Forget about ABCe; let’s have an old-fashioned fight about traffic numbers

The Telegraph.co.uk has ended the Guardian‘s long reign as the most widely-read national newspaper website. So writes my colleague Martin Stabe on the basis of ABCe data for April that run like this:

Telegraph.co.uk: 18.6m unique users
The Guardian: 18.54m unique users
Mail Online: 18m unique users
Times Online: 15.4 unique users
Sun Online: 14m unique users
Mirror.co.uk: 4.28m unique users

Now ABCe hasn’t been established as the gold standard of traffic measurement for very long. But already, the new regime has become boring.

Am I the only one who feels nostalgic for the days of near-impenetrable arguments about rival sets of traffic figures propounded by executives who themselves don’t fully understand the data? Surely someone could be persuaded to start an argument?

Admirably, ComScore seems interested in fomenting a barney. Just as ABCe’s hegemony seems to be solidifying, the panel-based traffic counting firm has decided to start publishing monthly figures for newspaper sites for the first time.

ComScore’s traffic analysis of the UK’s nationals for March looks promisingly discordant with ABCe — and (coincidentally) rather positive for News International:

The Sun: 4.3m visitors
The Guardian: 3.6m visitors
Telegraph Group: 2.8m
Times Online: 2.6m
Daily Mail.co.uk: 2.4m
Independent.co.uk: 1m
Mirror.co.uk: 990,000

OK — more seriously now. . . ComScore’s numbers are scaled up from a “panel” of UK consumers.

The UK bit is important. The Mail Online, for example, ranks high in ABCe, but boasts a large overseas readership. Cut that out influence — as ComScore claims to have done — and the site’s aggregate traffic looks much less impressive.

ABCe’s data is pulled directly from publishers’ server logs. Quite apart from its UK focus, ComScore suggests that its figures are lower than others because:

1) Its panel-based (research) methodology is not skewed by cookie deletion.

2) It counts as one individual the same person hitting a site from different locations (eg: home and the office).

These may well be persuasive arguments. But it’s hard to tell.

As the spat over ComScore’s data for Google’s Q1 clickthroughs proved, ComScore’s problem is a lack of transparency about how its research methodology works in practice.

Coincidentally, the same argument applied to Nielsen NetRatings. The last time I looked at its site, it contained the vaguest of descriptions of its methodology, lost beneath layers of corporate verbiage.

And Hitwise? I like the idea of pulling traffic numbers from ISPs’ servers (which is what the company does). In my humble opinion, Hitwise also does a slightly better job of explaining its methods to the outside world. . .

Of course, all of this is irrelevant if all you need are figures to bamboozle gullible clients during the first 20 seconds of a sales presentation. . .