve to remunerate the host directly.
In all cases we have classified these as 'non-profit sites' because the actual income of the 'owner' of the site bears no relation to
the costs of similar content sites.
   There is a wide diversity of personal sites on the Internet, from the presentation of individuals to the provision of general
information. One finds holiday photos or collective sites devoted to the organization of a shared activity. One also finds very rich
and up-to-date data bases provided on a voluntary basis and which are, moreover, sometimes used by search engines. Most
academics in the US have a personal homepage on which their most recent work can be obtained. Some personal pages are
visited a few times a day; others tens of thousands of times.
13 The figures presented in this article are drawn from a combination of diverse sources of statistics, all of which are somewhat
fragile. The reader is referred to the following sites, from which it is possible to form a personal opinion:,,,,,,, .
The statistics and projections used here are consistent with those of North American firms such as Forrester, Jupiter and the
Gartner Group. Their sites are:,,,, .

'Content    and 61                   14%               4027               245                49%               45%
retail' web
Total           435                  100%              1149               500                100%              100%

Thus, the non-profit part of the Internet is by no means subsidiary; surfers spend more than half their
time on it. If we take into account the fact that visits to portals and search engines are usually
motivated by access to non-profit sites, we conclude that the non-commercial part is the raison-d'être
of the Internet, its heart and indirectly the mainspring of its development. Net surfers go from non-profit
site to non-profit site, like the reader of a newspaper goes from one article to the next, skipping the
advertisements. If the Internet consisted only of 'profit sites' (i.e. content and retail sites) and lost the
externality of consumption between the profit web and the non-profit web, there would simply be no
surfers. In a sense, content sites benefit from the presence of non-profit sites just as advertising in a
newspaper benefits from the interest of the articles around it. Nobody would buy a newspaper or
watch a TV channel that consisted of nothing but advertisements.

Non-profit sites can, moreover, contribute directly to the development of content sites by serving as an
effective means of touting. Pornography, which accounts for about a quarter of the total audience on
the web , is a striking example. Only 10 percent of pornographic sites are paid for by subscription
fees. Most sites for adults provide information free of charge and are remunerated by content sites
because they promote them. Finally, totally free, voluntary sites (e.g. personal homepages) use
material from content sites for free, and refer customers back to them. The flow of information on the
web thus serves to prescribe consumption. Potential customers find what they want and can test the
quality of what they are going to buy ex ante. For information goods (and, more generally, for all
'experience' goods), this is essential.

The economic model of the pornographic industry and the sophisticated means it uses to guide web
surfers may well prefigure tomorrow's efficient web. Just as the flourishing industry of pornographic
sites is subjected to regulatory uncertainty, so too the existence of the web will depend on the rigour or
intelligence with which copyright rules are implemented in the future. A rigorous application of
copyright on information, texts, images and music would, in practical terms, stunt the development of
the non-profit web and in so doing sound the death knell of the content web. The Internet will develop
only if copyright owners realize that the direct consequences of copyright violations on the non-profit
web are likely to be offset by extra revenues on the profit web .

Role of the web in the communication sector

If we take into account the non-profit web, Internet in the U.S. already accounts for a substantial part
of the entire communication sector and that part will be multiplied by four in the next five years.

Web income has been defined in two different ways: either real income, effectively received, or total
income, the sum of this real income and fictive income that offsets the costs (also fictive because the
work concerned is voluntary) of non-profit sites. Note that we are referring here only to income of
Internet sites and not to income paid by Net surfers to telecommunications operators and access
operators (IAPs). Nor are we referring to firms' own markets, that is, their service sites for their own
intranets and extranets (internal portals, for example). Access to Internet at work, by employees, is
however included in the figures (irrespective of the purpose of such use).

These earnings have been compared to the entire income of the communication sector: advertising,
television, radio, music, press, publishing, video, cinema, video games, etc. (see Table 2).

   Cybersex (e-porn sites) accounts for 25 percent of the traffic, a third of all searches on Yahoo. One out of every four Net
surfers visits an adult-site every day. It is the main use at home (a quarter of the pages viewed) and the second most important
use at work (20 percent of the pages viewed). Ten percent of electronic trade today is pornography. These highly lucrative sites
indirectly finance other sites, for they are good customers of the web hosting providers, equipment providers (Cisco, Sun, etc.)
and portals (financing keywords). For more detailed figures, see the site:
   See in Shapiro, s.d., especially the chapter on the optimal implementation of copyrights on information (Rights management).
The recommendations in this part are synthesized as follows: "You should manage your intellectual property to maximize its
value, not to maximize its protection. The same goes for technology standards. Growing the market is usually more important
than extracting the last dime from your existing business model."

Table 2: Share of Internet in the communication sector (USA)
Internet / Communi-      1998               1999              2000               2001               2002            2003
cation sector USA
(G$ and %)
Income                   255.6              273               296.7              315.6              337.3           363.8
Total web income         11.9               20.2              30.9               43.5               60              81.7
Share of the web in      5%                 7%                10%                14%                18%             22%
the communi-
cation sector
Profit web income        5                  9.9               17.6               28                 41.9            60.8
Share of profit web      2%                 4%                6%                 9%                 12%             17%
in     communication

In the USA the Internet already accounts for five percent of the communication sector and this figure is
expected to rise to 22 percent by 2003 (only 17 percent if the non-profit web is excluded).

Figures in Table 2 correspond to:
- a tripling of the number of web surfers in five years. Thus, in Europe and North America this figure
is expected to rise from 120 million in 1998 to 382 million in 2003;
- a development of the Internet along the same lines as the current web. We have assumed that the
possible convergence of use of the Internet and television will be very slight in 2003. This does not
mean that broad band access such as cable or ADSL will not develop rapidly, but they will enhance
web surfers' comfort without triggering any major change in the way the web is used.

The above figures must therefore be considered as rough approximation. They correspond simply to
an extrapolation from current Internet use. In terms of these assumptions, the web is an emergent
medium that will occupy roughly a quarter of the communication market in five years' time.

Internet services eventually breaking even

The Internet, as seen above, is an original medium because of the important role of free access for
consumers and voluntary input by producers. It is also a medium that will eventually account for a
considerable share of all communication (in the sense of all the media). But is the Internet as a
medium profitable at least in the long run ? Will the players in this sector be able to remain
independent or will they have to rely on other firms: network operators, other fully-fledged media or
certain major distributors who are trying to expand their activities into 'virtual' distribution?

In this section we try to assess whether the current economic model of the Internet is viable, based on
the general assumption, considered above, of a web that will not converge rapidly with television
(even if, with the growth of cable access and ADSL, it does become more comfortable in terms of
bandwidth, to reach a peak bit rate of 1 Mbit/s per net surfer).

We have again divided sites into 'non-profit sites' (which fictively break even – by definition) and
'commercial sites' (or 'profit sites' which pay for their content and attempt to make profit). These
commercial sites are either 'content sites', which provide information and earn income from
subscriptions or advertising , or 'e-retailers' which live off on-line ordering (generally, payment takes
place on line and, for certain information goods, consumption may also take place on line).

Content sites are financed to a small degree by subscriptions. Today, these are practically only
pornographic sites and it is forecast that in 2003 they will still account for 70 percent of subscription
financing. Advertising is the main source of revenue of content sites. Here we have of course included
advertising from outside the Internet, which accounts for only a part of all advertising on the Net.
Advertising banners that refer users from one content site to the next correspond to financial

   In Europe and North America, the number of households linked to the Internet has risen from 51 million to 130 million. This
corresponds to 33 percent of households in the US and eight percent in Europe, in 1998. In 2003 these rates will have risen to
60 percent in the US and only 35 percent in Europe.
   Content sites will eventually be able to be financed in part by e-retail sites. By comparing the two types of business model in
our projection, we have tried to highlight the extent of the transfer required between retail sites and content sites.

exchanges between sites, which cancel out by aggregation . In these conditions, the economic
equilibrium of content sites depends to a large degree on their numbers since they are all practically
tapping into the same advertising expenditures (it was assumed here that in 2003 roughly five
percent of all advertising would be on the Internet). In Table 3 we have assumed that the number of
North American content sites would quadruple in five years (in 1998 there were about 1,000). These
sites are on the whole heavily in the red (their losses are equivalent to about 70 percent of their

Content sites draw their revenue from the economies of virtual distribution, compared to classic
distribution. With cautious assumptions of the market share of e-commerce (5 percent of all retail trade
in 2003), e-retailers are expected to show a substantial profit. They could thus finance content sites
since their profits will be greater than those sites' losses.

This type of subsidizing of content sites by e-retailers could be carried out in various ways:
- through integration between the actors managing the two types of site; thus, media-type sites would
balance their budgets by means of secondary e-commerce (just as the film industry balances film
budgets with rights on related products – a film is considered to be good when a large number of
secondary products such as video games and toys can be sold);
- by retroceding a part of the profit margin to 'business providers'. This can take the form of
remuneration for e-retailers' advertising banners on content sites, of remuneration for surfers sent to
the retail site ('click-through' payment) or direct sharing of profits (affiliation scheme).

Table 3: Economic equilibrium of web services (excl. non-profit sites)
Internet     USA    1998             1999               2000               2001               2002               2003
Income              5                9,9                17,6               28                 41,9               60,8
Content sites       2,9              4;7                7,3                10,5               14,5               19,3
of          which   1                1,4                1,9                2,7                3,7                5,1
of          which   1,9              3,4                5,3                7,8                10,8               14,2
E-retailers         2,1              5,2                10,3               17,5               27,4               41,5
Costs               6,4              13,4               23,6               33,9               45,8               60,2
Content sites       4,3              8,9                15,3               20,6               26,4               33
E-retailers         2,1              4,5                8,3                13,3               19,4               27,2
Balance             -1,4             -3,5               -6,0               -10,1              -11,9              -13,7
Content sites       -1,3             -4,1               -8,1               -10,1              -11,9              -13,7
E-retailers         0                0,7                2                  4,1                8                  14,3

Although the web does break even – contrary to what one sometimes hears on the 'casino' economy
of Internet – the situation nevertheless appears fragile.

The web's attraction is largely based on the its non-profit sites. But will it continue to develop as
assumed here? Will the invisible funding (non-application of copyright, free hosting paid by advertising
or offered by IAP, etc.) that enables it to exist today, allow it to multiply the number of its pages by
eight by 2003? If not, will the web not become a shopping centre where the only available
information will be there to promote sales? Visits to such a web would not increase as forecast here,
and would be reduced to e-commerce traffic.

Even if the web continues to benefit from the externality between non-profit sites (personal pages,
university sites, government sites, etc.) and content sites, we note on Table 3 that the 2003
equilibrium is preceded by a considerable overall deficit in 2000 and 2001, that is, when content sites
will develop without being sufficiently financed yet by e-commerce. We can therefore expect a difficult
period for content sites which will probably merge and try to rely on IAPs .

   Although the existence of advertising between sites does not change the overall equilibrium analysed here, it does impact on
the web economy. Internal advertising redistributes income from 'rich' sites to 'poor' ones, thus avoiding too much concentration
of advertising revenue on too few sites. That is the originality of the web as a medium: the audience and advertising revenue are
widely dispersed.
   Today this seems to be a very cautious estimate. In August 1999 Forrestier forecast 8% for 2004.
20 pays any site that sends it buyers, 5 percent of the sale (and up to 15 percent if there is prescription of a
particular book).
   In Table 3 the cost of the free web is multiplied by about three, for it was supposed that the production costs of one web page
were reduced in time owing to effects of learning and improved automation in the production of the HTML code.
   AOL has set an example of this type of successful merger between media functions and Internet access.

Clearly, the figures presented here are extremely fragile. They do nevertheless provide a rough
estimate and a framework of reference, both of which are indispensable. In particular, they show that
tomorrow's web is likely to break even thanks to e-commerce. We can therefore attempt an answer to
the question asked at the beginning of this article: what is the source of value today on the web?

The web economy is based on an externality between the quality of available information (due in
particular to the voluntary construction of this knowledge, in which North American universities play a
dominant role), which attracts net surfers, and the extraction of value from this audience by means of
sophisticated channelling through content sites (financed by advertising) towards e-retail sites. The
web thus appears as an immense newspaper featuring articles, classified advertisements, reader mail,
explicit advertising and editorial advertising. All this is free but is financed by the purchases induced
more or less directly through this information.
We are far from an Internet considered to be the industrial revolution of the twenty-first century. Nor
does such extrapolation explain the current increase in the share prices of web sites (e.g. Yahoo! and In the rest of this article we shall attempt to explain why the above analysis, because it
is based on current economic models (the advertising model; the mail-order model), does not enable
us to account entirely for the originality of the Internet, although it does set us on the right track. In fact
it is not only a matter of commerce but also of 'intermediation' , not only of productivity gains derived
from IP networks but also of an evolution of industrial structures allowing for more rapid innovation,
and not only of provision of free information but, perhaps, of a challenge to the distinction between
work and consumption.


The above analysis of the economic significance of commercial sites is partial. In fact the very term
e-commerce is misleading, for it does not only consist of selling on line, such as mail-order selling on
the Minitel (i.e. videotex, of which the development over the last ten years has been very modest:
today only 0.1 percent of all household purchases are made by Minitel or Audiotel ). In reality, it is an
evolution in the value chain of commerce and, more generally, a challenge to the nature of the
mechanisms which ensure mutual adaptation between new products and services on the one hand
and the tastes, needs, uses and wants of consumers, on the other. More precisely, the Internet could
be the place where the utility functions of consumers are collectively modified and where firms test
new products, particularly information products. These have to be amortized on broad markets
because of production economies of scale and, at the same time, adapted to each customer
segment .

Evolution of the commercial value chain

Estimations in the preceding paragraph were based on the usual definition of e-commerce,
characterized by a purchase and payment on-line in a virtual shop which has presented its products to
the potential customer. Customers choose the items they wish to purchase and put them in a virtual
shopping cart before checking out, that is, going to the financial transaction program (secure
transmission of credit card number, use of electronic purse, etc.).

With this type of definition, only goods of a particular kind can be bought on line, i.e. those that are
already purchased by mail, telephone, Minitel, Audiotel, etc. These include purchases:
- in which finding the best price is an essential element;
- with a low level of commitment and risk for the potential buyer;
- of items which are rare and in any case not available in neighborhood shops;
- for which demonstrations and assistance are not necessary for initial use.

   See the Department of Trade report cited above (The Emerging Digital Economy II) and Note 5 of this article.
   We use the term 'intermediation' here to denote all the services for networking and managing interaction, especially in the
context of new information technologies. See Gille, 1994.
   Distance selling accounts for 1 percent of household purchases. Over half of these purchases are by post and a third by
telephone. In ten years the Minitel and Audiotel have managed to capture barely 12 percent of distance sales, or 0.1 percent of
household purchases.
   The specific role of the Internet in the development of information economies, based on production with very large economies
of scale, is not considered here. The reader is referred to the following article: Gensollen, 1998: 197-227.

While it is true that the mail-order selling model can be applied on the web, with the added advantage
of more updated and easier to use catalogues, e-commerce is not limited to this aspect. Trade on the
Internet is characterized by more than just on-line selling . It does not even consist primarily of on-line
selling, in so far as a good ordered on the network can be paid for in a real shop. Moreover, once
buyers have made their decision, based on elements found on the web, all commercial acts, including
ordering and paying, can take place in the 'real world'. We could say that the commercial role played
by the Internet is not limited to electronic transactions; it is a new development in the commercial
value chain and the emergence of a new stage.

Consider (Figure 1) the various stages in a very simple retail value chain. There are six successive
steps, from production to customers: advertising, which provides information on brands and products;
display of the products in shops and supermarkets, allowing for comparison and providing customers
with advice (if only by the choice of the products being displayed); the order itself; payment; the
logistics of fetching the product or having it delivered; and, finally, after-sales service.

Figure 1 : Evolution of the trade value chain
                   Ordinary commerce                                                      E-commerce
                        Customer                                                           Customer
                    After-sales service                                                After-sales service
                         Logistics                                                          Logistics
                         Payment                                                            Payment
                           Order                                                              Order
              Presentation and reference list                                           Intermediation
          Information on brands and Advertising                                           Advertising
                       Production                                                          Production

In this type of value chain the decision to buy is based jointly on brand advertising promoting new
products and on the way in which the retailer lists and presents the products, possibly providing
elements of comparison or results of technical tests. Thus, producers and distributors compete to
influence the customer and benefit from the value created at this crucial stage. This competition
between distributors, who are in contact with the public, and producers, who define the products,
tends to impede the flow of information from customers to producers. The latter are consequently
forced to use expensive customer analysis techniques to define marketing strategies and market their
future products.

The commercial function of the Internet, at least as it appears today, consists precisely in informing
customers, recording their reactions, compiling data bases containing statistic or personal information,
and prompting customers to decide to buy. In a sense, e-retailers fit in between brand advertising and
the shop where one can order and pay. In so doing, they are in a position to recover the value
represented by product prescription to customers and marketing advice to producers. Intermediation,
which consists in adapting customers to products (decision to buy) and products to customers
(marketing) can be performed on the Internet without apparently affecting the other elements of the
value chain. Brand advertising will have less impact, supermarkets will lose part of their ability to
prescribe products and the reality of commerce will, in the end, take place on the web. However, the
value thus created still needs to be transformed into profits. That is what the firms that are trying to
create a clientele from the web are trying to do. As a result, their share prices are rising;
understandably so, for they are thus set to capture a large share of the value created by the
functioning of markets.

Furthermore, development of the Internet reduces some of the competitive advantages of large
distributors. The web enables small distributors not only to reach a wider clientele but also to group
together into communities and thus to have access to programmes for the development of customer
loyalty that were previously inaccessible. Intermediators , portals in particular, already offer

   Over classic distance-selling, the Internet has the advantage of allowing on-line consumption of certain information goods,
e.g. downloading of software or music (MP3), etc. Attributing a value to these information goods raises specific questions
addressed further on in this article.
   Quick adjustment between products and customers becomes essential with the fast pace of innovation. Such adjustment is
now possible via the Internet owing to the precision of data collected and the efficiency of data mining techniques.
   See, for example, the programme "MyPoints" by Intellipost
(, "ClickRewards" (, or the Yahoo service "ClickRewards". For its
subscribers AOL has also introduced a 'rewards program'. In all cases, the portal wanting to secure customer loyalty monitors

distributors this type of global couponing service where points and vouchers valid on all sites are

Of course major distributors such as certain leading brands will try to make their presence on the web
more visible. Starting by simply transposing their current activities (institutional sites and brand
advertising for producers and virtual shops for distributors), they will try to take over the next stage of
the value chain: intermediation. There is no guarantee, however, that web companies, which got going
earlier and have a clientele that, albeit transient, is already structured into communities and recorded
in data bases, will not be dangerous rivals in a field where the winner takes all. We note that they are
already starting to master the use of specific tools of intermediation: the web and e-mail.

The tools of intermediation: the web and e-mail

The networking of computers initially made it possible to exchange files by means of more or less
cumbersome procedures. This use still exists today and file transfer protocol (FTP) is currently used.
But the tools which now structure interaction on the Internet are more sophisticated. The web amounts
to more than just a practical file transfer protocol and e-mail offers more than a service for exchanging
private messages between individuals.

The web is a directed graph of pages, each containing:
- information (text, images, animated sequences, sounds) in HTML format which can be read by
means of browsers freely available on the web. Various peripheral programs exist for displaying
animated images, musical sequences, etc.
- information services on the web itself: search engines, site directories, etc. which enable users to
find their way around;
- information services on the real world: telephone directory, Yellow Pages, people search, maps, and
travel services;
- hypertext links: each site refers to a large number of pages, by means of hypertext links, so that it is
difficult and even futile to know whether one is changing to a different site and author. This fluidity
allows for the constitution of a collective discourse written by everyone for everyone;
- information acquisition procedures: each site tries to obtain information on the surfers who visit it;
they therefore propose questionnaires, interaction by e-mail, message boards, chat rooms or
instantaneous message services.

Some sites have 'easy access', either because they are proposed by the IAP or the publisher of the
browser (default page), or because they provide an essential service (directory or search engine).
These are general entrances , called portals. These hubs refer visitors to a large number of sites.
Some sites are visited more for the information they supply than for the links they propose, yet even
for these information sites, links play an essential part. One piece of information always refers to
others and has meaning only in that context. Moreover, the paths followed on the web are the service
that net surfers want most. When one is looking for something (practical information, a reference,
etc.), one never knows exactly what one is looking for. It is necessary to browse a little and be guided
by associated ideas. Search engines and especially the links proposed by the sites visited provide the
means for these hesitant searches and finally enable one to get a general idea of what one is looking
for. In fact, the web enables one less to answer questions than to be in a position to discover the exact
question one wanted to ask.

Browsing on the web thus leaves two types of track:
- the network (all the sites) registers net surfers' movements, either as a statistic or as individual
information, and can thus obtain a representation of demand (in the most general sense, i.e. what
interests people, what motivates them, what their associations of ideas are, what they want, etc.). This
representation, stored and processed in sophisticated and inter-connected data bases, can then be

customers' consumption, promotes affiliated sites, keeps a count of points, lists the sites that accept them and possibly
transforms virtual coupons directly into various types of 'real' advantage such as airlines' system of free miles.
   Audiences are widely dispersed on the web. Even general entry sites account for only a small part of overall use. Yahoo, the
largest site since its merger with Geocities, accounts for only 6.5 percent of the total audience. The second site, MSN
(Microsoft), which includes the message service Hotmail, accounts for only 3 percent of all audiences. The third, AOL after its
merger with Netscape, accounts for 2.5 percent. The eight largest sites together account for only 16.5 percent of the entire
Internet audience. These figures clearly show that the web is not a medium in the classic sense (i.e. essentially diffusing
information) but well and truly a means of communication between users.

used through data mining – for better (adjustment of supply to demand) or for worse (improper use of
personal data);
- net surfers discover what interests them in everything the web contains, that is, soon all knowledge
and cultural content. Due to the externalities of consumption of cultural goods, it will soon be as if a
work not accessible on the Internet did not exist at all (the case of information which that is valuable
precisely because it is not disclosed, such as firms' confidential information, will be considered in the
following paragraph).

There is thus a dual learning process: the surfer learns what there is on the web and the web learns
who the surfers are and what they want. It thus performs what liberal economists like Hayek or
Kirzner see as the main function of the market: rather than adjustment of existing supply to a static
demand, there is invention by trial and error of a new supply suited to a demand that the market
reveals. But the web does better than Kirzner's free market: it allows for learning that is not only global
and statistic (what the average consumer wants, what is technically possible) but also individual (what
each consumer wants and how to adjust production to each particular case). In this sense the Internet
is a sort of extended market that can be segmented ad infinitum.

Just as the web is not limited to file exchange procedures, so e-mail is not only a postal services,
similar to the one that has always been, except electronic. Two technical characteristics of e-mail
explain both its development and its function, in the web, in structuring communities :

- an e-mail message, unlike a real letter, can be sent to a large number of addressees without general
additional costs (time, money, looking up address, etc.). All e-mail software has the 'group' function
which enables the user to send e-mails to lists defined beforehand. These e-mails are therefore a type
of newsletter, since they are by nature more public than private (e-mail is used in push mode for
regular dissemination of information to vast numbers of people);
- an e-mail received can easily be forwarded to others; a message can thus be spread very easily and
reach a large number of people unknown to the transmitter, with consequences that even s/he could
not foresee . How can one talk of a private mode of communication in these conditions?

We thus see that the web and e-mail play complementary roles, based on dual organizations. Both, by
pooling information, allow the creation of communities oriented essentially towards learning and
consumption, in the case of the web, and towards the creation of a common frame of reference and
working in a group, in the case of e-mail. Moreover, the distribution of roles between the web and
e-mail is busy changing: intranet (firm's own internal networks) are used as a complement to e-mail,
which in turn is used in the push mode by web sites to lead communities. Today, forums, chats,
bulletin boards, mailing lists, newsletters, newsgroups or real-time message services (like ICQ)
represent intermediate states between the web and e-mail.

More precisely (see Figure 2), the web and e-mail:

- constitute complementary networks. The web is literally a graph of relations between pieces of
information, on which individuals 'travel'. E-mail is a web of relations between individuals linked
together by common concerns (hobbies, work, etc.), on which messages travel. These two networks
develop together over time, in relation to each other. Information on the web is never separated from
its author or the person that transmitted it because one can always correspond by e-mail with the
author of a page one has visited. Similarly, e-commerce on the web can lead to an exchange of
e-mails to prepare the decision to purchase. Finally, surfing on the web is a solitary activity in

   The Austrian school has stressed the fact that the market's main advantage was to allow the progressive discovery of the
demand by the supply, and of technical possibilities by the demand. This point has been treated, in particular, by Hayek, 1978
and Kirzner, 1985.
   It has not been possible in this article to analyse in detail the formation of communities on the Internet. Moreover, few studies
exist in this area. The reader is referred to Wellman, 1999, cited in note 34, and particularly to the introduction of this book which
shows that 'real' communities, when they are analysed in detail, are just as specialized, with ties as loose and unstable as those
of the virtual communities set up on the web. Moreover, virtual communities are part of a longstanding social trend (at least in
North America) from public to private sociability. The author notes that: "North Americans go out to be private – in streets where
no one greets each other – but they stay inside to be public – to meet their friends or relatives". The Internet fits perfectly into
the development of this sociability based in the private home.
   See, for example, the recent misadventures of a student whose email blunder found its way into e-mail boxes throughout the
world, to the point of being written about in the press (see the article in Libération, 29 May 1999: "David H., paria du village

appearance only . It is easy to send by e-mail a page one has visited, just as surfing together (for
example with ICQ) is common practice;
- in the case of the web, organize communities of consumers who, by surfing, discover what their
real need is (generation of demand) and, in the case of e-mail, organize work teams (even if this 'work'
is simply entertainment) based on a common language allowing for new routines;
- allow and accompany the fragmentation of firms into small production teams (see the following
paragraph) and the segmentation of demand up to the level of the community in which wants, tastes
and uses are defined.

Figure 2: Duality between the web and e-mail networks
                                              Web                                         e-mail
A graph                                       of information                              of individuals
of which the arcs are followed by             Individuals                                 messages
which leaves a trace                          in the files of the sites                   in mail boxes
and organizes                                 individuals into communities (clientele)    Messages into common discourse
Result                                        decisions to buy                            a common reference list
Function                                      generation of wants (individual uses)       generation of norms (collective uses)

We see that the Internet, independently of commercial sites in the strict sense of the term, that is, sites
where one orders and pays for goods or services, is the place where all commercial relations are
initiated. It is on the web that consumers discover both what they want and what is offered to them; it
is from the web that firms can offer goods and services to suit everyone. The Internet allows for such
adjustment because the web and e-mail lead to the formation of small interacting communities. These
represent the right level of articulation between a supply in progress and a demand before its critical

It may seem that this view of the future web / e-mail extrapolates unduly to the population as a whole
the behaviour of today's net surfers. These users constitute a particular group that is active and
enterprising because young, educated, rich and interested in new technologies. When today's
television viewers are linked to the Internet will they not reproduce the same passive behaviour that
they were used to with the broadcasting mode?

We note that behaviours as regards broadcast information have evolved with the proliferation of
channels. More active ways of using television are spreading (e.g. zapping, use of the VCR,
pay-per-view, etc.). Furthermore, use of the Web / e-mail, as described, can be 'active' to a greater or
lesser degree. For example, on Geocities one can be the voluntary leader of a community in the city
one has chosen, one can create a personal homepage which one updates more or less frequently, or
one can limit one's participation to gleaning a few bits of information by browsing through other
'inhabitants' ' pages. Similarly, in forums and chat rooms, the number of 'passive' participants who
read or listen but send no messages, is not insignificant. Finally, one can limit one's activity on the web
to visiting a few information pages selected from the list of bookmarks. The web thus allows for varied
modes of interaction, which facilitates progressive learning of more active behaviour.

Externality between non-profit sites and commercial sites as a source of value

The concrete analysis of Internet tools in the preceding paragraphs enables us to address the initial
question of the source of value on the Internet in more detail. In this section we discuss two classic
types of value:

- welfare (collective surplus) defined as the sum of all the differences between consumers' willingness
to pay and the costs of the products offered;

   See the paper by Wellman & Gulia, 1999. The authors discuss the nature of virtual communities formed on the Internet, their
differences compared to real communities, the intensity of the relations they spawn and the degree of cooperation they
   The pioneers of this type of marketing (sometimes called 'collective filtering') are, for example: (i) Amazon, which started with
the sale of books and now focuses its added value on purchasing advice provided by the customers themselves (readers
provide useful critiques for potential buyers); (ii) Geocities, which started with the hosting and monitoring of communities and is
now introducing selling points into each of its 'cities'.
   We are of course referring to the marginal cost. We are not dealing here with the specific questions raised by production with
high economies of scale. That is precisely the case of information goods, available on the Internet. It is therefore an important
aspect, treated in detail in the article Internet : une nouvelle économie de l'information ? – see Note 26.

- the producer's profit: in the market a price (or several, if the market can be segmented) is
established, and the value created for the producer is the difference between the price and the cost
(the profit, of which the growth prospects are reflected in the firm's value in financial markets). The
difference between willingness to pay and the price constitutes the consumer's surplus.

The Internet participates in the creation of value in four ways:

- by better segmentation of clientele the network makes it possible to increase firms' profits and
(particularly in the case of increasing returns) to serve groups of customers who would have been
excluded from a single-price market;

- through the role of intermediation played by Internet sites, firms active on the network divert and
increase the value created today by brands and distributors, by adjusting the product offering to
consumers' needs and vice-versa;

- through the creation of new markets, the Internet increases both the social value and the value
created by firms. That is the main source of value, the one which, as we have seen, characterizes the
free market, inventor of new forms and process of joint invention of new uses and the techniques
required to satisfy them;

- through a better organization of production, the Internet, intranets and extranets will enhance
productivity and consequently reduce production costs. This controversial point will be addressed in
the following section.

We have seen the concrete means – organization of the web /e-mail – through which the Internet
realizes the three stages of the value creation mentioned above. In all cases it is a matter of
constituting active communities whose main aim is clearly not to define tomorrow's products nor to
modify their functionality. The functioning of a free, dynamic and inventive market is a secondary,
invisible product of the visible activities that take place on the network. To take a simple example, one
does not surf on the web to find a cheap car (that is the visible part of e-commerce, the tip of the
iceberg); one surfs to find information or talk to someone. One goes back to a familiar site, gives
advice, interacts in a forum and finally discovers the product or service one was looking for, possibly a
car, while intermediators on the look-out realize what product or service will soon be in demand.

Hence, value on the Internet is not created primarily on content sites. On the contrary, it originates on
non-profit, voluntary sites where people can participate, interact and form ephemeral or stable
communities. Services offered by the Internet clearly constitute a sort of public facility with substantial
although indirect economic spin-offs. The Net develops the technical and social means for subsequent
value creation; trying to establish whether it will eventually create value directly is like trying to
ascertain whether commercial law creates value by attempting to identify signs of profits on the sale of
commercial law handbooks.

In so far as value is not created on the Internet in the place and at the time that it can be collected, the
very existence of the network – of which the usefulness derives essentially from the articulation
between non-profit and commercial sectors – is based on economic agents' ability to perceive their
long-term interests.

Until now, besides the fact that certain public organizations invested in making useful information
available on line and facilitating administrative procedures on the network, the Internet has also
benefited from various economic transfers and regulatory exemptions that have allowed its
development and still characterize its fragility:

- as regards network equipment (a point not considered here), the Internet has benefited from
transfers by local operators ;
- copyright rules are not rigorously applied . If they were, not only the amounts of these rights but also
the transaction costs would make it impossible for most non-profit sites to exist. More generally, the

  See Note 7 above.
  For example, the application of copyright to technical copies made in IAP caches would be fatal for the Internet if only
because of the transaction costs involved. Such regulations were nevertheless proposed in France in the autumn of 1997. The

collective discourse, consisting of the borrowing and tinkering that characterize the web today, could
not develop. Lastly, the concept of copyright itself is contradictory to the logic of hypertext links, the
very essence of the web;

- financial markets, which have proved to have amazing insight by financing net enterprises, might end
up extremely concerned about the extent of losses directly recorded during the next two years and the
current financial bubble could well pop. Spin-offs are spread throughout the economy and are not
limited to profits made on commercial sites;

- control over content could be stepped up. Today the Internet is a relatively free medium compared to
the others because of its international character. Here again, strict application of laws on the display of
'indecent' or illegal content, and of laws making all operators who allows such content responsible
(IAPs, host sites, telecom operators) would sound the death knell of the Internet. Adult sites account
                            39                                                                40
for a quarter of all traffic and laws generally have an extensive definition of pornography .

- local taxes and VAT-type taxes are not always scrupulously collected by e-commerce due to high
transaction costs, at least with current procedures. The fact remains that states cannot leave
transactions to develop beyond a certain volume if they evade taxes.


The analysis of value created by the Internet cannot be limited to the services available on the public
network. IP private networks (intranets and extranets) with varying degrees of openness to the
Internet, will play an essential role in the restructuring of the industrial fabric and the recomposition of
production processes. Yet in order to understand the role that the professional Internet can play, we
first need to characterize, from an economic point of view, the information consumed primarily by the
public at large and the information which firms use for business purposes. We then consider three
elements successively: the characteristics of 'information' as a good, depending on whether it is by
nature hedonistic (an 'experience good') or pragmatic (a 'search good'); the use of pragmatic
information by firms to develop routines and collective occupations that soon become their main
assets; and, lastly and more precisely, the role that IP networks could play in the development of
innovation and the organization of manpower suited to the rapid renewal of knowledge.

Hedonistic and pragmatic information

Information is an economic good only by way of a sort of misuse of language. It has been possible to
consider that information goods were traded in a market because information was always linked to
physical goods that were expensive to produce (books, records, tapes, etc.) or to services (cinemas,
etc.). Information can now, however, be reproduced and distributed at very low costs which bear no
relation to the costs of its creation. Laws and regulations designed to prevent breech of copyright
between publishing firms do not fit the final market. Consumers become aware that the information
they have bought cannot be used freely; they do not have the right to transform, adapt or redistribute
it. Whereas one can do what one wants to with an ordinary good, the same does not apply to
intellectual property. Even if the material medium is sold, the author remains the owner of an
immaterial right to the content. A person who has bought information can read, watch or listen to it but
cannot really use it because s/he has no control over the content.

The Internet has thus spawned the need to possess information by demonstrating the possibilities
available to everyone if they could just cross the boundaries of the 'information rental' contract. This
question concerns very different aspects, depending on whether it relates to information consumed for

prime-minister commissioned the Conseil d'Etat to conduct a "study on legal questions raised by the development of the
Internet". The final text, "Internet et les réseaux numériques" (the Internet and digital networks) was adopted by the general
assembly of the Conseil d'Etat on 2 July 1998. These regulations provide for a mechanism of remuneration on the basis of a
flat-rate, to be applied to technical copies made by IAPs on their servers. At European level, the "directive relative to copyright in
the information society" proposes that "acts of temporary reproduction with the sole aim of allowing the use of a work, with no
economic implications, are exempted from reproduction rights".
   For more details on the role of 'cybersex' see Note 14 above.
   See, for example, the U.S. Communications Decency Act of February 1996, at the following address: This law was eventually judged contrary to the first amendment. See also the
text of the law recently passed in Australia, available on:

entertainment (hedonistic information) or information used in the context of a productive activity
(pragmatic information) whether it is by a firm or an individual (e.g. purchase of a do-it-yourself guide).

Hedonistic and pragmatic information have different economic characteristics. While the value of
pragmatic information can be estimated before the purchase, this is not the case with hedonistic
information, for consumers are not fully aware ex ante of how useful the information will prove to be ex
post. Before buying it they have access only to limited information, supplied free of charge to give the
consumer an idea of the product; for example, in the case of a movie, the title, a synopsis and a few
pictures. Hedonistic information is thus an extreme case of an experience good ; since repeated
consumption is rare (in general one does not repeatedly consume the same content ), consumers
cannot progressively reduce the uncertainty related to purchases. Potential customers can be guided
in their choices only by the opinions of those who are already familiar with the information. It works
best when the utility functions of those who have already consumed are correlated with the utility
functions of those who are still hesitant.

In the past, a small number of critics had the role of giving opinions that broadly enabled consumers to
get an idea of the ex post utility of information. The Internet has enriched this procedure, for anybody
can give an opinion. Hence, a very strong positive externality, the formation of communities based on
lifestyle, and the exchange of opinions accompanied by a brief profile (users present themselves) are
characteristic of the web. Observers of the web are often surprised by the fact that the messages
exchanged seem highly personal, almost intimate, without any intention of forming lasting relations
such as those that would be formed during face-to-face interaction in the same conditions. The
intimacy thus created is required to assess the correlation of tastes, which in turn is necessary for
estimating the ex post utility of the information for a potential consumer, based on the opinion of an
actual consumer. Generally, however, neither party has the intention of creating a lasting relationship.
No matter how intimate a relationship on the Net may be, it uses the other person as a means for
estimating ex post utility more precisely, and not as an end in a context of friendship .

The willingness to pay for comments that enable people to form a sufficiently precise idea of the utility
of information yet to be consumed broadly equals the surplus of ex post consumption. Critical
commentary is often the only way of guiding potential customers towards satisfactory consumption, at
least in the case where information is so abundant or difficult to find that chance consumption is
ineffective. That is to say, the value of information could also be extracted through interaction between

Yet if the value of information stems from comments on that information, the collection of that value,
i.e. its transformation into profits, is complicated. On the Internet at least, it is a matter of
auto-production by consumers. Today, value created by these interactions contributes to the profits of
those who supply the means for this collective discourse. These means range from storage capacities
or hosting to interactive software (ICQ, cat software, etc.) and the organization of communities
(Geocities). This value, as we have seen, is not directly transferred from consumers to sites since it is
collected through the value derived from audiences (through advertising and sales).

Comments on information goods cannot be formulated without more or less explicit reference to the
content itself. It is therefore necessary to have extensive rights to fair use, particularly with regard to
sound and image. Content suppliers are often aware that they stand to gain if, by allowing long
citations, they promote the formation of a common critical discourse that sometimes borders on a sort

    Experience goods are defined as goods whose essential characteristic, for example the quality, is unknown before it is
purchased. By contrast, the use that a customer will be able to get out of a search good can be evaluated precisely ex ante. The
main source of uncertainty for the buyer lies in the conditions under which it is supplied (place where the good is available,
price, terms of sale, etc.). On the characteristics of goods (experience and search goods) and on the fact that the Internet
transforms 'experience goods' into 'search goods', in a sense, see Ward Hanson's (Stanford) lecture: "Marketing on the Internet"
    It is nevertheless true that certain content lies somewhere in-between. For example, the next episode of a TV series is not
known ex ante but all potential viewers know that it will resemble the preceding ones. We can thus consider it as the repetition
of previous consumption. Similarly, musical works are often known before they are purchased. These are intermediate cases,
between experience goods and search goods.
    This crude description of sociability on the web focuses on the most frequent relationships. Others do exist, such as support
groups, political forums, etc. Even if these communities are not intended to promote the commercialization of experience goods,
the fact remains that they do allow it and are thus able to generate value. The social advantage (or political risk) represented by
support groups, virtual lobbies or on-line political assemblies are a subject beyond the scope of this paper.

of collective consumption. They are nevertheless concerned about the current trend towards a real
model of purchase of information which, if it replaced the pseudo-rental contract model now in place,
would be characterized by:

- the right to redistribute the work or large parts thereof. Since this distribution could take place at no
cost, consumers' welfare, at least in the short term, would clearly be enhanced. The content industry
uses sophisticated time segmentation techniques for the extensive dissemination of information.
Prices decrease over time and adjust to the willingness to pay of different customer segments. For
example, consumers willing to pay a high price are channelled away from delayed consumption
because of the various existing 'club externalities' (they prefer to see a film when it is released, that is,
when everyone is talking about it). The Internet will thus challenge the current strategies of media
companies who will have to turn towards the organization of comments to benefit from the content
they possess;

- the right to modify information and distribute the work thus transformed. For example, a community
might want to adapt content to its own standards. This would undermine copyright in its moral
dimension – a point that will be discussed in the fourth part of our article, in so far as it relates to the
undermining of a distinction between consumption and production.

From the point of view of the collective welfare, one also has to take into account the fact that most
hedonistic information constitutes cultural goods. These goods are characterized by the fact that the
social utility of their consumption does not amount to the sum of individual utilities, which justifies state
intervention in the domain. In this sense, access to the Internet and the free consumption of
information on the network must be encouraged and could play a cultural role similar to that of public
libraries in the early twentieth century.

We note, finally, that the system that has progressively been set up on the Internet, primarily for
information goods, is naturally suited to selling. This includes the sale of all experience goods –
especially since their ex post utility is difficult to assess ex ante – and all cultural goods since the
comments attached to them help to enhance the utility of their consumption. These two criteria could
make it possible to identify the goods and services to which e-commerce is perfectly suited.

Figure 3: Comparison between hedonistic and pragmatic information
                                                   Hedonistic information                     Pragmatic information
Type of use                                            final consumption                        means of production
Type of good                                            experience good                             search good
Main problem to solve                             estimation of ex ante utility              use of information ex post
Source of value                                              critique                          cumulative knowledge
Externality                                            generally positive                    negative in the short term

Pragmatic information is similar to a factor of production. Scientific descriptions of phenomena,
operating modes and algorithms are the type of content that reduces uncertainty and has
unambiguous utility for whoever buys it. Provided the label is not deceptive, the utility is known ex ante
as well or as little as it is known ex post. In a sense, reading this type of information does not reduce
uncertainty on its utility. In these conditions, the critical comments needed for the consumption of
hedonistic information play an extremely limited role when it comes to pragmatic information, at least
in the case of rational buyers. Moreover, it often happens, especially in the case of rival firms, that the
utility of certain pragmatic information is as high as the access to the information is limited. Unlike
hedonistic information, for which critical comments are the base of positive externalities, pragmatic
information has an essentially negative externality. We can thus expect the Internet to develop
differently for this type of information, used primarily in the production context , and the concepts of a
general public Internet, such as portals and communities, not to be directly transposed to corporate IP
networks .

   Some pragmatic information is also part of household consumption in so far as households assemble raw materials to
'produce' certain products they consume (e.g. kits). Do-it-yourself manuals and scientific books are examples of pragmatic
information for private individuals. Some information goods occupy an intermediate position, e.g. video games of which the
content (the scenario) is an experience good while the effectiveness of the software can be assessed by the information
provided before its purchase. Video games are classified as experience goods or search goods depending on whether the
potential buyer's decision is based primarily on the scenario or on the quality of the animation.
   To be sure, some information or services exchanged between firms are more a matter of experience goods than search
goods. Trade between firms (B to B) is heterogeneous: some transactions resemble those of retail trade (B to C) while others
constitute more lasting and strategic relations, more similar to intra-corporate relations.

Since pragmatic information is a search good rather than an experience good, it should be able to be
traded on the market more easily than hedonistic information. The sale of information concerning
manufacturing processes, especially in the industrial sector, is extremely frequent and is legally
controlled by means of patents. Yet computer programs are protected by copyright, although a trend
towards patent rights is also under discussion. Economists have always considered that protection of
inventions by patents, which give firms a monopoly position, is an obstacle to competition and
impedes progress. Yet this ill is considered necessary if adequate incentives for research are to exist.
It is tempered by the relatively short duration of exclusive rights granted by patents, especially
compared to royalties (20 years, compared to 70 years after the author's death), and by the nature of
patentable inventions (manufacturing processes rather than general results).

Rapid technological progress is challenging the balance of this legal structure and making it less
effective. The length of protection, suited to the industrial world of the nineteenth century, seems
inappropriate for today's fast pace of innovation. Moreover, manufacturing processes are currently not
as separable from basic results as they were previously, so that there is a risk of protecting general
algorithms or scientific theories. Support for the constitution of monopolies hinders the functioning of
markets, while secrecy on advances in knowledge handicaps research and innovation and further
promotes the constitution of monopolies by providing in-house research with synergy that other firms
do not have.

In so far as the initial aim of the Internet was to link up public research laboratories and universities, it
is hardly surprising that this network spawned a sort of new encyclopaedism consisting of the
provision of free access to algorithms and software to promote innovation – just as the French
Encyclopaedists of the eighteenth century wished to widely disseminate all manufacturing techniques
in order to promote the economic and social development that had been impeded by closed
corporations formed around secrecy. This evolution towards 'free' information will be considered in the
fourth part.

Hence, the development of IP networks and Internet in firms has led to a contradiction: in-house IP
networks are really useful only if they are open to the outside (otherwise why transform current
networks?) but this opening constantly raises the question of the distinction between information
belonging solely to the firm (i.e. it is in the firm's interests to keep it secret) and 'scientific' information
(i.e. it is in everyone's interests to make it public).

Some firms choose a strategy of closure: they isolate their internal networks, practically prohibit their
employees from using the Internet and merely transpose into IP their former LANs. Others, especially
in the US, adopt a policy of openness and integration of their intranets in the global network. The
former argue in terms of the value of the pragmatic information exchanged; the latter consider that in a
rapidly changing economic sector the value lies not in the flow of information itself, but in the rapidly
renewed stock of knowledge. It is this point that we shall now develop in order to identify the value that
firms could extract from IP networks.

From pragmatic information to core competencies
Traditionally, at least since Coase's seminal article in 1937 , firms are considered to develop where
markets are ineffective or non-existent. If the legal or regulatory framework does not allow the normal
functioning of certain markets, the firms that internalize these markets benefit from a considerable
advantage. Thus, from the 1950s to the 1980s, conglomerates developed to recreate pseudo-markets
within the group, especially labour and capital markets. These heterogeneous multinationals
eventually broke up when administrative costs became too heavy following the liberalization of
external markets.

More generally, firms develop where markets prove to be unable to efficiently manage complex
information, for example when there is a high level of asymmetry of information (principal-agent
model), when information is badly distributed (team theory) or, lastly, when the cost of managing the

     See Coase, 1937 and Réseaux N° 54, 1992.

information is too high in a market (high transaction costs         in the case where, for example,
opportunistic behaviour and the existence of irreversible investments require complex contracts).

This view of the firm as the manager of a flow of information reaches its limits when the pace of
innovation increases and the processes it is supposed to regulate change quickly. It is then necessary
to take into account the fact that pragmatic information is meaningful and useful only when there is
already existing knowledge. If this knowledge evolves slowly, models in which the firm is a processor
of information are relevant. If, on the contrary, technological progress is rapid and challenges
industrial, commercial or financial processes, it is necessary to consider the firm as a processor of
knowledge and expertise .

Expertise, which is used to process pragmatic information, can be formalized and codified to a greater
or lesser degree. Some know-how remains tacit and can be transmitted only through long and costly
learning. A distinction must also be made between individual and collective know-how, consisting of
routines of actions and processes. In a fast-changing environment firms are characterized by their
capacity to transform tacit and personal knowledge into collective know-how. The capacity to manage
knowledge is a firm's main asset and defines its core competencies.

Knowledge management is more complex than simply processing a flow of data. The change from a
firm that processes information to one that processes knowledge (see Figure 4) calls into question the
organization of labour:

- more than in the past, firms have to obtain from their employees extended collaboration, for
knowledge is above all individual and difficult to formalize;

- due to the rapid evolution of knowledge, employees have to undergo training more often and tend to
remain with the same firm for shorter periods than in the past;

- a firm's specific know-how is difficult to imitate since that would require the transformation of
individual knowledge into collective routines. Consequently, questions of ownership of information and
secrets are less important, while employee training and their openness towards the outside become

Figure 4: Comparison of firms as processors of information and knowledge
Firm characterised by
                         control of the flow of information              construction of specific skills
Nature           of   Flow                                                        stock
Use              of   Transactions                                                 action routines
Individual limits     limited rationality and opportunism                          capacities for alertness and learning
Problems         to   asymmetry of information                                     change from tacit to codified knowledge
solve                                                                              individual and collective knowledge
                      transaction costs
                                                                                   skills that are difficult to appropriate

Strategic assets
                      information considered as secret

Internet and IP networks will be in a position to create value in so far as they will serve as tools suited
to the new industrial structures being set up:
- The Internet should allow for the creation of a category of 'knowledge workers' (also called 'symbol
analysts' ) not confined to the information sector as such (media, software, etc.), who play a crucial

   In his 1937 article, Coase defines the firm as a set of activities which are less expensive to coordinate in a hierarchy than on
a competitive market. Transaction cost analysis refined this intuition by trying to characterize the efficiency of transactions
according to whether they take place in a market or an organization. The specificity of irreversible investments (which make it
tempting for the contracting parties to try to extract rents) and the capacity to manage long-term contracts that are often
incomplete (a capacity that depends on the effectiveness of the legal system) play an essential part. See the analyses of
Williamson, 1975 and 1985.
   On this subject see Cohendet & Llerena, 1999.
   Sometimes referred to as "wetware"; see Note 50.
   On the subject of knowledge workers see Romer, 1995. This author defines hardware, software and wetware as follows: "The
total output from all of the industries that could conceivably send their products over a wire still account for a small fraction of

role in all sectors of the economy. Their on-going training will be made possible by the constant
interaction allowed by a public network such as the Internet;

- IP networks will have the role of structuring efficient collective routines based on the knowledge of
each person. This may occur either within large corporations that recreate internal pseudo-markets, or
between small independent rival firms that form a dynamic industrial fabric such as Silicon Valley

Corporate demand is therefore likely to evolve. For classic firms, the Internet represented a threat, so
their demand focused on isolation and protection. High-tech firms are more interested in software
allowing for collective learning, that is, both individual learning and the passage from individual
know-how to collective routines.

 Collective learning as a source of value

The main role of the productive system is to provide the link between the possibilities opened by new
technologies and the aspirations of end users. This process, which is extremely complex when
technological progress is rapid, can be broken down into three phases: the invention of products, the
development of collective occupations adapted to their production and, finally, the maintenance of
manpower capable of participating in the organizations set up.

The invention of products no longer consists of adapting production to existing needs or to utility
functions supposedly defined ex ante and analysed by classical marketing techniques. The fast pace
of technological progress demands that products and their uses be defined simultaneously. The
development of the productive activity of final markets thus challenges the role of intermediaries and
distributors. As we have seen, the Internet is progressively inventing a tool (the web/e-mail twosome)
aimed at allowing this elaboration of new functions of consumption and based, in particular, on the
externality between the commercial web and the non-profit sites. This is mentioned in the present
section (see Figure 5) only as a background for the description of the two other phases: the
'production of the firm', that is, the passage from individual expertise to collective routines and the
'production of manpower' or individual learning.

We note, moreover, that although it is already possible to identify the applications being set up in the
production of final markets, industrial organization has still not adjusted to the pace of innovation that
the information sector is forcing on the economy as a whole. It is therefore not surprising that the
professional applications that will create value on IP networks have still not been entirely developed.
Analysis is therefore more difficult and forecasting future sources of value more uncertain.

Figure 5: Role of IP networks in the various phases of production
Phases of production Function                      Main actors              Source of value           IP tool
Production of final New uses                       Firms/                   Free/voluntary            Web / e-mail
markets                                            Distributors/            externality
                     intermediation                Intermediators/
                         New products
Production               New products              Big corporations/          Articulation of individual Groupware
of firms                                           Work teams/                expertise and collective
                         inter-functioning         Industrial fabric of small occupations
                         New expertise

Production           of New expertise              Firms / professional Production of individual Courseware
manpower                                           associations           expertise
                         inter-training            Unions / welfare state


economic output. Relatively few knowledge workers produce information that a consumer will enjoy watching, hearing or
reading. […] The computing metaphor replaces the traditional categories of inputs (capital, raw materials, production and
non-production workers) with three broad classes of inputs: hardware, software and wetware. Hardware includes all the physical
objects used in production – capital equipment, computers, structures, raw materials, infrastructure and so on. Wetware
captures what economists call human capital and what philosophers and cognitive scientists sometimes call tacit knowledge. It
includes all the things stored in the 'wet' computer of a person's human brain". See also Zack, 1999.
   As formulated by Reich, 1991.

The second phase of the global production process, the development of collective occupations,
consists in transforming individual know-how into routine actions, that is, defining tasks which have to
be executed directly by the firm because they provide a lasting competitive advantage.

Most firms are moving from a cybernetic view of production in a stable environment (the firm as a
processor of information) to a biological view of a fabric of innovative teams in a turbulent environment
(the firm as a processor of knowledge). At the same time, because production techniques have high
economies of scale and the goods produced contain an increasing proportion of information, size
represents a competitive advantage and extended, even global, monopolies emerge. Subjected to
these contradictory demands, firms have to be managed, in a sense, in a quasi-chaotic state .
Competitive advantages do not last, and their strategies must not only change frequently but also be
defined in a flexible and adjustable way. We are thus far from the perfect static market where the
invisible hand does the job of an omniscient planner. The image is more one of adjustment and
selection. Complex winning strategies, often radically new, are generally derived from simple
organizational structures. The entrepreneur's task consequently consists more in firmly establishing
these structures than in abstractly defining long-term strategies.
Many scenarios have been developed to forecast the future of the industrial fabric of an information
economy where the pace of innovation is very fast. Depending on the case, sets of small rival firms
emerge, organized by venture capitalists, or else hegemonic firms which try to recreate the dynamism
of independent rival teams within the group. Constant industrial restructuring in the domain of new
information technologies provide examples of such developments .

Networks and IP services will create value, either in the framework of intranets for large firms or in that
of extranets and Internet for small firms, by providing them with networks, software and procedures for
transferring frequently updated individual expertise to collective practices, thus facilitating innovative
production. This offer does not exist today and the groupware currently used probably gives only a
very vague idea, if only because it was developed for 'information-processing' types of firm and
therefore emphasize control of information and supervision of the participants.

The third phase of the global production process, the production of manpower, concerns the
organization of a class of knowledge workers capable of rapidly acquiring the know-how necessary to
manipulate new techniques. The constitution of this capital of skills and knowledge is at the root of a
twofold evolution: first, relations between firms and their employees are becoming looser and,
secondly, training, which is playing an increasingly important part, can no longer be limited to the early
stages of life.

In so far as firms no longer manage their employees' careers, workers, who need some economic
security (continuity of salary during periods of unemployment) and social security (continuity of good
social coverage), try to reconstitute an environment in which they are able to benefit from recognition
of their skills and supervision of their on-going education. Frequent job changes necessitate the
provision of placement services and information on the job market. Furthermore, knowledge workers
need lasting working relations with experts in the same branch as themselves, that is, a place for
identification and social interaction between peers. Organizations centred on occupations, such as
professional associations or guilds , might therefore develop as a substitute for the social

   See Brown and Eisenhardt (s.d.). Based on concrete examples, these authors try to define what the rules of strategy might
be for these firms which, like life itself, are bordering on chaos: "Organization drives strategy, not vice versa. […] The most
effective firms are structured at the 'edge of chaos' where a few simple structures create complicated behaviours, there are
surprises and mistakes, and managers must work to stay there. The edge of chaos is more than a bland mid-point on a
continuum between organic and bureaucratic structures. The structures are not arbitrary. Staying on the edge of chaos pushes
out the efficient frontier for the trade-off between innovation and efficiency."
   Thus, the two scenarios on twentieth century organizations developed by the "MIT Scenario Working Group", have the merit
of describing two types of organization in detail: 'small enterprises' and 'virtual countries'. See Laubacher, Malone et al., 1997.
   For example the different buy-outs and mergers of firms such as Yahoo, AOL, Netscape, Geocities, Microsoft, Excite, etc.
   This type of organization is described in Laubacher & Malone, 1997 (see note 53). The authors describe what these guilds
might be: "The widespread adoption of flexible employment practices is potentially troubling, because the current system is set
up to accommodate full-time workers. Non-traditional work arrangements typically do not include important features which come
with a 'regular( full-time job: (i) economic security, assured through a reasonable expectation of employment continuity and
company-subsidized insurance and pension plans; (ii) career development, through access to internal career ladders and
company-supported professional networks, training courses and counsellors, and the informal professional accreditation which
service at an established firm confers; (iii) social connection, through interaction with co-workers, and the sense of identity which
comes with a position in a firm. These were all needs which have traditionally been provided for by employers within the context

responsibilities of big companies and the state, and allow the innovation process to benefit from
qualified manpower.

Because training must be on-going, at least for knowledge workers, the tools available on IP networks
and that we have put under the generic term 'courseware' should facilitate effective diffusion of
expertise. But they will also have to constitute mediums for the various functions of human resource
management that big companies and governments have progressively abandoned: certification of
knowledge, social solidarity, management of labour market, etc. We thus see that these applications
will constitute not only learning programmes but also tools for knowledge management, employability
and labour markets.

Faced with the development of this partial solidarity, since it is limited to a particular occupation, the
state will have to provide overall solidarity, in more difficult conditions, in its area of competence. In
this sense, IP networks will have to provide coherence between applications that are a priori very
different but have the same aim of manpower production: training, distribution of work, access to
medical treatment and social solidarity.


We have successively considered the development of IP networks and of the structures of
consumption and production they accompany or generate, with a view to defining how or where value
on the Internet is created. Some current developments suggest that in the longer term, as the
information sector develops, the two movements described previously might converge. In other words,
there might be a merging between the communities or consumers who participate, without always
being aware of it, in firms' marketing efforts, and small or even micro enterprises that produce the
innovations needed to make new products.

This evolution, which is natural in so far as it perpetuates the current trends towards the expansion of
the service sector into a sort of fourth sector for information services and goods, is accelerated by the
reduction of time devoted to work and the progressive confusion between leisure and productive

Because the distinction between work and leisure is fading, the model of the large capitalist firm is
becoming less relevant. As a result, the liberal model of coordination by the market rather than by
hierarchies is taking over. It is, moreover, in the name of freedom that the first examples of
convergence between production and consumption were developed with the distribution of freeware
as opposed to the dissemination of information protected by patents or copyright. In this type of
economy, value is created by networks since they are capable of providing all the services allowing
effective interaction between individuals who are at once producers and consumers.

Challenging the leisure / work distinction

The distinction between leisure and work as two distinct periods in life was born during the industrial
revolution even if initially some people had very little leisure and others very little work. That pleasure
cannot be derived from work and that it is necessary to consume in order not to be bored seems
obvious to everyone; it is a belief that structures current social organization. The fact that it is recent is
forgotten, and few are aware of its strangeness.

In everyone's life the productive phase is characterized by a time of constraint, when one is
encouraged to cooperate in hierarchical structures and activities are subjected to strict rules to ensure
their productive efficiency. Working time is bought at a price that is not clearly dependent on its
productivity or difficulty. The capitalist firm, because it provides the means of production which
represented an extremely heavy load for the secondary sector, thus owns the production from which it
may derive profits subsequently distributed to investors.

By contrast, the leisure phase is characterized by free time occupied by the purchase and
consumption of goods, that is, by the destruction of their market value since there is practically no

of the employment relationship. With the growing adoption of flexible employment practices, there will be a greater need for new
kinds of organizations to fill the vacuum and assume these roles for workers who will no longer have an affiliation with a firm."

second-hand market between consumers. In a sense, work which creates value because it is forced to
do so is the opposite of free time during which that value is destroyed for pleasure. Everyone's
production time buys wasting time.

This schema, more suited to an economy where the primary and secondary sectors are dominant, is
being undermined by a twofold change: first, at least in the information sector, it is becoming possible
to produce outside the framework of the firm; secondly, the schema that opposes costly production
and the willingness to pay in order to consume, corresponds less and less to reality.

In the information sector production outside the usual framework of the firm is made possible both by
the reduction in costs of equipment and by the increase in leisure time:

- the reduction of costs (and prices) of equipment needed to produce, enables small firms and even
individuals to engage in a productive process. Whereas substantial capital was required to build and
use a blast furnace, the tools required by the information economy have very low costs and, above all,
marginal costs that are almost nil (microprocessors, software, etc.), provided that these tools are not
protected by copyright or patents. The only really costly investment remains the know-how
accumulated by each person;

- the increase in the leisure time of individuals capable of producing provides the manpower needed
for the development of 'leisure-work'. Today, the fact that young people enter working life later due to
an extended period of studies or training and that older people retire earlier, leaves everyone long
periods in which they are able to produce outside the context of a firm. Finally, with the substantial
increase in mass unemployment and a third of the country's manpower idle, the U.S. – which has the
lowest unemployment rate – multiplies incentives to the development of a parallel economy not
based on the work/leisure schema.

Furthermore, the opposition between production and consumption is progressively being undermined,
in practices and social representations alike:

- first, the development on the Internet of effective second-hand markets, for example by means of
auctions, reduces the phase of destruction of the market value of consumer goods. At the same time,
consumers are in a position if not of co-producers (although some second-hand goods may have been
'improved' by possession) then at least of actors in the trading process. In any case, consumption and
destruction no longer necessarily go hand in hand;

- the tertiarization of the economy, that is, the migration of value into the services accompanying
goods (maintenance, rental, insurance, etc.) often leads to a confrontation between the supplier's
work-time and the buyer's leisure-time. When this exchange was mediated by the production and
purchase of a good, the model appeared less clearly; when it is at the same time – so to speak
face-to-face – that the producer and the buyer of the service cooperate in the market act, the arbitrary
nature of the distribution of roles is likely to appear;

- the development of computer technology allows active leisure activities at a low cost, that are less
and less distinct from production. Thus researchers at the MIT Media Lab are preparing intermediate
musical instruments between classic instruments and records so that everyone can have access to
the degree of creativity that suits them;

- the development of consumption that is free or even for which the consumer is paid, can sometimes
be effective for goods distributed at zero marginal cost, like information goods. It remains, however,
that the fact of being paid to consume, as in the case of the net surfer who receives a salary
   The real unemployment rate is well above the official rate. In the U.S., the official rate in 1995 was 5.7 percent but if we add
those individuals who would like to work without being statistically defined as jobless, we would obtain a figure around 14
percent. Finally, if we take into account the fact that a large number of 'self-employed' are in fact people with neither work nor
income and that a large number of individuals of working-age are not included in manpower statistics, we obtain a rate well over
30 percent. See the details of this estimation in the section 'the real unemployment rate' in the article of Thurow, 1996.
   See, for example, Gershenfield, 1999, chapter two (Digital Expression). The author is explicit about the creative possibilities
offered by 'virtual instruments': "It used to be that many people played music, because that was the only way to hear it. When
mass media came along, society split into a small number of people paid to be artistically creative and a much larger number
that passively consumes their output. Reducing the effort to learn to play an instrument, and opening up the space between a
PC, a CD player and a Stradivarius, points to the possibility that far more people will be able to creatively express themselves.
Improving the technology for making music can help engage instead of insulate people."

proportional to the time spent surfing, helps to muddle the contradictory concepts of paid consumption
and wage-earning production;

- finally, the willingness to pay to perform a productive activity is demonstrated with the development
of leisure activities. Is it not paradoxical that airline pilots are paid to fly when there is a strong
willingness to pay for the right to fly during leisure hours? Why pay royalties to authors when they are
prepared to pay so that we can read their work? Why pay a professional football player when playing
football is, for some, a pleasure? In all these cases and many others, the schema of
production-consumption is maintained by a complex social construction: professionalism must be held
up as a value and 'amateur' work disqualified, prohibited for the sake of quality, safety or insurance.

The growing uneasiness about mass unemployment or the 'entertainment society', for example, stems
from the impression that these dysfunctions could be reduced if they were not required to maintain the
current organization of production and consumption , an organization that is less and less suited to
an information economy.

Liberal network or capitalist diffusion

The eighteenth century industrial revolution was triggered by two major innovations that are
sometimes confused under the vague term 'liberalism': the liberal doctrine in the strict sense of the
term, and the large capitalist firm. In the rest of this paper we shall reserve the term 'liberalism' to
Mandeville's initial intuition , taken up by Smith, according to which it is futile to inculcate people with
moral rules of cooperation and to develop complex regulations to guarantee the smooth functioning of
society. Leaving the selfishness and opportunism of each person free reign can, under certain
conditions, lead to the optimum situation or at least to a better situation than that obtained by carefully
controlling everyone. Yet economists have not completed the liberal programme : establishing the
general political rules that allow laissez-faire (characterized by the fact that the state does not
intervene in private affairs other than to enforce the law) leads to a situation that is not too bad and is
acceptable by all.

Critical theories of industrial development have attacked both innovations equally, rejecting liberalism
while the focus of their criticism was essentially exploitation of labour by large capitalist firms.

In fact, big firms have always struggled against the constraints that liberalism placed on their
development, the expansion of their markets, and their striving for a dominant position to increase
their weight in a competitive game. This constant tension is particularly marked in the U.S. where the
liberal tradition has always been stronger. Anti-trust laws are the main armistice in this long war; for
example, the 1890 Sherman Act and the 1914 Clayton Act . Today, media groups and software
producers such as Microsoft have replaced the railways in anti-trust lawsuits.

With the development of the information economy, that is, the increase in the value of software
development and the realization of information content, big companies are in crisis. The progressive
confusion between work and leisure is undermining their existence, while the economies of scale of
the production and processing of information promote the formation of monopolies.

   The strange impression that society needs unemployment to reproduce itself as it is has been commented on in Vivian
Forrester's book L'horreur économique (1996). The author strongly criticizes the fact that society, instead of reducing working
hours and even preparing a distribution of wealth that is not based on wages, clings to employment as the only source of social
integration and distribution of goods. She denounces the current grotesque situation: on the one hand those who are so
over-worked they do not have the time to consume their income and to whom it is necessary to supply very expensive goods
and services, and on the other hand jobless or badly employed people who do not have enough to live on and are excluded
from society and prevented from working (even on a voluntary basis) while looking for a job. In a sense, the scandal of
unemployment, is that it is necessary in order to maintain the production /consumption model, just as sport needs to be
transformed into a show to distinguish the professional from the amateur, and just as employment needs to be practically
inaccessible outside big firms, etc.
   The title of the book by Bernard Mandeville (1714) sums up the liberal programme: "The fable of the bees or the rogues who
became honest men. With the commentary, where it is proved that the vices of individual people tend to benefit the public".
   In this respect, current research focuses on modelling incomplete contracts. See the introduction to the recent article by
Tirole, 1999, p.741-781: "Indeed, I would argue that the difficulties encountered in conceptualizing and modelling incomplete
contracting partly explain why the normative agenda of the eighteenth century political scientists – namely addressing the
question of how one should structure political institutions – has made little progress in the last two centuries".
   On North American anti-trust legislation, see for example:

Data processing, the media and telecommunications networks, which account for an increasing
proportion of value creation for all sectors, are thus set to be the battlefield between big companies in
the process of integration, and a more liberal type of organization suited to information economies.
North American liberals (like those of the Electronic Frontier Foundation) have pointed out the
interests at stake and above all the threat to freedom of speech and even the development of
knowledge, implicit in the extension of property rights .
How could mathematics have developed if theorems had been patentable ? How could we speak if
we had to pay for the right to use every word in the dictionary? These examples may seem ridiculous,
yet discussions are underway on the extension of patents to algorithms used in software , and
expressions as commonplace as 'Olympic Games' are protected by copyright in the U.S. .

The big corporation in the information sector, which creates or transforms information (content,
software) or provides the technical means for information processing, could not develop towards a
dominant position unless:

- it is able to create and defend extended property rights on the information produced, the strict
application of copyright, patents on algorithms, and processes of marking information for this purpose;

- it is able to appropriate the intellectual creation of its employees; in this respect, the French droit
d'auteur is less unfavourable to the interests of employees than North American copyright (with the
exception, however, of software);

- it is able to prohibit consumers from free use of information, that is, to prohibit the reuse of
information. Now, creation feeds on interpretation, reuse and comparison; how could one create a web
page if one had to find and pay all the authors of the content used?

- it is able to promote techniques of asymmetric dissemination, to the detriment of techniques of
symmetrical exchange. Technological progress is a threat to information corporations in so far as
innovations under way all tend towards greater interactivity and more symmetrical relations (TV
broadcasting is being enriched with a feedback channel, so that TV programmes will be able to be
seen, recorded and edited on a microcomputer; the broad band web will allow the exchange of
animated images, and so on).

Tension between two schemes of organization in the information sector – one which favours property
rights and the other free use – is already perceptible in the diffusion of content (e.g. the role of the
Internet in the diffusion of music in MP3) and especially the design and dissemination of software. The

   See, in particular, the often cited text by Barlow, 1993. The following extract highlights what has just been said: "Furthermore,
the increasing difficulty of enforcing existing copyright and patent laws is already placing in peril the ultimate source of
intellectual property, the free exchange of ideas. That is, when the primary articles of commerce in a society look so much like
speech as to be indistinguishable from it, and when the traditional methods of protecting their ownership have become
ineffectual, attempting to fix the problem with broader and more vigorous enforcement will inevitably threaten freedom of
speech. The greatest constraint on your future liberties may come not from government but from corporate legal departments
laboring to protect by force what can no longer be protected by practical efficiency or general social consent."
   On this subject see Callon, 1999.
   On the project to extend patent legislation to software, see the site and for the opinions of free
software developers on this point, see: thegnuproject.html. For example, at this address: "The worst
threat we face comes from software patents, which can put algorithms and features off limits to free software for up to twenty
years. The LZW compression algorithm patents were applied for in 1983, and we still cannot release free software to produce
proper compressed GIFs. In 1998, a free program to produce MP3 compressed audio was removed from distribution under
threat of a patent suit. There are ways to cope with patents: we can search for evidence that a patent is invalid, and we can look
for alternative ways to do a job. But each of these methods works only sometimes; when both fail, a patent may force all free
software to lack some feature that users want."
   See the Supreme Court Judgement, 1996. On these questions of intellectual property rights on the Internet, see also the
papers of Pamela Samuelson at:
   We are not considering the case here where information corporations 'tertiarize the information sector', that is to say, manage
to develop a business model in which most of the value is derived from services (and possibly non-information goods) related to
the content, software and information they produce. This indirect attribution of value to information helps to delay, at least for
some time, the effective establishment of an information economy model in which value shifts not only towards services related
to content but also towards the provision of the means required for the consumption/production of information. On the shift of
value towards services related to the consumption of information, see Gensollen, 1998. See also Shapiro & Varian (s.d.) for
examples of the indirect attribution of value to information.

GNU project for the writing of an operating system and 'free' software is explicitly aimed at social
organization focused on exchange rather than diffusion.
Free software is not a free-of-charge program ; it is characterized by the fact that the user has free
access to the source code and is therefore free to modify, enrich or redistribute it so that everyone can
benefit from everyone else's work. Free software moreover uses copyright rules at first to prevent the
successive authors from developing protected versions (a technique called copyleft ). It is, in a sense,
a paradoxical use of copyright which is used to limit property rights on information and not to enforce
them. The writing of 'free' software is less the result of a rational economic choice (in the sense that
non-cooperation is rational in a prisoner's dilemma) than of a rule of conduct that would lead to the
optimum outcome if everyone complied with it .

The diffusion / cooperation dialectic as a source of value

In the preceding paragraphs we attempted to define two opposing schemes for imposing an
information economy, schemes which we could, to be brief, call 'liberal' and 'capitalist' information
economies, respectively. Just as in the event of war value develops with arms dealers, so value on the
Internet will form wherever the network is able to provide everyone with the tools they need to impose
their view of the development of the information economy.

It is therefore a matter, on the one hand, of providing big companies with the means to develop their
own property rights on information and, on the other, to provide content creators and consumers (often
considered as the same people) with the technological means to bend restrictive rules until they have
been dismantled. We note, moreover, that for historical reasons the Internet is a 'free' network in the
sense of free software; no corporation is responsible for running it, programs for using the web have
always been provided free-of-charge to users and their source code is often available. In many areas
the Internet sets the example of spontaneous coordination and cooperation not imposed by a
hierarchy. This type of behaviour makes economists say that it has still not reached maturity, that it is
in a pre-economic phase.

We saw above that the liberal view and the project of big information companies are opposed on three
main points:

- the distinction between creation and consumption of information. Big firms seek a monopoly position
on the creation of content and wish to confine their customers to the role of passive
consumers-disposers of information. By contrast, the development of tools, equipment and software,
which allows for sophisticated information processing by individuals, is in keeping with the liberal
model. The manufacturers of terminals are primarily interested in this type of development . The

    See the site: ; the aim of the project is defined as follows: "The GNU Project started in 1984 to develop a
complete free Unix-like operating system. Variants of the GNU system, using Linux as the kernel, are now widely used; though
often called 'Linux', they are more accurately called GNU/Linux systems. The first test release of 'the' GNU system, using the
GNU Hurd as the kernel, was made in August 1996."
    See Stallman (s.d.): "The term 'free software' is sometimes misunderstood – it has nothing to do with price. It is about
freedom. Here, therefore, is the definition of free software: a program is free software, for you, a particular user, if: (i) you have
the freedom to run the program, for any purpose; (ii) you have the freedom to modify the program to suit your needs (to make
this freedom effective in practice, you must have access to the source code, since making changes in a program without having
the source code is exceedingly difficult); (iii) you have the freedom to redistribute copies, either gratis or for a fee; (iv) you have
the freedom to distribute modified versions of the program, so that the community can benefit from your improvements. Since
'free' refers to freedom, not to price, there is no contradiction between selling copies and free software."
    "Copyleft uses copyright law, but flips it over to serve the opposite of its usual purpose: instead of a means of privatizing
software, it becomes a means of keeping software free. The central idea of copyleft is that we give everyone permission to run
the program, copy the program, modify the program, and distribute modified versions – but not permission to restrictions of their
own. Thus, the crucial freedoms that define 'free software' are guaranteed to everyone who has a copy; they become
inalienable rights."
    This type of cooperative choice is presented as the choice of a freer society and thus fits into the liberal tradition. This is clear,
for example, in the following extract from Richard Stallman in The GNU Project: "The easy choice was to join the proprietary
software world, signing nondisclosure agreements and promising not to help my fellow hacker. Most likely I would also be
developing software that was released under nondisclosure agreements, thus adding to the pressure on other people to betray
their fellows too. I could have made money this way, and perhaps amused myself writing code. But I knew that at the end of my
career, I would look back on years of building walls to divide people, and feel I had spent my life making the world a worse
[ gnu/thegnuproject.html]
    For example, in the past it was in the interests of manufacturers of radio-cassette players to introduce machines on which
audio tapes could easily be copied; this, however, was contrary to the interests of record and pre-recorded tape companies.

reduction of data processing equipment costs makes it possible, for the same price, to extend the
performance of terminal equipment towards network processing capacities. Similarly, the provision by
networks of various intermediation services is congruent with the liberal scheme, whether these
services consist of selected directories, reputations management , insurance, trusted third party, etc.

- the attribution of value to information. Big information companies try to give value to content through
the broadcast mode, since information is protected against copies and especially against attempts to
reuse it. The liberal model is characterized, by contrast, by the free construction of a common
discourse. All software or hardware that prohibits copying or makes it possible to mark information for
subsequent control strengthens the position of major information companies. This type of restrictive
policy, if implemented, would considerably reduce information flows since current use of the Internet
would be profoundly undermined. Telecommunications firms, whose profits today depend on the
volume of information transported, would then have to extend their activities towards information
control. We thus see that big information companies have a strategic interest in moving towards
network operators. We also note that the increase in the information flow, especially in the case of a
high level of asymmetry (cable networks and, to a lesser extent, ADSL) favour the dissemination mode
and the model of large information corporation;

- control of phases of elaboration of information by knowledge workers. Today, apart from certain
small companies which form an industrial ecosystem, like Silicon Valley, where content and software
benefit their authors, most big information firms try to control the information exchanged both in-house
and with the outside , to ensure control over the content at the time of its elaboration. Of course, this
control is not incompatible with a policy of profit-sharing with employees, in the form of shares (as in
the case of Microsoft). Big firms will therefore ask for software (e.g. groupware) which allows very
close supervision of the work carried out, while certain small, more open, firms will ask for
sophisticated screening software between intranets, close entities linked by extranets and the outside

The role of networks and the way in which value is created on the Internet will depend on whether the
final organization of the information sector is closer to one or the other of these two models. The two
possible outcomes are very different (see Figure 6).

Figure 6: Comparison for information economies of the
"Capitalist" and "Liberal" models
                                Broadcast model                                                     Liberal model
Industrial structure            Global oligopolies                                                  Small rival firms
Costs of equipment              Expensive production equipment (patenting)                          Inexpensive           production
                                                                                                    equipment (p=Cm)
Nature of trading               Dissemination of protected information                              Trading and enrichment of
Control of information          Information belongs to firms (copyright)                            Free     intellectual   creation
Property rights                 Copyright, patent                                                   Free software (information)
Status of work                  Leisure / work contrast                                             Confusion leisure = work
                                Professional / amateur contrast                                     Confusion         producer     =
Strategic advantage             Media/software firms have an advantage                              Hardware manufacturers have
                                                                                                    an advantage
Role                   of Profits on access control                                                 Profits on traffic
telecommunications firms Role of policing                                                           Role of intermediator
Internet                  Mass media (transfer of value from TV)                                    Virtual       consumption       /
                                                                                                    production area (development

   Thus, on auction sale sites the background of the activities of each buyer and seller can be consulted to 'know who one's
dealing with' before buying or selling.
   For an example of the way in which firms can justify control over information used by their employees, by hiding behind the
risk of legal action for sexual harassment or violation of copyright, the reader is referred to the book by Overly, 1999, from which
the following extract is drawn: "Circulating inappropriate jokes via e-mail or downloading pornography from the Internet can
result in sexual harassment suits, discrimination suits, and copyright violations. […] The best policy is to monitor employees'
computer use. The company's written policy should make it clear that employee messages and files could be reviewed at any
time, without notice. The policy should expressly state that employees shouldn't expect anything they create, store, send, or
receive on their computers to be private. […] The average employee spends three hours a week surfing the Net for
non-business reasons. In addition to wasting a company's time and money, this non-business surfing can make a company
vulnerable to copyright violations. […] Publicly posting summaries of employees' Internet activity – what sites each person
visited and how long he or she spent there – can be an effective way to discourage unnecessary Net surfing."

                                                                                of value of intermediation)

In the model of diffusion by the big information firm, value is linked to content and its consumption. In
this scenario the web will evolve towards a medium whose economic model will perpetuate that of
television, characterized by the consumption of identical information by a large number of passive
consumers. The effectiveness of this model derives from the fact that the cost of copies and
dissemination of information is almost nil. But that is also its weakness, for media firms have to use
legal and technical means, including close surveillance, to fight against unauthorized copies, while
such Malthusian actions are seen running counter to a maximization of the collective welfare.

In the case of the liberal model, where each consumer is free to use and enrich information received
and thus to constitute different content, value is linked to the means available to active consumers to
use this information. It is thus created on networks from sophisticated terminal equipment and by the
provision of software for the acquisition, processing and exchange of information. Thus, big media
firms and equipment producers have opposing strategic interests, while network managers are in an
ambiguous position. In the liberal model their profits will derive from traffic and networking services,
while in the broadcast model it will come from the technical control of traffic in order to enforce
property rights on the information transported.


In this article we have tried to answer the question: 'What is the source of value on the Internet?',
without attempting, at this stage, to formulate the strategies that particular actors ought to implement
to transform that value into profits.

In the event of relatively slow development of the Internet, and if production processes and modes of
consumption remain the same, value will be based on e-commerce. The Internet will become a highly
effective means of mail-order selling. Content sites will become less important and will be limited to
free information (university, public, institution sites, etc.) and to some sites financed by advertising,
along the same lines as the press. A share of the value created will thus derive from that sector. In this
scenario the web economy will break even in three or four years.

If the Internet continues to develop in all countries at the explosive rate witnessed in the U.S. – which
would probably require that property rights on information be adapted to the web – the generator of
value will reside primarily in externality between non-profit and content sites. Consumers, encouraged
to surf the web for quality information available free of charge, will provide the information needed for
the design of new products and services as they simultaneously learn, in communities, about new
uses and new types of consumption.

The pragmatic information processed by firms is by nature very different from the hedonistic
information consumed today on the Internet. The latter type of information constitutes experience
goods, which explains the usefulness of the Web / e-mail. Pragmatic information, by contrast, is added
to individual knowledge and facilitates the development of collective occupations. Professional IP
services will therefore not be limited to the transposition of current Internet tools (e.g. in-house
portals). In order to create value, they will have to provide the means for frequent updating of
individual know-how from public networks (courseware), as well as the organization of manpower into
communities. Professional IP services will also have to accompany the development of routines
allowing for innovative production (groupware).

Finally, if the differences between work and leisure fade and consumers can enrich information and
disseminate it freely, the Internet will create value by bringing together communities of consumers and
producers. If information corporations impose a broadcast model, the Internet will become a medium
similar to television as we know it today and will harness part of the value currently created by

To top