Good things come to those who wait, but sometimes you have to ride out rainstorms to get your rainbows. In 2011, Facebook launched its Messenger platform and people gladly hopped in, forming an active community of around 600 million chatty humans by 2015. In April 2016, the social network opened Messenger to chatbots, conversational software programs driven by algorithmic scripts.
The announcement triggered much hype, but early chatbots bombed in the arena of public perception. Pundits were divided into those who graciously lauded chatbots’ “potential” and those who just couldn’t ignore the disappointing experiences most chatbots delivered. The headlines read “Please, Facebook, don’t make me speak to your awful chatbots” (The Guardian); “Facebook’s new chatbots still need work” (TechCrunch); and “Facebook Chatbots Are Frustrating and Useless” (Gizmodo). Keep in mind that Facebook builds bot developer tools, but not the majority of end-user bots on Messenger, so perhaps you can split the blame with the brands and third-party developers who do.
The first-generation chatbots on Messenger included services that gave updates about the weather, took flower or food delivery orders, and sent airline travel documents to your phone. Those were the good ones. The rest stoked a burning wish that you were speaking to a human customer service agent or just using a regular website or app instead. Amid the negative publicity, even David Marcus, Facebook’s VP in charge of Messenger thought chatbots were overhyped, describing to Business Insider the stark similarity of crappy websites and annoying mobile apps in the early days of the web and the smartphone ecosystem.
But, the conversation has shifted since those early days. Nearly 1.2 billion entities talk on Messenger now and not all of them are humans. More than 100,000 conversational experiences have been built by a thriving developer community for Facebook’s Messenger platform. Every month, these chatbots generate 2 billion messages with the people they engage. Many of the bots have demonstrated cost-saving capabilities for the businesses they serve while others – such as SnapTravel – reel in significant topline revenue.
The Key Drivers Behind Messenger’s Turnaround
We talked with Anand Chandrasekaran, Director of Platform and Product Partnerships for Messenger at Facebook, to learn how the platform has evolved. Chandrasekaran attributed the recent success of chatbots to three key driving factors:
- Availability of design elements that support rich conversational experiences
- Enhanced service discovery and entry points for consumers
- Clear paths to monetization for developers, brands and businesses
Thanks to a lively feedback loop between Chandrasekaran’s team and the developer community, Facebook added a number of rich visual elements to the Messenger platform to improve consumer experiences. “When I first started talking to bot developers, they told me that a lot of e-commerce and retail is visual,” he explains. “I want to know the shade of blue of the dress I’m buying or the color of the roses I’m buying. Just talking about them via text isn’t enough.”
The need for visual support led Facebook to quickly add rich media and webview interfaces for Messenger bot developers. One of Chandrasekaran’s favorite bots, Epytom, uses rich imagery as well as machine learning and natural language processing (NLP) to give style tips and suggest new outfits and purchases.
The 80+ bot developers Chandrasekaran talked to also highlighted critical challenges with discovery. Interacting with Messenger bots usually require that you access them via a brand’s Facebook Page, but unless you’re already both familiar with both the bot and the brand, you might not think to hit that “Message Us” button. Even when a user is already interacting with a bot, the conversation often gets lost in Messenger history because they’re mixed in with conversations with friends and family. Users who couldn’t easily engage a bot where they left off often gave up in exasperation.
“We knew discovery was important, but talking to developers alerted us to the urgency of solving this challenge and also gave us insights into the type of discovery people wanted,” reflects Chandrasekaran. To solve the issue with losing track of bot conversations and discovering new bot experiences, Facebook’s Discover Tab keeps recently used bots up at the top of the screen and also suggests featured bots to check out.
Another discovery mechanism is what Facebook calls “M Suggestions”, or the in-line contextual pop-ups that suggest useful bots or built-in functionality. For example, saying you’re “on your way” pops up options to share location with your conversation partner. You may also have seen pop-ups from Lyft or Uber that enable you to order a car in-line.
Finally, many successful bot use cases are inherently social, such as when you’re shopping for concert tickets, new outfits, or making travel plans. Chatbots that are commonly used socially are included in the Messenger UI’s chat extensions, right next to where you type messages and upload photos.
Now you can easily pull bots like Fandango into a Messenger thread if you want to discuss movie times with friends, or pull SnapTravel in to find hotel deals together.
Paths To Growth & Monetization Materialize
Early adopters saw the potential of chatbots to encourage desired business activities such as driving video consumptions, offline footfalls, and recurring subscriptions. Those opportunities to monetize are now being proven out.
Hotel booking startup SnapTravel recently attracted $8 million in Series A funding. The company implements a human-bot hybrid approach to enable users to locate and book rooms via SMS text and Messenger. According to Chandrasekaran, SnapTravel rakes in $1 million in monthly revenues from Messenger alone.
Meanwhile, brands such as KLM, T-Mobile and Sephora have deployed Messenger chatbots to ease customer congestion and provide seamless brand experiences. Chandrasekaran reports that Sephora drove 11% more conversions via chatbot than other digital channels. In addition to driving increased revenue, Juniper Research projects that chatbots may cut business expenses by as much as $8 billion by 2022.
Some companies, such as Epytom the chatbot stylist, have grown their business almost entirely on the back of Messenger. According to Chandrasekaran, the fashionable chatbot receives 5 million messages a month. Rather than selling users new clothes they don’t need, the bot focuses on showing you how you can re-style your look with the typical pieces you likely already have in your wardrobe.
Chatbots have been a bit slow living up to their hype, but they’re definitely doing it in style.