Does the OECD publish its criteria for collecting broadband subscriber statistics?
Yes. The OECD publishes the criteria for the broadband subscriber data collection here (Methodology 2010). The methodology was adjusted for the December 2009 data release and removes wireless broadband subscribers from fixed broadband statistics. All wireless broadband subscribers will be included in a forthcoming wireless broadband indicator.
For most countries this has had no noticeable impact on their total numbers or penetration rates because fixed-wireless subscriptions make up only a very small portion of total broadband. There are, however, a few countries with significant fixed-wireless subscriptions that may see total figures fall or rankings shift as a result of removing the fixed-wireless subscriptions. These include, the Czech Republic, the Slovak Republic, Ireland, Australia, New Zealand and Canada. The methodology for data collected and published through June 2009 is available here (Methodology through 2009).
What is the difference between the subscriber and user data that the OECD publishes?
One of the best ways to illustrate the difference between "subscriber" and "user" statistics is to consider another telecom example: the number of payphones in a country. Telecommunication operators know exactly how many payphones they have in operation and they can provide this data very easily to regulators if asked. This data can be a timely and accurate assessment of the extent of the payphone network in a country.
While operators do know the total number of payphones they cannot provide data on how many different users make calls from the payphone each month. Telecommunication operators have no good way to distinguish one payphone caller from another. They only know when a call is made. If regulators want to know how many people make use of payphones the best way to approach the question is by using surveys. It is possible to get an idea of the number of payphone users in a country by asking a subset of people whether they've recently made a call from a payphone and then extrapolating the results for the country as a whole.
Broadband subscriber and user data are much the same. Operators know how many "subscriber lines" they have in their network. The OECD collects and publishes this data from telecommunication regulators twice a year. The data give a very good measure of the physical lines in a country. The subscriber data does not however, provide any information on how the lines are used.
OECD member governments also conduct surveys to find out how broadband is used, particularly by businesses and households. These surveys are relatively infrequent and the dates may not correspond well across countries. The surveys provide important information on the number of households with a broadband connection, the number of people with access to broadband at work and the number of students with broadband access through their schools.
The OECD collects and publishes this data as broadband "users". The data is available on the OECD Broadband Portal and provides a more detailed view of broadband use by households and businesses, but at less regular intervals.
Subscriber data is available for all OECD countries but household survey data may not be. There are a number of countries which do not have national surveys on broadband use which can be used for comparisions across countries. These surveys are very important for understading some of the key trends which are not captured in subscriber data.
Why does the OECD express sibscribers "per 100 inhabitants"instead of as a percentage of households?
OECD subscriber data contains the total number of business and residential subscriber lines in a country. Normalising the number of broadband subscribers by the population provides an idea of relative penetration of subscriber lines. Expressing the number of subscriptions in terms of households would be misleading because some connections are to businesses. Normalising subscribers as a percentage of total households would consistently over-estimate broadband penetration.
How is business broadband measured?
OECD “subscriber” statistics count the number of broadband lines provided by operators in the country. This includes business and residential lines, with residential making up the vast majority. For clarification, the high-capacity leased line to a business counts as one subscription. Subscriber data does not count the number of business employees who may use that connection. Employee usage is captured by OECD user statistics. This user data is collected by national surveys in OECD countries and then harmonised and published by the OECD. In short, subscriber statistics include any business subscriptions and user statistics look at access by individuals in these businesses.
How does the OECD count Wi-Fi in broadband statistics?
Wi-Fi is a wireless technology which is typically used to share a broadband subscription among multiple users. OECD broadband "subscriber" statistics count the broadband subscriptions which supply connectivity to a Wi-Fi hotspot in a home or office. The OECD also counts any leased lines or business DSL connections which furnish Internet access to commercial Wi-Fi hotspots in a cafés, etc. OECD subscriber statistics DO NOT count the number of users who access the Internet via a Wi-Fi hotspot. People connecting to the Internet via Wi-Fi are counted as users, not subscribers. These users can be picked up in national Internet surveys. One of the reasons household survey data are so important is because they can capture trends such as Wi-Fi use which are not available from subscriber data.