Almost one year in, Facebook’s Messenger bot platform is still looking to find its groove.
The Information (paywall) has exclusive information on what Facebook is working on for the next iteration, which will most likely ship or at least be announced around April’s f8 developer conference:
After two years of trying to figure out ways to get commerce flowing through Messenger, with little success, Facebook is taking a new tack. Messenger is trying to find ways to inject buttons into conversations between friends when it thinks they intend to do things like make restaurant reservations, buy tickets to something or order food.
The buttons would give users a way to act on those ideas, by linking up with an outside company to buy something or make a reservation. Currently, users on Messenger have to search for businesses, which mostly operate on the app through automated bots. Those bots, as well as Messenger’s personal assistant M, have had difficulty fulfilling requests from people without the help of humans. […]
What The Information describes makes sense for Facebook but one wonders how this -privacy invading- feature will be advertised to users:
Driving the new approach has been a new version of M—called “omni-M” internally, which is now being tested among the small subset of users with access to it. It is designed to pop into chats between friends to try to understand what people are trying to do. The effort, run primarily out of Facebook’s applied machine learning division led by Joaquin Quiñonero Candela and product manager Ragavan Srinivasan, has gotten higher priority in recent months, said three people involved in those discussions.
According to The Information’s sources, this new strategy is still early in development. Which means a lot can happen until it will ship.
To make discoverability and on-boarding and switching between e-commerce services easier, Messenger will handle payments itself. Facebook Messenger launched this in September 2016 for its then 30,000 chat bots:
To support payments in Messenger, Marcus says that the company is working with all the major players in the industry including Stripe, PayPal, Braintree, Visa, MasterCard and American Express — not just Stripe and PayPal which the Facebook developer blog post mentioned.
Separately, leading up to F8, executives at Messenger have also told developers to prioritize commerce-driven experiences to test out new “native payments” systems that let users quickly buy things without having to enter their credit card information. Some developers have tested native payments and it’s expected to roll out more widely by F8 or later this year. So far, Facebook has given some brands free credits for ads in News Feed to drive users to chat with them in Messenger.
Wether -and how- this all will work is still very open.
The Information on another roadblock:
One potential problem is whether people will drop enough hints in Messenger conversations for the approach to work. Messenger employees have researched in past years how much “explicit intent” is evident from conversations between people on the app, like when users tell friends they want to grab a ride or order food. Those instances weren’t as frequent as the Messenger team expected, one person familiar with the tests said.
In the meantime, Facebook Messenger is slowly but surely going away from decision trees, aka bots within the traditional chat interface, and towards webviews. (Like WeChat)
Mikhail Larionov, Engineering Manager at Facebook’ss Messenger platform wrote about the thinking behind this in an article on Medium:
If the people using your bot lose their previous selections and have to start over, that negatively impacts your conversion rates and makes the overall experience subpar. It also illustrates that if a tree is longer than 2 steps and people want to revisit an earlier step in the flow, a tree may not be the best choice.
In general, even having a “Restart Flow” button in the menu is a good indicator that things will go wrong. Consider using a “Change Your Selection” button instead — a button that simply invokes a webview that displays their previous choices and gives them the opportunity to make edits. […]
If you’re still debating between decision trees and webviews, here are two main use cases where webviews usually make more sense:
Rich browsing experiences. Your content can be more engaging in a webview, incorporating larger images, additional scrolling, and specifically structured data.
Entry of multiple pieces of data. Here webviews are the only feasible way, as cost of reentry and inability to change previous selections with decision trees results in a really poor UX.
A lot of my reservations regarding bots and conversational commerce go away with rich, visual webviews. (And no, this, then, is still not the same as an stand-alone app or a website. Runtimes matter.) In the end it all comes down to the right interface for the right shopping situation. Pure decision trees make only sense for a very narrow subset of shopping situations. For those they are perfect, for everything else they fast become cumbersome, at the least.
One can say a lot of things about Facebook Messenger, but it certainly is not a niche product, nor does it want to be a niche platform.
Hence the push to build a platform that serves as many needs as possible.
- At Least 27 Shopping Bots in the Wild, Germany based chatShopper a User Favorite
- How BuzzFeed, Time Inc. Are Using Facebook Messenger Bots to Create Gift Guides
- The bad User Interface of Today’s Facebook Shopping Bots
- First Bots on Facebook Messenger are Like First “Social Shopping” Pages on Facebook
- Zalon Messenger and the Question about Dedicated Apps in Conversational Commerce
- The Sweet Spot for Conversational Commerce
- Conversational Commerce: How Messaging Will Enrichen E-Commerce