Facebook launches analytics for Messenger chatbots
Facebook is addressing one of the biggest missing pieces of its chatbot offering: analytics.
Facebook is addressing one of the biggest missing pieces of its chatbot offering: analytics.
In April, Facebook opened up its Messenger app to brand chatbots, and recently took steps to monetize Messenger by allowing brands to pay for ads that aim to drive more interactions with those brand chatbots.
Now, it is addressing one of the biggest missing pieces of its chatbot offering: analytics.
Yesterday, Facebook announced that it is extending its Facebook Analytics for Apps tool to Messenger so that brands jumping on the chatbot bandwagon can access insights about the Facebook users who are interacting with their chatbots and better understand those interactions.
Without adding additional code, brands that have built chatbots using Facebook’s Messenger Platform will be able to track a variety of metrics related to Messenger chatbot activity, including messages sent and messages received, as well as user blocks and unblocks.
Additionally, Facebook is offering brands access to reports that provide aggregate, anonymized demographic data of users who interact with their chatbots. This data includes age, gender, country, language, education and interests.
Brands that run multiple Facebook Pages can filter their reports to view analytics data associated with a specific Page.
Those that want to define their own metrics can also do that by configuring custom events.
For example, according to Facebook software engineer Sridharan Ramanathan, “a travel business can see how often people are transferred to a human agent” or “an ecommerce business can build cross-platform funnels to see what percentage of people interact with its bot also make a purchase on its website or app.”
While Facebook’s Messenger Platform has been used to build tens of thousands of chatbots since its debut earlier this year, the jury is still out on just how productive they will be for brands. Gizmodo’s Darren Orf, for instance, has called them “frustrating and useless.”
Making them less frustrating and more useful will in many cases require chatbot developers to employ better natural language technology, but analytics could help brands better understand where their chatbots’ interactions with users are falling short.
It could also help them identify the user groups that they should focus on building chatbot experiences for.
In some cases, using analytics and making revisions informed by analytics data could be the difference between chatbot failure and success as brands invest in this new space.