More Facebook users than Internet users in South East Asia?


Posted on August 30, 2014  /  3 Comments

Internet & Facebook users, 2014

I have always been intrigued by the differences between South and South East Asian countries. We saw this over and over again when we did the Teleuse@BOP surveys. But playing around with some numbers for Facebook users in four South and four SE Asian countries, I was astounded. In all the SE Asian countries, there are more Facebook users than there are Internet users. In the case of Myanmar, the multiple is 4.

Common sense will tell you that you have to be an Internet user to use Facebook. But the SEA data refute common sense.

What are the possible explanations?

Something wrong with population numbers? But we used the most recent national censuses, except for Myanmar. In any case, the same numbers were used to calculate Internet users/100 and Facebook users/100. So this cannot be the source of error.

Is it possible that Facebook is reporting false data? We obtained these rounded numbers by going through the advertising door. This would be a serious offense since advertising rates are set on the basis of these numbers. We therefore do not think this explanation is valid, but we will check with Fb. Facebook does not ask people whether they use the service. They know.

It is possible that the Internet user numbers are wrong. We used the ITU method which we have expressed serious concerns about. But the sources of the ITU data for Indonesia, Thailand and Bangladesh are from the National Statistical Organizations and are likely to be from demand-side surveys.The others (India, Philippines, Pakistan, Sri Lanka, Myanmar) are ITU estimates. So unlikely that this is the source.

So the conclusion is that the cause is that Facebook users simply are not reporting that they use the Internet. This is what we found in our qualitative research. It is now appearing in the quantitative data too.

3 Comments


  1. Myanmar census numbers have come in. Its population is 51.4 million. This does not change the rounded Internet users/100 and Facebook users/100 numbers, since we had not gone with the high estimates.

  2. From Christoph Stork (slightly edited):
    “I can think of several additional explanations for this:

    1) The main one is similar to active subscribers vs owners of mobile phones.
    a) M2M: corporate or activity based Facebook accounts. I have for example two Facebook accounts, a personal one and another.
    b) Duplicated SIM: private Facebook account and one official one
    c) orphaned accounts that are not being used any longer but counted as active.

    2) Facebook data is realtime, census data several years old.

    3) Facebook is available on feature phones as an app. Users may use it but not use any web browser or email, so they are not aware that they are using the Internet. One can register on Facebook either with email or with mobile phone number.”

  3. Indeed Just to add to what Christoph has said. RIA found in its 2012 ICT access and use survey pilot, that several people had indicated they did not use the Internet, in some cases even know about it, but were using Facebook Zero, which with other social networking apps was the main driver of Internet take up across most of the 12 African countries surveyed. We subsequently adjusted the questionnaire so if you were using Facebook, then you your default answer to whether you used the Internet was yes. The questionnaire included a further a set of questions on what one used Internet for that cross checked this and of course indicated other Internet use, browsing, e-mail etc if any.
    Just to add further that the national census done around the same time as our survey in South Africa, has similar high-level results but just does not go into use sufficiently (at all)
    to begin to answer these kinds of questions. Further like NSO this data is collected at the household level with very broad definitions, not the individual user level where mobile communications, certainly for the cordless generation is an extension of Self.