There are plenty of discussions about how, when and what to measure online, and nobody's claiming ComScore is inaccurate in its numbers. On one point, however, it's plain wrong. It hasn't measured how many people are connected to the Internet. It's measured how many people are on the Web--and the Web is not the Net.
That distinction matters. It matters because getting it wrong confuses the past and misrepresents the future. The Internet was around for decades before the Web, but the Web is the technology that made the Internet accessible to ordinary people. Invented in 1991, it's taken exactly the same time for the Web to spread from Tim Berners-Lee in his Geneva cubicle to a billion people worldwide as it has for Guns N' Roses to record a new studio album. That's almost a record.
In the history of the world, the only technology to spread faster is GSM, the digital mobile-phone standard. That's another European invention--and another one that went live in the wonderyear of 1991. This one, however, is in the hands of 3.5 billion people. Between them, the Web and wireless have democratized data: since 1991, you haven't needed approval from anywhere to store or share data with the world, and the world hasn't needed an appointment anywhere to retrieve it.
But that two billion disparity between GSM and the Web highlights why the Web is not, and never will be, all of the Net. Every GSM user can be considered part of the digital online community: GSM is a perfectly good digital wireless network with copious connections to the Internet, and the 2.5 billion users who aren't on the Web are most certainly connected--albeit at dial-up speeds.
Raw GSM is too slow for the Web, and a good thing too. That means handsets and infrastructure are cheap enough to get to the people who need them most. GSM is not too slow for the Internet-—text-only dial-up created global communities in the 1980s--and those two billion deserve decent services from us and the chance to develop their own. But if we count the Web as the Net, they vanish from sight.
The GSMA is already trying to get the next two billion connected, with development funds aimed at those earning around than US$2 a day. It will take decades to get Western-style broadband to 4.5 billion people: it takes next to no time to build services that will run over GSM to the farthest corners of the earth.
History tells us that ignoring disenfranchised majorities does nobody any favors. Common sense tells us that 3.5 billion trumps one billion. Foresight says the global status quo will change, with or without our consent, and sooner than we'd like. Such ideas deserve be taken seriously: getting the numbers right will be a good first step.