Buzzfeed’s news numbers

Digiday has a piece today announcing that 17% of Buzzfeed’s web traffic goes to news. Here’s the key part:

Of BuzzFeed’s 76.7 million multiplatform unique visitors in April (comScore), 17 percent were coming for news. The publisher historically hasn’t broken out its content by vertical to comScore, like other top news sites including CNN, Yahoo and The Huffington Post do. But it started to on a limited basis as of last month, when it began breaking out its Entertainment and Life coverage (43.7 million and 20.1 million uniques, respectively) to comScore. Stripping out those verticals leaves 13 million uniques for the rest, including hard news.

Ignoring analysis for the moment, let’s just look at the reasoning here. Unique browsers can visit more than one section of a site, so it’s possible that there’s overlap, and that simply subtracting the known verticals from the known total traffic isn’t a useful way to start. (I’m not certain of comScore’s methodology for vertical breakouts, but would be surprised if it doesn’t let sites count a user twice in different verticals, given that audiences overlap.)

So that 17% could be higher than it appears at first. Then, that 17% includes hard news plus everything that doesn’t fit into Entertainment or Life, so the actual audience for Buzzfeed’s news could be smaller than it appears at first.

The other element here is that unique browsers are the broadest possible metric, and likely to show news in the brightest light. In March, again according to comScore, Buzzfeed averaged 4.9 visits per visitor and 2 views per visit in the US, for roughly 10 views per visitor. If, among the 17% who visited news at all, five of those views are to news, then news is very, very well read with an exceptionally loyal audience. If just one of those views is to news, then news is much less well read than “17% of traffic” might suggest.

This post has been brought to you by early morning web analytics pedantry.

Update: Digiday has now altered the headline of its post and the text of the paragraph posted above.