Those Millions On Facebook? They Actually Visit, And It’s Not A Huge Deal Anyway.

Traffic numbers provided by companies should always be questioned — I mean, of course each company is going to try to present the data in a way that makes them look as good as possible. Which is what New York Times finance writer Andrew Ross Sorkin has understandably done, going to town on Facebook for how it counts its active users in an article out tonight called “Those Millions On Facebook? Some May Not Actually Visit.”

His main criticism is that Facebook counts 845 million monthly active users and 483 million daily active users, but gets to these numbers by including people who click “Like” or take another action on the web or mobile devices — but don’t visit Facebook.com during that time. Because they’re not visiting the home site, where the ads are, he suggests Facebook might not be making as much money off of them.

First, I’ll look at what third party data says about actual Facebook on-site usage, then at the idea that these users not visiting the site is a problem, anyway.

The article cites Nielsen, a well-regarded web measurement firm, to draw a contrast with Facebook’s own numbers. The filing said the social network had 161 million monthly active users as of December. Nielsen said 153 million unique visitors in the same period. From there, Sorkin goes on to guess that this difference might be due to the Like button and other off-site Facebook usage: “Assuming that Facebook’s United States traffic accounts for only about 19 percent of its business, that means the numbers are off by at least 40 million users from the 845 million Facebook defines as “active.”

First, Nielsen is just one data source, which itself disagrees with the numbers provided by competitors. For example, direct rival comScore showed that Facebook.com actually had 162.5 million uniques in December. Nielsen and comScore use similar types of methodologies, which involve doing things like tracking a sample of internet users, and there’s no reason (that I know of) to think one is more right than the other in this case. So if Sorkin went by comScore’s numbers, he apparently would have guessed that Facebook was actually undercounting site usage.

Another point on the comparison. Both of these companies track “unique visitors,” which are standard units of measurement that they separately define for all sites they track. The measure is a rough equivalents to the active visitors that Facebook tracks to Facebook.com, but we don’t know that for sure.

But Sorkin does have a fair point in noting the differences between the results. There are probably some users, especially the all-important daily active users, who don’t actually visit places with Facebook ads every day. ComScore provides worldwide Facebook numbers, and it shows 794.3 million monthly uniques and 297.1 million daily uniques. That could indicate Facebook’s monthly numbers are high by a relatively small 50 million MAU but a huge 186 million difference in DAU versus daily uniques. But that point comes with its own qualification: comScore may not be able to track Facebook data equally in every country based on local factors, like lots of users getting on a single computer at an internet cafe.

And there are also comScore data points in Facebook’s favor on this issue. There average Facebook user worldwide spent 11.6 minutes on the site per visit and December… and get ready for the kicker: visited 32.6 times. So if you’re an investor and you were worried that lots of users were clicking Like but not going to the site every single day, comScore seems to be saying that, well, they are visiting the site multiple times on some days even if they’re not on at every point of the calendar. This means they’d still be seeing a bunch of ads.

But this is just comScore data, possibly as right or wrong as Nielsen’s.

There’s a bigger point to Sorkin’s article, which he sort of addresses, which second-guesses the premise. Likes are actually quite valuable in and of themselves, because Facebook can use them to target ads, and provide the data to developers so they can build products that use Facebook to customize user experiences. Likes and other actions also generate content in the news feed that in turn makes the site more engaging. It’s hard to know exactly how valuable all that targeting and engagement activity is.

But the Like buttons and other web-focused products are only part of what could be going on with Facebook’s web-wide play. It could turn on an ad network or a payments system that is available across the web at some point in the futre. These possibilities have been speculated about for many years now in tech circles, and Facebook has tried to avoid saying anything definitive. But the idea of an ad network for publishers — like, uh, AdSense, which Facebook chief operating officer helped build in her previous job at Google — seems pretty straightforward. Facebook could sell ad inventory on other sites on behalf of publishers, and give them a cut just like Google does with its Adsense publishers. And on the payments front, Facebook could expand its Credits payment system to the web as well. In fact, it already has in the form of games like FarmVille.com, Zynga’s web version of its Facebook game. That game relies on Facebook as the login credentials, so it counts as off-site, but it also monetizes via Credits, so Facebook still makes money.

All in all, it’s reasonable for Sorkin to question the original numbers, and he might have a point. But third party data indicates that it might not be a meanginful one. And as far as this relates to prospective stockholders, Facebook is already monetizing traffic on the web via targeting, and could have ad plans for the future that would make this discussion moot.

[Top image via NASA.]