gian@gandalf.it

Once again, we see that the net is still in its infancy and with a patchy pattern; it's very unlikely that it will be a common tool worldwide "by the turn of the century". Several more years will have to pass, and probably some major changes will have to take place, before it is really understood and used globally.

A few general observations:

• The growth of the net does not follow any "logical curve".
• At this time it's slower than in past years.
• There are big differences between areas and countries; over time they change, but they don't decrease. Some countries have a diminishing number of internet hosts, and it's not always easy to understand why.

Each reader, I think, can draw conclusions from some of the detailed data, according to his or her specific interests. Broadly, I think a simple truth is confirmed: there will be more disappointments for generic efforts, big opportunities for entrepreneurs who understand the evolution and the way of using the net to fit their specific needs and resources.

After the "quantitative" analyses I shall quote the comments by the author of a recently published book. Other interesting opinions will be reported in future issues of this newsletter. And also a few more considerations based on numbers - there is more to be learnt from data than can be crammed into one issue.

Things are beginning to change. Techno-hype continues to prevail, but experience and deeper thinking are beginning to make their mark: more and more often we read comments that are precise and considerate, with a sound analysis of facts and a healthy dose of common sense.

The growth rate is wobbling up and down. After a sharp acceleration in 1994-95 (probably due to the spreading of the World Wide Web and hypertext) it's decreasing. Here are the growth rates in the last five years by six-month periods (number of internet hosts worldwide):

% change in half year

In the first half of 1997 the number of hosts in the United States grew 18 percent, versus 21 percent worldwide. Several countries had fast growth, e.g. Singapore (more than doubled), Venezuela (93 per cent), New Zealand (84), Korea (73), Malaysia (60), Argentina (50). In the next chapter we shall see data on European and Mediterranean countries.

The two largest countries in the world are still far behind. China in this period grew "only" 30 percent (but threefold by adding Hong Kong). India (that has enormous potential) grew 53 percent, but its numbers are very small. I think the situation of large countries with low net presence is worth a deeper analysis, which I shall include in the next issue (this one already has too many "numbers").

By major geographic areas, this is the situation:

As we can see, net penetration is still unbalanced. It's even worse if we consider the dominant status of a few countries in each area. In spite of growth in other parts of Asia (especially in the South-East) Japan still has 66 percent of all hosts in the continent. 95 percent of African hosts are in South Africa. Half the hosts in South and Central America are in Brazil - and another 30 percent is in two countries, Chile and Argentina. We shall see Europe in the next chapter.

The internet is still concentrated in a small part of the world. Twelve countries, with 13 percent of the world's population, have 90 percent of the net, as we can see in this graph (which does not include the United States, for the sake of readability).

"Big" countries in the net
Countries with over 200.000 internet hosts according to Network Wizards - July 1997
(United States not included)

It's pretty obvious that some "small" countries are big in the net, and vice versa. This becomes clearer if we look at per-capita density.

Hosts per 1000 inhabitants
Countries with more than 50.000 hosts and density over 3

As seen from Italy, this is not encouraging. We have density marginally above the world average, less than half the average in Europe. Italian hosts are 2 percent of the world's, while Italy's is approximately 4 percent of the global economy.

RIPE (Réseaux IP Européens) July 31, 1997
(this list includes only 22 countries with over 20.000 hosts - out of 52 in the RIPE area)

Growth is fast - but less than generally expected. Britain and Germany dominate Europe, with 40 percent of the net (20 percent of the population). Germany shows fairly strong growth (above European average) but appears to be concentrating on itself and less dynamic than other countries in exchanges with the rest of the world. Italy had strong acceleration between the end of 1996 and the beginning of this year, but is slower now; other European and Mediterranean countries are more dynamic. Turkey has the fastest growth in this period (overtaking its neighbor and rival, Greece) followed by Portugal, Hungary, Spain, Poland and Russia.

The French peculiarity

In all these analyses, France is "underestimated" because hostcounts don't include the "French peculiarity"; widespread use of the minitel. In many countries, of course, there are relevant online activities that don't use the internet: BBSs, community networks, intranets, etc. But some sources estimate that in France there are 14 times more users of minitel than of the internet - I can't think of any other place in the world with anything comparable. Of course the two systems can't be compared mathematically; but with this "peculiarity" online activity in France could be one of the highest in Europe - though minitel communication is within the country and not with the rest of the world.

Here is the European picture in per-capita density.

Hosts per 1000 inhabitants
This graph includes countries in the RIPE area (Europe and Mediterranean) with density over 2.

Italian host density is less than half the average in the European Union (and the lowest among "G7" countries). But if we look, for instance, at the high level in the little Republic of San Marino... it seems very likely that there are "pockets" of high density in our country, as in others (not necessarily in the big cities) though so far we have no reliable data on where they are.

Another interesting picture is density in relation to income (GNP).

Internet hosts in European Union countries in relation to GNP
Elaboration on RIPE data - July 31, 1997

Germany in this analysis is barely above the EU average: its presence on the net is not in proportion to its general economic dominance. Italy's weakness is sadly obvious.

Two small statistics

"Small numbers" can be relevant, if seen in relation to broader data. Few sites worldwide publish statistics of hits by country of origin. Here are two examples that, I think, are interesting.

The first is the percentage of inquiries received by RIPE - the European internet control and information service - from September 1996 to July 1997:

(This list does not include 96 countries with a percentage below 0,5)

It's easy to understand the relatively low level of queries from the United States and other non-European countries, as RIPE information concerns Europe; not so easy to explain the relatively low frequency of high online density countries such as Finland - maybe it's because they have better information services locally.

Another example is hits on a commercial site, House of Ireland, in a short period - April 20 to 26, 1997.

Of course such tiny statistics have no general significance, but I think it's interesting to see what happens on an individual commercial site.

In this example - 80 percent of hits on a European commercial site come from North America, less than 3 percent from the home market and 10 percent from the rest of Europe.

Yes. - Hagel answered - Initially virtual communities were a spontaneous social event in a variety of network environments. What people fundamentally want is to connect with each other and create ongoing communities. But this represents an interesting commercial opportunity as well.

Even more concisely - he explained - "Community precedes commerce". In earlier network environments, there was a notion that commerce and community were somehow at odds with each other, that as soon as you introduced commerce, you undermined community. Certainly that can be true if abused, but if it's thoughtfully done, commerce can actually reinforce community and extend value to members.

Is commerce - asked the interviewer - where the Web is headed?

In the near term - Hagel answered - we're setting ourselves up for a major backlash against the Web. Business plans have been put together by large companies and investors that say these early ventures have to be big businesses very quickly. But the unappreciated implication of increasing-returns models is that you go through an early period of very slow buildup before you hit accelerated growth. The expectations have been set too high.

Hagel observed also that too many strategies were set up without proper regard for the way information is collected and used. If the use of personal information starts to be systematically abused, we shall see a huge backlash around privacy.

He also said: When I compare how much money is invested in the technology itself versus that spent on understanding social dynamics - it's a hugely baffling imbalance. And yet so much depends on understanding the latter.

So... even in the United States, where the net environment is much more advanced than in Europe, the perspective is still seriously unbalanced. Most people and companies fail to understand that what matters is not the technology, but how people use it. Net marketing, as any use of the net, cannot be handled with blind automatism, vague strategies or hasty choices; it needs depth, time and patience.

In mid August (when most retailers are closed in my country) the old hard disk of the computer I use more often, after years of faithful service, showed the symptoms of a terminal crisis. Luckily I had backup of almost everything, and time enough to save most of my latest work before the unavoidable catastrophe. But collapse was close... where could I find a replacement?

Easy, I thought: order it online. In a few minutes I found what I wanted and placed the order. In a few hours the vendor, very efficiently, handed the parcel to the courier. Delivery was expected the next morning. But it never arrived.

The system provides online tracking. After a day of waiting, I found a statement: "nobody available to sign". That was not true. We were here waiting. The courier came to the right address, but (quite mysteriously) left without delivering. They promised to deliver the next day. Everyone was on the outlook... but this time no van from that courier appeared in the neighborhood. That night, online tracking repeated the same wrong statement. And so on...

During days of waiting and frustration I had several telephone conversations with the courier's central system. They had the same data that I could read online; but no way of reaching the people making the deliveries and correcting the mistakes, or even finding out what and why had gone wrong. They had my telephone number and may e-mail address, but nobody ever called or wrote. Eventually I had to cancel the order and buy the hard disk from someone else.

I think I know what happened. During August the courier was operating with reduced, maybe temporary, staff. The person making the delivery was inexperienced, nervous and confused. Such things can happen... but the courier's office could not track down the problem and fix it.

These inefficiencies cost them repeated unfulfillments, cancellation of the order and loss of a customer - it's pretty obvious that I shall never again order from that supplier, or anyone using that courier.

The moral is simple: if we put too much trust in automatic technology, and fail to provide human support in case of trouble... the consequences can be much more expensive than the savings we thought we could make by reducing the level of service.