Following Facebooks IPO a number of documents have been released covering the misgivings the Security and Exchange Commission (SEC) had around their user data, in particular around mobile devices. These have been expertly summarised in this fascinating story on Bloomberg. This is not just important insight of investors; it is critical knowledge for marketers that are building businesses on Facebook and for our perception and application of analytics and social listening to developing strategy.
The key points from the article include:
- They critically had very little idea how many mobile users they had and were double counting users
- Facebook had no concept of how many mobile only users there were. They estimated that 69 million, or 44 percent, of the 156 million new users might be mobile- only. But these were finger in the air numbers.
- They had no reliable count of the geography of mobile users. For instance, the company counted as Canadian many BlackBerry (RIMM) users around the world because the servers are based in Canada.
This is of course an issue for Facebook as the value proposition is based on the data. But it’s more important for the marketing industry as whole. Whole agency models are built on the validity of Facebook and publisher data.
Exploring their recent numbers
This also brings into question their recent 1bn active user claim (600m mobile). No question that Facebook is still growing, we’ve tracked over 1bn accounts and 645m active users (from 87% of the global internet population) as of Q2 2012, and forecast 2bn (thanks to emerging market growth) by the end of 2014.
Crucially we track Facebook as having the highest level of active usage (defined as contributed or used) versus number of accounts outside of Chinese services, which stands at 64%. This it should be noted is an incredibly high figure.
If we take the 1bn active users as gospel truth, this would suggest they have 1.56bn accounts based on our active user percentage. If we also buy into the claim that they have no users in China and exclude China, the global internet universe now stands at 1.7bn. This would mean that 89% of internet users on the planet have a Facebook account.
This clearly implies a number of things:
- The real active number is much lower (but still hugely significant)
- Mobile users are being double counted
- To generate such scale there would have to large number of users in China (being registered as other markets in official numbers – mainly in the US)
- Users are being recorded as active without their participation or knowledge (i.e auto logins in via mobile)
- Duplicate and fake accounts are a very real issue
The wider issue with analytics
This is not an issue that is going to go away and it reaches far beyond Facebook. As consumers become more multi-platform, the task of tracking users with reliability becomes near impossible. Add to this growing user awareness of tracking, initiatives like Microsoft’s “Do Not Track”, the shift of internet behaviour into apps, and you have an increasingly murky future if you entirely rely on analytics and tracking to drive insights for strategy development or planning.
This is not a commentary on optimisation or campaign reporting (we use analytics here for our marketing), however the simple truth is that relying on web and social analytics to solely drive business decisions is increasingly dangerous. We believe research provides the most accurate and detailed view of how people are behaving online. These insights back this up.
It also underlines why we exist and why we brought GlobalWebIndex to market and why through client adoption we’ve been able to create the largest multi-market syndicated research service on the planet.