Is Facebook Use Declining in North America?

Before anyone gets their panties in a bunch, this post is meant mainly to ask the question: is Facebook still actually gaining new users? And if so, how many of those are really in North America? I've been trying to make sense of their user growth lately to hopefully find an area where the actual data given explains the empirical data I see around me... that no one who doesn't have a Facebook is just now making the move to get one, and if anything it's more long-time users leaving. More from my blog: huntingforcommoncents.wordpress.com

Facebook recently announced its Q4 earnings and gave some insight into how the company is faring in the often talked-about mobile world, the Facebook Gifts space, and other tidbits that gave us a look into the growth trajectory of the company in the next year or two. Facebook reported revenues slightly above expectations (1.58B vs 1.52B estimate) and earnings of $0.03 (or $0.17 on an adjusted basis, excluding stock-based compensation), but what people really care about is how well mobile is doing.

Reason being, mobile is the highest growing segment of Facebook users, yet only 14% of revenues had come from mobile ads last quarter. Fortunately for Facebook, this quarter they announced that mobile’s contribution to revenues increased to 23%, rightly in line with what many in the market expected. However, why is the important number the % contribution rather than the absolute number (or even the % growth) of mobile ad revenues?

Looking at it that way, mobile ad revenues just about doubled, or grew somewhere in the neighborhood of $160M to $320M, rough estimate. Anyway, it looks like Facebook has at least started to find a way to grow their mobile revenue… even though half their daily and monthly active users (DAUs and MAUs, for short) are on mobile. Put plainly, they need to eventually see that mobile revenue number in line with their user mix.

But the purpose of this writing is not to postulate on what Facebook is going to be able to do with mobile revenue in the future. What I’m interested in here is looking at how Facebook is actually representing their user base, what it means, and how that may affect the company once other people actually figure it out. So let’s do that.

From their Q4 presentation, here’s a graph of the monthly active user (MAU) growth over time, meaning how many people log in to Facebook at least once a month.

Most of the growth here is obviously coming from the East, and if we look at the dark blue at the bottom (US & Canada), we see the growth has more or less slowed to a halt by Q3’12 and a slight uptick in Q4’12 but still pretty flat. While there’s still growth in the rest of the world, Facebook has very little growth in the areas where people are most likely to actually respond to their ads. But growth nonetheless. Or is that growth possibly not even real?

Facebook has some cute fine print at the end of its presentation that doesn’t command much attention, but should be considered by the inquisitive investor. Let us indulge.

This first slide is telling us that Facebook recognizes that many user profiles are for non-real and non-unique people, examples being Pets, fake celebrities, or simply made up people… Lennay Kakua, anyone? Facebook estimates this number of users to be about 5% of the population, and thus it’s possible MAUs are inflated by about 5%. If you believe it’s only 5% I have some oceanfront property in Kansas to sell you, but let’s humor them and say it is.

Now before I go any further, we need to take a look at a bit of Facebook’s history regarding fraud accounts. Facebook reported total MAUs for fake users to be estimated at about 5-6% on total MAUs of 845 million as of December 2011, repeating this estimate in May 2012 when it filed its s-1. By the time they disclosed again in June 2012, they estimated 8.7% to be fake on an increased total of 955 million MAUs. On an absolute basis, that means the fake accounts number had went from 42.25M to 83.1M, or a 64% to almost 100% increase. Keep in mind that if Facebook actually cut out their estimate of fake users from their reported MAU figures, the growth in fake users would be astronomically higher – in the 170% to 290% range. What’s clear from the past, no matter how you look at it, is that Facebook had previously under-reported the fake user estimates… or fake users are growing dramatically quicker than real ones.

Let’s pretend for a moment that these numbers, put face to face with the possibility of fake users outpacing real user growth, are legitimate. So our US and Canada demographic saw quarter-over-quarter MAUs increase from 189 to 193, representing a 2.1% increase. Net of Facebook’s “fake user” estimate, this would be a 2.9% decrease in real-people MAU growth in the US & Canada. In Europe, a 1.9% decrease. In Asia, only a 2.5% increase, and only 0.5% in the rest of the world. In aggregate, this would suggest a slight 0.2% decrease in MAUs coming from real users, for a company that the Street expects to grow revenues in excess of 30% a year… forever. For reference, the actual numbers for real MAU growth were about 8.6% last quarter, and 4.6% today.

The MAU decline, net of fake users, could actually be much harsher if Facebook was just playing fast and loose with its estimates. In the prior quarter, the estimate of 8.7% fake MAUs was reported higher than previous on higher growth. In the quarter just reported, the number dropped to what appears to be 7.2%. Facebook made this downward adjustment in fake and wrongly-classified, but actually increased duplicate accounts slightly. This would suggest total fake MAUs went from 87.6M to 76M, or about 14.5% less. This could be thanks to some sort of amazing effort by Facebook to “clean up the streets,” so to speak, or it could simply be a more lax estimate by Facebook, who would like to believe most users (particularly coming from the East, where most growth is anyway) are not only real people, but are mostly using their classification system correctly. Perhaps I’m too cynical but why wouldn’t a company who sets up ridiculous press conferences to announce its new stalker technology not be inclined to brag about eliminating almost 12 million fake accounts? Just saying – on an 8.7% estimate our new MAU number for fakes would be about 92M.

Given it touts over a billion MAUs in total currently, it’s not unreasonable to suspect a good portion of Facebook’s MAU growth is no longer coming from real ad-viewing people… and likely something closer to 10% of MAUs are currently fake users. Companies, brands, pets, duplicate profiles, etc. are all very common no matter where you are, and given the increasing commercialization of social media and Facebook’s history of upward revisions, that number is likely increasing faster than unique user growth. At some point this needs to translate to the realization that Facebook is done growing organically and could slowly be in decline.

Now let’s look at one other issue surrounding user geography and access methods which could be even more troubling.

The gist of this disclosure is that Facebook’s servers have limitations on identifying exactly where their website was accessed from geographically speaking, and also in identifying the authenticity of its mobile (and even desktop) user base. First, Facebook discloses that mobile devices with the Facebook app installed will frequently ping the FB servers to check for updates, even if the user doesn’t use Facebook or even have a profile anymore! Without getting too in-depth, additional research leads one to suspect that the mobile DAUs Facebook reports are also inflated significantly. Not only does that push the Street to be more upbeat about FB popularity and the future of mobile, it could result in justification for charging more to mobile advertisers.

In the bottom portion of the disclosure, Facebook mentions that “We estimate that the number of MAUs as of March 31, 2012 for the United States & Canada region was overstated as a result of the error by approximately 3% and these overstatements were offset by understatements in other regions.” Tacking on an additional 3% in MAU inflation, that would bring the number of MAUs in the US & Canada even lower, further increasing the growth of fake MAUs on a % basis if we take Facebook’s word for it and assume less-developed countries have more fake accounts (I question this, but it works to support me here). While this means there is some understatement in the numbers of other regions, those are regions where revenue per user is going to be lowest (and even lower after adjusting for these under/overstatements), so the net outcome is that Facebook use is dying in the developed world, and at a likely mid single-digit decline in real users in North America, it’s only the tip of the iceberg, and will result in lower ad revenues as time goes on. Even in the most conservative circumstances, Facebook growth is clearly overstated. Another piece worthy of note in the disclosure is that Instagram users who log in using their Facebook information (something that Instagram encourages), every picture they post will leak through to the FB servers, registering them as a DAU or MAU depending on their frequency. With all this extra counting, it’s clear that these numbers have to be much higher than the actual. And I don’t know about you, but I don’t have too many friends who are telling me they just got a Facebook – if anything, I’m seeing quite the opposite.

So that’s all from me for now. Chew on that a bit and feel free to leave a comment or two with your thoughts on the matter. Do these numbers make sense? Is Facebook still all the rage, or is it dying? And how would this end up affecting ad revenues as more and more shift goes to mobile?