timone NetMarketing
disponibile anche in italiano

Marketing in the internet – as seen from Italy

No. 72 – August 2, 2004

Other articles on similar subjects
are published in English
in the Offline column


loghino.gif (1071 byte) 1. Ten years ago – and now

Ten years ago I was asked to start writing about the internet. There were interviews, articles and books that were published only in Italian. A monthly column started in 1996 (it’s only partly online in English.)  In following years it was replaced by others in different magazines.

This series has been appearing, in both languages, since 1997. Only one of the other columns, Offline, is available in English (it started in 1998.)

(On a different, but not unrelated, subject The Power of Stupidity started, in English, in 1996 and continued in following years.)

Of course it would be boring for readers to re-consider old thoughts and comments if they were just the opinions of one person. But I never wrote anything without checking a variety of sources and discussing issues with all sorts of people. Including some that have depth of knowledge and competence on the subject – and others who don’t, but ask interesting questions.

For a variety of reasons I’ve been looking back at things that were published years ago – and that I had almost forgotten. In some cases they could be written today as they were then, but in others the perspective has changed.

Neither I, nor my best advisers, ever attempted forecasts or prophecies. But what we were learning from trends, as we could see them ten or five years ago, wasn’t always confirmed by what happened later.

We knew that the net was growing, but not as fast as most sources were claiming eight or ten years ago. On a global scale, and in Europe as a whole, facts are in line with what we expected. There was no exaggerated acceleration in the years of hype, nor any relevant slowdown a few years later, when there was a widespread feeling of gloom. (See international and European data.)

There were, however, some changes that didn’t fit our expectations. In some parts of the world there is, so far, slower development than we thought would come (see low density countries.)  While in some places (including Italy) there has been faster growth, in recent years, than could be “guessed” by any projection of previous trends.

We never believed in a “new economy” – or in a “network society” as separate from human culture as a whole. We were quite firm in believing that there wasn’t, and there isn’t, any “cyberspace”. But, to some extent, we were naive in expecting a faster and greater evolution of the “information age”. The potential is there, but our way of using our new information and communication resources isn’t improving as fast, or as widely, as it could.

We never believed in the repeatedly announced “death of printed paper”. However we did, for a while, expect hypertext systems on hardware supports, such as cd-roms, to provide useful alternatives in the case of complex specialized databases – but “electronic publishing” has been very disappointing (and web resources are often much more effective, if and when they are properly updated.)

We were concerned that attempts to censor, control, restrict, scrutinize and “tame” the internet would continue – and unfortunately we were right. We were quite concerned about spamming and online swindles (see Spam and scam) but we didn’t expect them to grow so enormously, with no serious efforts to bring them under control.

We were hoping that the expansion of fatware would find a limit (see The Hermes pendulum and the art of lightness) though we never made any assumptions about when it could happen. It isn’t surprising, but it’s quite disappointing, that we are still running at full speed in the wrong direction (see Less is better and The stupidity of technologies.)

We believed then, as we do now, in the importance of opensource resources. That notion is somewhat more widely perceived than it was ten years ago, but it isn’t as extensively applied than it should.

We were skeptical about financial manipulations – and things went even worse that we expected (see The “bubble” misconception.)

Sometimes we were hoping for too much – or too soon. Not in growth or size, but in quality. Fact proved us right, most of the time, when we were cautious or skeptical. We were more often wrong when we expected human improvement, evolution in culture and society. We knew that changes don’t happen just because we have the tools. But it wasn’t unreasonable to hope.

My friends and I have been discussing this for a while. We are disappointed – but we are learning from our mistakes. It wasn’t just daydreaming. Human nature and cultural evolution have ups and downs. There is no reason to stop trying to find new openings, new thoughts, new paths, new hopes. But we must realize that it’s much more complicated than “reasonable” expectations lead us to believe ten years ago.

The future is unpredictable. But we can learn by following trends and developments, keeping our eyes open for those small signals that often are more relevant than the “big picture” as painted every day by the big homogenized information system.

back to top


loghino.gif (1071 byte) 2. China – and other problems

On July 26, 2004, Reporters sans frontières issued a statement about censorship of the internet in China. It was one of many on the same subject. Over the years they have been published many times, by several human rights organizations, but rarely reported in mainstream media. This one became front-page news in newspapers, and was covered in peaktime news on television, because it involved two large “western” companies.

There is nothing new about information control and censorship in China – as in many other places. But this time the news is about two US-based international online services cooperating with the Chinese authorities in blocking access to “forbidden” sources.

In the case of Yahoo, it’s been going on for years. The “news” is that apparently also Google is “complying”. In June 2004 Google bought into Baidu, a Chinese search engine – that, as others in the country, doesn’t allow access to websites that aren’t approved by the government. It is reported that, under pressure from Chinese authorities, in July Google agreed to “filter” also its own system. Sites that can be found with Google in the rest of the world are not reachable by internet users in China.

Several American companies have been cooperating with Chinese censorship. For instance Reporters sans frontières explains that Cisco Systems has supplied thousands of routers that are used to spy the use if the internet by people in China.

“Compliance” by international search engines or information services doesn’t change the practical circumstances. Chinese authorities can (and do) block sources that don’t support censorship. And anyone in China who is found bypassing the rules can be sent to jail. But cooperating actively isn’t the same as being victimized. This isn’t just a serious violation of principle. It also paves the way for all sorts of other compromises.

China is the biggest, by sheer size – but it isn’t the only problem. The Global Internet Freedom Act, approved by the US congress in July 2003, seeks to fight censorship in countries such as China, Burma, Cuba, Syria and Saudi Arabia. But the list of governments censoring information as whole, and specifically the internet, is much longer. It includes countries with a large population, such as Iran and Vietnam, but also many others, that aren’t less important just because they are smaller. (A report by RSF The Internet Under Surveillance covers 60 countries).

The situation around the world is very complicated. There are countries with low internet activity because of censorship, others for basically different reasons. And there are countries (e.g. in South-East Asia) with a high quantity of online presence, but poor quality, because of severe restrictions of freedom. (Things are likely to get worse in Hong Kong, that has a separate TLD and a three or four times larger hostcount than mainland China, if central authorities continue to enforce tighter control – and there isn’t much more freedom in neighboring countries.)

The question is... why did this particular item, in the vast and serious subject of censorship and human rights, gain special attention in mainstream media?  The answer is that Google is in the limelight because of its announced stock offer. There is something weird, and perplexing, in this way of approaching the subject. Media coverage tends to ignore, or to treat very superficially, the complex and unhealthy ties that link finance and repression, censorship and terrorism, compromise and manipulation. How can we trust media that touch (vaguely) on such subjects only when (maybe) they can have an influence on the stock market?

Are we to expect that Google, driven by greed and financial pressures, will fall into the tangle of compromises that drove its competitors to failure?  That, so far, seems unlikely. But, if there were more symptoms of unhealthy “compliance”, we should be ready to drop Google (as we did a few years ago with traditionally established search engines) and move on to whoever will offer us free, open and reliable resources.


back to top


loghino.gif (1071 byte) 3. Large low-density countries

In issue 31 (February 1999) I published the results of an analysis, that had started two years earlier, on “large low-density countries”  It wasn’t updated in following years because other documents, that were being added in the data section, helped to understand the global situation. Another reason was the dismaying fact that, in some of those countries, the situation wasn’t changing – or it was getting worse in proportion to the general growth of the internet (most readers, anyhow, seem more interested in what is happening at the higher end of the scale.)

But I guess it’s time to take another look. Some things did change in five years. The most obvious fact is that the gap has increased. Some of those countries no longer have “low density”, while others are lagging behind. Of course it’s arbitrary, as it was five years ago, to draw the line at 50 million inhabitants. And of course what is happening in many other countries, regardless of size, is important. But this is just a way of showing, with a few examples, how different the situations are around the world.

Here is a summary of how things have changed in five years, from 1998 to 2003, in the 18 countries considered in this analysis. (The first four are no longer in the “low density” category, but they are included here for the sake of comparison with five years earlier.)

  number of
hosts 2003
times growth
in 5 years
per 1000
Brazil 3,163,349 x 14.7 18.6
Mexico 1,333,406 x 11.8 13.7
Russia 800,277 x 4.8 5.5
Turkey 359,500 x 11.1 5.7
China * 160,241 x 9.3 0.12
Ukraine 130,596 x 8.3 2.6
Thailand 103,700 x 5.0 1.7
India 86,871 x 6.6 0.09
Indonesia 62,035 x 4.0 1.0
Philippines 27,996 x 3.0 0,4
Egypt 22,452 x 10.7 0.3
Pakistan 15,124 x 4.9 0.1
Nigeria 1,172 x 2.9 0.01
Iran 496 x 2   0.008
Vietnam 340 x 10   0.004
Congo 153 n.a.   0.003
Ethiopia 9 n.a.   n.a.  
Bangladesh 2 n.a.   n.a.  

* Not including Hong Kong  (see Chinese area)

Worldwide hostcount in 2003 is 5.4 times larger than it was in 1998.
Average density is getting close to 15 hosts per 1000 inhabitants.
Seven of these countries have faster-than-average growth,
six are slower, in five figures are too small to be comparable.

There is growth in several countries, but only two (Brazil and Mexico) have reached a level that can lo longer be called “low” (in the data section there a document on large laguage communities that includes an update of internet development in Latin America.)

Russia and Turkey are far below the world average, but they are considerably above the level of less developed countries (see European and international data.)

In other countries, even when the growth factors are high, total numbers and density remain very low.

The following graphs include the 12 countries, of these 18, that five years ago had over a thousand internet hosts and now have over 20,000.

They can’t be all placed in one graph, because differences are too large. The first group includes four countries with densities that are no longer “low” – between 5 and 20 hosts per 1000 inhabitants.

Four countries
(thousands of internet hosts)

4 countries
The red part of bars shows growth in five years (from 1998 to 2003)

There is remarkable growth. The fastest is Brazil – now one of the eight largest countries in the internet worldwide. Mexico has overtaken Russia (and also sereral countries that were traditionally strong online – see international data.)

In the second graph the situation is different. We see three “large countries” with densities between 1 and 3 hosts per 1000 inhabitants. One in Europe, two in Asia.

Three countries
(thousands of internet hosts)

3 countries
The red part of bars shows growth in five years (from 1998 to 2003)

There are different speeds of growth – all retalively high, in a five-year period, as a percentage. But levels remain low in relation to population.

In the third graph we see five countries with very low density – between 0.09 e 0.5 hosts per thousand inhabitants. Four in Asia, one in Africa.

Five countries
(thousands of internet hosts)

5 countries
The red part of bars shows growth in five years (from 1998 to 2003)
Hong Kong is not included in China

The fastest growth, as a percetage, is in Egypt. – followed by China and India. There are different reasons why internet development is very low in the world's two largest countries. (In the analysis of language communities there are data about Chnese-speaking countries – and at the end there is a map of the situation in Asia.)

While in China, as we have seen, there is severe political repression, there are other problems in India (not only poverty, but also bureaucratic and organizational problems that stand in the way of internet development.)  This is a very complex situation, but it’s hard to understand why a country with a billion inhabitants, more English-speaking people than the British islands, and high levels of technical and cultural competence in larger numbers than most countries worldwide, has fewer internet hosts than Iceland with less than 300,000 inhabitants. The entire Indian “sub-continent” has just over 100,000 internet hosts.

In the other six “large countries” the level of online activity is extremely low. In Nigeria the number of internet hosts increased from 419 in 1998 to 1172 in 2003, but density is 0,01 per thousand inhabitants. In a large part of Africa things aren’t any better. But, as we shall see in the following analysis, Nigeria, with over 100 million inhabitants, has fewer internet hosts than some much smaller African countries.

In Vietnam the number of hosts increased, in five years, from 34 a 340, in Iran from 244 to 496. These would be tiny numbers even in countries with a much smaller population. We know that is both these countries, as in many others, freedom of speech and communication are heavily repressed.

Internet activity in Congo is extremely low, and it’s even worse in Ethiopia, where the number of hosts appears to have decreased from 78 in 1998 to 9 in 2003. No internet hosts were found in Bangladesh in 1998, and now there appear to be two. Several other countries, in Africa and Asia, are close to zero hostcount.

These, of course, are only a few examples of a great variety of situations around the world. There is considerable global development of the internet, with some countries growing much faster than average, while others are slower. A large part of the world still has very limited access to the net (see international data).


back to top


loghino.gif (1071 byte) 4. The internet in Africa

In ten years of studying the development of the internet I never published an analysis of the situation in Africa. I tried several times to learn something from available data, but (except in the case of South Africa) numbers were too small to be meaningful. Though some of the figures aren’t as tiny as they used to be, even now, in most cases, they are only marginally relevant. But maybe we can begin to take a closer look to what is happening in the continent with the lowest general level of internet activity.

Two facts are clear. Most of the continent has a very low hostcount. And (see international data) eight tenths of the total are in one country, South Africa, that has 5 percent of Africa’s population.

This table shows the situation in the 13 African countries with over a million inhabitants and more than a thouand internet hosts.

  number of
hosts 2003
% growth
in two years
per 1000
South Africa 288.633 + 21 6,7
Egypt 22.452 + 284 0,34
Kenya 8.325 + 277 0,28
Morocco 6.517 + 378 0,26
Tanzania 5.534 + 272 0,16
Zimbabwe 4.501 + 29 0,43
Mozambique 3.249 n.a.   0,18
Namibia 3.164 n.a.   2,09
Botswana 1.920 + 51 1,14
Zambia 1.880 + 72 0,20
Rwanda 1.495 + 32 0,20
Swaziland 1.495 + 23 1,36
Nigeria 1.172 + 24 0,01

As explained in international data
worldwide growth of the internet
was 58 % in the last two years
and world average density is close
to 15 hosts per 1000 inhabitants

In a general situation of very low density there are relevant differences. Egypt, while still far behind South Africa, is considerably ahead of all other African countries. There is high percent growth in Morocco, while in other North African countries internet activity remains extremely low. Libya has the same density as Nigeria, Tunisia only slightly better and Algeria much worse.

In the rest of Africa there are some indications of growth in a few countries, such as Kenya and Tanzania, and there is relatively high density, compared to most of the continent, in a few small countries such as Namibia.

In this graph we see density (hosts per 1000 inhabitants) in the same countries as in the chart – excluding South Africa.

Internet hosts per 1000 inhabitants
in 12 African countries

12 contries

This chart can’t be taken too seriously because (as pointed out at the beginning) figures are too small to be accurate, or even approximately close to reality. And these are only a few examples of the different, and complex, situations in many countries. But it’s clear that, even in a low-development environment such as Africa, local situations vary – ad some change over time.

In the analysis of “large countries” we have seen the extremely low level of online activity in Congo and the disastrous situation of Ethiopia. Somalia and Eritrea aren’t doing any better. No active internet hosts were found in Sudan and Zaire. Countries with a lower density level than Nigeria include Uganda, Cameroon, Chad, Burundi... and several others.

Let’s look at this picture also as a map.

Internet hosts per 1000 inhabitants


Two “blue dots” would appear in the ocean
if this map included the Seychelles and Mauritius,
that have over three hosts per thousand inhabitants.
Density in the São Tomé and Principe islands
is close to South Africa’s. But, of course,
small numbers aren’t very reliable.

Compared to the maps in three documents in the data section  (WorldEuropeLatin Anerica and Asia)  the scale in this case is based on much lower density levels.

Also in other parts of the world it is noticeable that geographic “vicinity” has some influence – as in the case of countries close to South Africa. But otherwise the spots of (relative) development are scattered – ad most of Africa has extremely low intenet activity.



back to top back to summary


Gandalf homepage