We started publishing computer statistics in 1986. At that time we were with Future Computing, which was the leading market research company in the PC industry. Future Computing had collected the most PC industry statistics of any organization in the world. Over the next 15 years our country-specific computers-in-use statistics became the main source of such data. As the eight Computer Industry Almanacs were sold across the world we received better data from our readers. Many organizations publish yearly computer shipment data for major countries, but there is little country-specific computers-in-use data. We are now doing the same for Internet user statistics.

To test the data we calculate computers per capita, Internet host computers per capita and Internet users per capita. These ratios tend to fall in ranges depending on a country's industrial, economic and political development. At a minimum these ratios show data points that need further checking and analysis. We also calculate the ratio between computers and Internet hosts and the ratio between computers and Internet users. These ratios are also helpful tools to spot anomalies in the data.

It is speculative to project the number of Internet hosts and Internet users for year-end 2005 through 2010. To do so we first projected the computers-in-use for each country. The increases in the installed base of computers per country happen at a reasonably predictable rate. Using four guidelines we projected the Internet host and Internet user numbers.

The first guideline is the current growth rate for each country. The second guideline is the relative maturity of each country's adoption of the Internet. The maturity is indicated by the ratio between computers-in-use to Internet hosts and to Internet users, the Internet hosts per 1,000 people and Internet users per 1,000 people. Some countries have been very active in their adoption of the Internet and their growth rates have already dropped. Other countries are just starting to use the Internet and their growth rates have not peaked yet.

The third guideline is the economic climate, telecommunications infrastructure and pricing, gross domestic product per capita and political and religious freedom. If the economy is poor, if the telecom network is limited or expensive and/or the average income per capita of a country is low, then the potential for Internet usage is low because relatively few companies and few consumers can afford the cost of Internet hardware and usage fees. Another important factor is the political and religious freedom in a country. If the personal freedom in a country is limited, either by politics or religion, the potential for Internet usage will be much lower than without these restrictions. Examples are Cuba and North Korea that have limited political freedom and Iran, Iraq, Saudi Arabia and others where religious restrictions put a damper on Internet usage.

The fourth guideline is the surveys and projections by market research companies, associations and government organizations across the world. There are numerous Internet user studies and surveys that provide useful information. However, these statistics must be used with care as it is often unclear what was measured and how the measurements were done. All these factors are used to project the year-end 2005 through 2010 data. As we gain additional information we expect to update these estimates and projections.

To summarize the forecast methodology, there are basically four different ways of estimating the number of Internet users in a country:

1. Projections from a well-designed and executed survey. An increasing number of surveys are now available from government agencies in many countries.
2. Estimates based on the number of Internet host computers. The Internet user/Internet host computer ratio can be estimated from previous survey data or by comparison with countries with similar Internet/computer industries.
3. Estimates based on the portion of computers-in-use with Internet access. The computers-in-use/Internet user ratio can be estimated from previous survey data or by comparison with countries with similar Internet/computer industries.
4. Projections from the number of Internet access accounts and the average number of Internet users per account. Such data is available from ISP associations in a few countries.

In preparing the estimates we primarily used the first three methods. Internet access account estimates from ISP associations are a promising methodology for the future, but few statistics are currently available.

To make an estimate and forecast of computers-n-use and PCs-in-use for many countries is a daunting task. Ideally such data would be based on comprehensive and identical surveys in every country and such surveys would have been conducted on a regular basis over many years. However, such data does not exist and will not be available for a long time. The reason is that such surveys are too expensive to do for a single organization. If such data existed the price would only be affordable to the largest companies.

Instead there is a variety of survey data and computer shipment estimates from many different organizations. These surveys and sales estimates use different methodologies and are primarily done in the major countries. Some surveys are excellent, some are fair and a few are actually misleading. The result is that there is lots of conflicting survey data and shipment estimates and it takes a major effort and knowledge to determine what is good information and what is poor data that should be discarded.

The strategy of this project is to collect as much existing data as possible and use a common PC product definition across the geographic regions. A lot of time has been spent analyzing a wealth of data. The resulting forecast is an inexpensive product that has all the top-level statistics for the worldwide PC industry.

The historical part of this forecast is based on several types of information as explained below:

1. Survey data from computer resellers. The survey data is especially plentiful for the U.S. market where monthly survey data of computer resellers is available since 1984. Lesser amounts of survey data are also available for the European market. This data includes PC product mix, prices and PC vendor sales information.
2. PC manufacturers' annual, quarterly and federal financial fillings. These documents include product mix data, revenues by geographic regions and often contain PC shipment figures by geographic regions.
3. PC manufacturers' press releases, which often list their sales, market shares, rankings, milestones and other useful information.
4. Government statistics from many different countries. Useful information includes survey data that estimate how many households have PCs, number of PCs per household, future buying intentions, how many workers have PCs and so on.
5. News reports and articles from the computer and business press and their websites. There is a wealth of information nuggets buried in thousands of articles.
6. Research reports by Wall Street analysts. These reports often have useful shipment information.
7. Internet user statistics. The recent flood of Internet user statistics is helpful in estimating the number of PCs sold and PCs-in-use.
8. A personal library of computer industry statistics that goes back to the beginning of the PC industry. These statistics started with Future Computing, which was the premier PC market research company in the 1980-1986 timeframe. Future Computing collected the most PC industry statistics of any organization in the world. The data collection continued for the Computer Industry Almanac and Internet Industry Almanac, which are reference books that contain lots of computer market statistics. For over 15 years country-specific PCs-in-use and computers-in-use statistics were collected for over 50 countries. As the eight Computer Industry Almanacs and one Internet Industry Almanac were sold across the world additional data was received from some readers.

The forecast for future years is more difficult to do. However, 20 years of experience in forecasting the PC industry is helpful. The following factors were used in the forecast.

1. Ultimate PC penetration estimates is a useful technique. The ultimate PC penetration rate is different for each country and region and basically depends on the wealth of each country and region. More information is shown below.
2. PC sales per capita and PCs-in-use per capita are calculated to track how far the PC penetration has proceeded. Increasing PC penetration invariably slows the PC shipment growth rate.
3. PC replacement rates and PC product lifetime. Existing PC owners replace their PCs at regular intervals. The replacement rates depend on the market maturity, economic conditions and the useful life of PCs. In each country and region a portion of new PC sales replace older and obsolete PCs. The forecast assumes a PC lifetime of four to five years in the industrialized countries and five to six years in the less developed countries.
4. Internet market dynamics. Since the Internet has become the main driving force for PC adoption, the market dynamics of the Internet growth is a major factor in PC growth. The results from another forecast product, which estimates of the number of Internet users for the same geographic regions and 50+ specific countries, was used as a guideline.

The computers-in-use forecast was developed using four guidelines. The first guideline is the current growth rate for each country. The second guideline is the relative maturity of each country's adoption of computers and the Internet. The maturity is indicated by computers-in-use per 1,000 people, the ratio between computers-in-use to Internet hosts and to Internet users, the Internet hosts per 1,000 people and Internet users per 1,000 people. Some countries have been very active in their adoption of computers and the Internet and their growth rates have already dropped. Other countries have computer penetration and are just starting to use the Internet and their growth rates have not peaked yet.

The third guideline is the economic climate, telecommunications infrastructure and pricing, gross domestic product per capita and political and religious freedom. If the economy is poor, if the telecom network is limited or expensive and/or the average income per capita of a country is low, then the potential for computer usage and Internet usage are low because relatively few companies and few consumers can afford the cost of Internet/PC hardware and usage fees. Another important factor is the political and religious freedom in a country. If the personal freedom in a country is limited, either by politics or religion, the potential for Internet usage will be much lower than without these restrictions. Examples are Cuba and North Korea that have limited political freedom and Iran, Iraq, Saudi Arabia and others where religious restrictions put a damper on Internet usage.

The fourth guideline was the surveys and projections from press releases, publications, websites, associations and government organizations across the world. There are numerous Internet user studies, computer shipment estimates and surveys that provide useful information. However, these statistics must be used with care as it is often unclear what was measured and how the measurements were done. All these factors are used to project the year-end 2004-2009 data. As we gain additional information we expect to update these estimates and projections.

Today there is a wealth of market research data for the PC industry. The problem is not a scarcity of data, but too much data that often shows conflicting results. Data disagreements are usually caused by different definitions of product categories between different sources. It is especially difficult to compare data between different geographic regions from different research sources. The result is that there is lots of conflicting survey data and shipment estimates and it takes a major effort and knowledge to determine what is good information and what is poor data that should be discarded. The next sections explain the forecast methodology and assumptions.

The strategy of this project is to collect as much existing data as possible and use a common PC product definition across the geographic regions. A lot of time has been spent analyzing a wealth of data. The resulting forecast is an inexpensive product that has all the top-level statistics for the worldwide PC industry.

The historical part of this forecast is based on several types of information as explained below:

1. Survey data from computer resellers. The survey data is especially plentiful for the U.S. market where monthly survey data of computer resellers is available since 1984. Lesser amounts of survey data are also available for the European market. This data includes PC product mix, prices and PC vendor sales information.
2. PC manufacturers' annual, quarterly and federal financial fillings. These documents include product mix data, revenues by geographic regions and often contain PC shipment figures by geographic regions.
3. PC manufacturers' press releases, which often list their sales, market shares, rankings, milestones and other useful information.
4. Government statistics from many different countries. Useful information includes survey data that estimate how many households have PCs, number of PCs per household, future buying intentions, how many workers have PCs and so on.
5. News reports and articles from the computer and business press and their websites. There is a wealth of information nuggets buried in thousands of articles.
6. Research reports by Wall Street analysts. These reports often have useful shipment information.
7. Internet user statistics. The recent flood of Internet user statistics is helpful in estimating the number of PCs sold and PCs-in-use.
8. A personal library of computer industry statistics that goes back to the beginning of the PC industry. These statistics started with Future Computing, which was the premier PC market research company in the 1980-1986 timeframe. Future Computing collected the most PC industry statistics of any organization in the world. The data collection continued for the Computer Industry Almanac and Internet Industry Almanac. For nearly 15 years country-specific PCs-in-use and computers-in-use statistics were collected for over 50 countries. As the eight Computer Industry Almanacs and one Internet Industry Almanac were sold across the world additional data was received from some readers.

The forecast for future years is more difficult to do. However, 20 years of experience in forecasting the PC industry is helpful. The following factors were used in the forecast.

1. Ultimate PC penetration estimates is a useful technique. The ultimate PC penetration rate is different for each country and region and basically depends on the wealth of each country and region. More information is shown below.
2. PC sales per capita and PCs-in-use per capita are calculated to track how far the PC penetration has proceeded. Increasing PC penetration invariably slows the PC shipment growth rate.
3. PC replacement rates and PC product lifetime. Existing PC owners replace their PCs at regular intervals. The replacement rates depend on the market maturity, economic conditions and the useful life of PCs. In each country and region a portion of new PC sales replace older and obsolete PCs. The forecast assumes a PC lifetime of four to five years in the industrialized countries and five to six years in the less developed countries.
4. Internet market dynamics. Since the Internet has become the main driving force for PC adoption, the market dynamics of the Internet growth is a major factor in PC growth. The results from another forecast product, which estimates of the number of Internet users for the same geographic regions and 50+ specific countries, was used as a guideline.

The PC shipment forecast was developed by using three more guidelines. The first guideline is the current growth rate for each region. The second guideline is the relative maturity of each region's adoption of PCs and the Internet. The maturity is indicated by PC sales per 1,000 people, PCs-in-use per 1,000 people and Internet users per 1,000 people. Some countries have been very active in their adoption of computers and the Internet and their PC growth rates have started to drop. Other countries have limited PC penetration and are just starting to use the Internet and their growth rates will remain high.

The third guideline is the economic climate, telecommunications infrastructure and pricing, gross domestic product per capita and political and religious freedom. If the economy is poor, if the telecom network is limited or expensive and/or the average income per capita of a country is low, then the potential for PC usage and Internet usage are low because relatively few companies and few consumers can afford the cost of Internet/PC hardware and usage fees. Another important factor is the political and religious freedom in a country. If the personal freedom in a country is limited, either by politics or religion, the potential for Internet usage will be much lower than without these restrictions. Examples are Cuba and North Korea that have limited political freedom and Iran, Iraq, Saudi Arabia and others where religious restrictions put a damper on PC adoption and Internet usage.