Facebook may be killing its virtual assistant, but not all brands have given up on the platform’s bots.
This week’s announcement that Facebook is dissolving its M platform — a virtual personal assistant for messenger, designed to complete tasks like booking appointments and ordering products — in many ways mirrors the difficulties afflicting the retailers that are testing bots. Ultimately the social platform found M was too complex to manage, plus it struggled to automate certain actions, a challenge retailers are also experiencing as a result of a lack of backend resources. Despite this, many brands continue to trial the technology in an attempt to not just tout their tech capabilities, but also drive consumers to their e-commerce sites and increase sales.
However, much like the demise of M, many retailers, in the rush to implement the technology, are failing to establish the appropriate resources needed to maintain them, said Maya Mikhailov, CMO and co-founder of mobile commerce platform GPShopper. It doesn’t help that messenger bots are also increasingly facing competition from other automated services, like voice-activated style programs from Amazon Alexa and Google Home. (Perry Ellis, for example, became one of the first retailers to test voice commerce in September, adding yet another channel for commerce.)
“It’s not as easy as just ‘insert chatbot here.’ Chatbots were initially positioned to retailers and brands as a magic bullet, where all you needed was a bot,” she said. “As retailers think about what they’re going to do with bots, they need to remember it needs to be usable and scalable. Bots aren’t a bandaid solution for customer service solutions.”
Despite publications like Wired going so far as to call M’s demise a harbinger of the end of bots, Mikhailov said she doesn’t believe bots are on their deathbed. However, she said brands and retailers need to develop a better understanding of how automation and virtual assistants operate. The best use cases for retailers is still in customer service related to inventory and delivery status, she said, because it’s relatively easy to program and requires less resources than other capabilities.
Ad position: web_incontent_pos1
“If the other side of your bot is primarily a human being executing a task, then scalability of that bot is going to be really challenging, because you still have to throw bodies at it,” she said. “But that’s not the promise of bots; it’s that they are attached to some sort of intelligence and automation, and not that it’s just a chat interface.”
Though Facebook chatbots continue to present several challenges for retailers, they haven’t prevented brands like Louis Vuitton, David Yurman and Levi’s from continuing to experiment on Facebook.
Carey Krug, CMO at David Yurman, said the brand’s decision to launch its holiday Facebook bot late last year was inspired by high engagement rates among its consumers on Facebook. She said five percent of the bot users clicked over to the website, and more than 50 percent clicked on the “browse for gifts” option that featured products within the chat.
Ad position: web_incontent_pos2
“We have a very loyal and engaged Facebook audience, and we know we need to be where our customer is. They are definitely interacting with us and converting on Facebook,” she said. “We also like to test new technologies, especially when we know our customers are early adopters.”
Eitan Sharon, founder and CEO of mode.ai, the company that built the Levi’s and Louis Vuitton bots, said that although mode.ai is platform agnostic, Facebook messenger remains the primary draw for retailers. While Sharon was not able to provide sales or conversion numbers, he said Louis Vuitton and Levi’s both found success on the social media platform, because it was focused on styling, rather than the advanced options available to users on Facebook’s M.
“With engaged users on the bot, not only do they buy, but they buy more than they do on traditional channels, like apps,” he said. “Engagement is significant. On average, users are spending double-digits minutes on bots.”
Still, looking to the future, Mikhailov said the end of programs like M can serve as a time for retailers to reassess their goals with bots and how they aim to use them to connect with consumers.
“The chatbot hype has almost reached its peak, and now we’re seeing the trap of disillusionment, when it comes to the hype cycle of chatbots,” she said. “Personally I think that’s great, because we need to get beyond the use of chatbots as a gimmick and make them useful to retailers in automating functions.”
