M Recommendations, or How Messaging Replaces Search

Amanda Felson
Assist Blog
Published in
3 min readSep 27, 2017

--

Yesterday, Facebook added 3 recommendations to M, Messenger’s virtual assistant. M will now prompt users to share gifs, use pre-populated Quick Replies, and summon Fandango’s Chat Extension to view trailers and book movie tickets.

Chat Extensions allow users to launch and share branded experiences within the Messenger platform. They appear in Messenger’s Composer drawer, and clicking on a Chat Extension launches a webview with content or additional functionality that can easily be shared back into a message thread.

Image credit: Facebook

The announcement adds Fandango to a handful of brands (including Spotify, Food Network and Lyft) at M’s virtual fingertips, offered into message threads when prompted by keywords like “movie.” The steady growth of M’s roster hints at a not-too-distant future where thousands of brands move in and out of Messenger conversations, their Chat Extension a key entry point for discovery.

M + Chat Extension = Brand Discovery
Historically, consumers have found brands on Messenger through 1 of 3 methods:

  1. Messenger’s search function
  2. Messenger’s Discover tab
  3. Branded marketing (e.g. Promoted Post with Messenger code.)

The combination of M + Chat Extension is an important 4th discovery tool because it inserts brands into message threads when users might not otherwise know to engage them. I may have no idea I can chat with Food Network until I reference “recipe” and see an M recommendation — but their Chat Extension fundamentally changes the way I think about Messenger’s ability to offer me relevant content at exactly the right moment.

M + Chat Extension = Messaging is the new Search
The impact M’s recommendations, though, extends beyond just discovery: M + Chat Extension connected me to Food Network at a time when I might have left Messenger and initiated a search. With M + Chat Extension, brands harness a conversion tool that connects them to users while they move through the conversion funnel. Just as Google built a business monetizing keywords that suggested intent, M’s recommendations (triggered by keywords) seem to hint at future opportunities for paid brand promotion, and Facebook monetization.

And let’s not forget, too, that this comes with a better user experience. For example:

Without M Recommendation or Chat Extension:

  • I kick off a Messenger thread about seeing a movie on Saturday night
  • A friend suggests we find showtimes for “Kingsman: The Golden Circle.” She searches for the trailer and shares the top result — a YouTube link — into the thread
  • I click on the link which launches the trailer in my YouTube app
  • I search for showtimes and types our options into the thread. We agree on a time and destination.
  • I launch my browser and use search to navigate to Fandango.com to purchase tickets

This conversation requires us to navigate between several mobile apps and web pages, and to copy/paste information into the message thread. It’s an awkward, fragmented experience that requires me to search and chat simultaneously to complete my purchase alongside a group.

With M Recommendation and Chat Extension:

  • I kick off a group chat about seeing a movie on Saturday night
  • M recommends that I “Find Showtimes” in Fandango’s Chat Extension, where I can watch the trailers and find local theaters
  • I watch the Kingsman trailer, share showtimes into the message thread and we agree on a time and theater
  • We each complete our purchases from Messenger’s webview

This group conversation took place exclusively within Messenger and its webview, with no need to search or launch other apps to retrieve information. M’s recommendation simplified the transaction process, and reduced the likelihood of our abandoning purchase. And, naturally, conversion is easily attributed back to Messenger: everyone wins.

If you are interested in automated solutions for messaging or voice, please send us a note: hi@assi.st. Learn more about Assist here: http://assi.st

--

--