With 1M apps to date, Facebook’s analytics tool adds features but hasn’t fixed measurement error
Developers can compare two audience segments at a time and break down their audiences based on the sites that sent them.
Facebook’s rival to Google Analytics has piqued developers’ interest, though it’s unclear to what extent it’s been able to sustain that interest. And at the same time Facebook adds new features to its measurement tool, the company has not yet fixed as measurement error disclosed last year.
More than 1 million mobile apps, websites and Messenger bots have used Facebook’s Analytics for Apps at some point since Facebook rolled out the free tool in March 2015. That means 200,000 new apps, sites and bots have used Facebook’s analytics tool since September 2016, when the company said that more than 800,000 apps had used it since its debut 18 months earlier.
But both stats are measurements of lifetime, not active, usage. That makes the million-plus mark a nice milestone but more of a signal of interest than sustained usage. The growth in interest between September 2016 and now — with 200,000 new apps using it at some point in that period — likely has to do with Facebook extending Analytics for Apps beyond mobile apps to also measure audiences for websites in September 2016 and for Messenger bots in November 2016. Developers have to actively implement Analytics for Apps into their mobile apps and websites, but Analytics for Apps requires no setup for Messenger bot makers and is immediately available, which could have boosted its usage. As of November 2016, there were 34,000 Messenger bots on the market.
While it’s unclear what kind of sustained adoption Facebook’s Analytics for Apps has received among developers, it’s more clear why it would attract at least initial interest. Like any analytics tool, Facebook’s product reports stats about how many people are using a developer’s app, site or bot and how they’re using it. But Analytics for Apps also provides information about who those users are, at least for those users who are also Facebook users. By incorporating Facebook’s own anonymized audience data and the anonymous data Facebook sources from third parties like Oracle and Acxiom, developers can see information like what Facebook Pages their users like, their estimated annual household income, what city they’re in and their job titles.
In addition to announcing the million-plus milestone, Facebook is adding two features to Analytics for Apps that aim to give developers a better grasp of who their audiences are and how those people are coming to their apps, sites or Messenger bots.
First, developers will now be able to compare two audience segments side-by-side for some measurement categories. Developers can already create audience segments that break down their user bases based on categories like age, gender, the city they’re in and what type of device or operating system they’re using. Developers can then see information in Analytics for Apps specific to these audience segments, like how many people using iPhones and iPads visited an e-commerce checkout page or the average size of their shopping carts. But developers have not been able to see this information for more than one audience segment at a time, so an app could not easily compare purchase activity for people using its iOS app versus people using its Android app. Now they can, though developers can only compare two audience segments at a time, and for now, the comparisons can only be made when viewing events, or actions that people take in an app, on a site or with a bot.
Second, developers can create audience segments that take into account which third-party sites directed people to their mobile apps, sites or Messenger bots. That means a developer would be able to learn more about the people visiting its site or app from Google.com, for example, based on those people’s Facebook profiles. And with Analytics for Apps’ new comparison option, the developer could compare that audience with people coming from Bing.com to see how many sales a site or app generates from the respective search engines or the average size of their respective shopping carts. Then the developer could then use that information to inform the app’s or site’s search strategy.
At the same time Facebook adds new features to its analytics tool, it has still not fixed a three-month-old measurement error. In November 2016, Facebook disclosed that it had incorrectly counted referral traffic from Facebook to developers’ apps and sites, resulting in a 30 percent inflation on average. At the time, the company said it was working on fixing the error, and a Facebook spokesperson said the company is still working on a fix.