The Week When Digital Marketing Entered the Age of the Agent Layer
By Barry Levine, Marketing Land
It’s not just Facebook Messenger’s new bot platform, but that huge userbase bends the tipping point.
When the history of digital marketing is written, this week may be seen as the point when marketers went from delivering media to becoming mediators.
Facebook’s launch this week of its Messenger platform and bot ecosystem did not, of course, invent the category. It builds on what WeChat, Kik and other messaging apps have already launched successfully and arrives shortly after Microsoft’s announcements of a smarter Cortana and a new Bot Framework.
But when a bot platform launches with a user base of about one-seventh of the world’s population, it tilts the tipping point down into critical mass.
A tipping point not just for chat bots, but also for their wiser cousins, intelligent agents like Siri, Cortana, Amazon’s Alexa and Google Now. In short, users will increasingly expect that brand interactions will employ what might be called the Agent Layer.
“This is the first time we’ve seen something go from hype to reality this fast,” mobile marketer Michael Richardson told me, referring to “bots as validated by Messenger.”
CNN bots that distribute news through chat, Sephora bots to help you buy the right beauty products, Messenger bots for buying flowers and customer service bots for Sprint. Soon, there will be thousands of these utilitarian helpers, most likely managed by more intelligent master agents like Facebook’s emerging M virtual assistant.
And that means digital marketing is entering a new phase.
Up to now, digital marketing has largely been about the direct delivery of messages — website pages, push notifications, ads, content marketing, emails, even user experiences that support ideas the marketer wants to convey. While marketing agencies and brands like to tout their “conversations” on social media, they have mostly been shared posts and targeted tweets.
Now, much — if not most — of brands’ interactions with users will be mediated through automated operators like bots, virtual assistants and super-smart Watson-types.
There’s conversation… and conversation
Imagine that you walk into a store in a shopping mall, and you have to find everything on your own. That, digital marketing agency SapientNitro Vice President David Hewitt pointed out, has been the classic user experience for websites and many apps.
Now, chat bots — as well as chats with, you know, actual humans — have begun to populate websites, like helpful sales associates promptly showing up as you walk through the door of a store in a mall. It’s so much more convenient to ask a sales associate/bot what you really need, than having to figure out yourself what that is and where to find it.
Next, imagine that instead of walking into each store in a shopping mall, encountering a store-specific sales associate in each, you just walk into the shopping mall itself and are greeted by an agent who, when necessary, becomes or calls up the store-specific bot you need.
Of course, you can still visit any store, but this is a level of convenience that master agents like M and Cortana will probably provide and, one assumes, that will become expected.
Andy Shirey, Senior Product Manager at mobile engagement firm OpenMarket, told me he “completely [agrees that] user expectations will change” because of bots/agents. His company currently uses SMS and MMS bots to parse texts and issue prestructured replies.
But, he noted, keep in mind that there’s conversation… and there’s conversation.
“I have a conversation when I buy a (movie) ticket” at a ticket booth, he said, “but it’s worthless.”
“If I ask an opinion about the movie — that’s valuable,” he added. “It’s less about buying a ticket than getting information.”
He gave another example of how Home Depot might use bots. “I could say [to the online bot], ‘I want to buy lumber,’” Shirey envisioned. “That’s good for Home Depot, but not a good example of the power of bots.”
Discovering what you need
Instead, he noted, a user could tell the bot: “I need help building a deck.” The bot figures out what you need.
That’s a “tremendous opportunity for marketers,” he said, although “it’s less about traditional pricing and offers and more about the conversation around the value I can deliver.”
It’s also a new way of thinking for digital marketers: Imagine the ways your products can be used, and then create interactions through a bot/agent that support them.
And marketers may have to imagine interactions more subtle than “I need help building a deck.” SapientNitro’s Hewitt points to Sephora’s bot on Kik that launched recently.
Instead of a site that offers a product catalog, or even a bot that helps you explore a product catalog, Sephora’s bot asks the user to “tell a little bit about yourself, in a convenient way,” Hewitt told me.
“If this was a survey on a website,” he noted, “I’d skip it.”
“But because it’s asked in a conversational way, it doesn’t feel intrusive. [The bot] is helping you help me. It’s a lot more like a guided selling experience.”
With two or three questions, the Sephora bot determines what kind of makeup interests the user. This, of course, not only allows the bot to suggest certain products but also provides something almost as valuable as a sale: customer data.
“Conversation as the UI”
“My experience with [the Sephora bot] was a lot more natural than a lot of customer service reps, who are reading from a script anyway,” Hewitt said.
Aside from the challenges of designing marketing efforts that accommodate all the common scenarios of using a product or related products, or of coaxing mutually useful information from a would-be customer via an exchange of messages, the marketer’s new opportunity is succinctly summarized in one sentence in a developers’ document that Facebook issued on Tuesday:
“Think of the message thread as your app.”
Similarly, Microsoft recently proposed “conversation as the UI [user interface].”
Facebook, Microsoft and other platforms will template the bots and provide back-end services. The real app, then, is the give-and-take between users and the bot/agent.
This arguably requires a different kind of thinking than working with the flowcharts, storyboards, decision trees, user data maps and the other tools in digital marketers’ toolkits.
It requires understanding how to represent your brand, and your brand’s products, through the expressions of a virtual entity.
Just as we’re hardwired to recognize faces, we’re hardwired to characterize conversational exchanges. Expecially when they include — as they will — all kinds of media, including generated animation and video that depicts the bot/agent in real time.
Sometimes, the bot/agent will be a clear and intended personality, like Imperson’s Miss Piggy chat bot for Facebook Messenger.
At other times — even if it’s a utilitarian bot for helping you order flowers — the message thread itself will have to be tuned for its tone, its manner and what it implies about the brand.
Is Sephora too nosy about what kind of makeup the user wears? Does Home Depot act like it understands I really don’t have the slightest idea how to build a deck?
Since at least the time of “Mad Men,” brands have strategized about their personalities. But customers’ frequent interaction with a layer of bots/agents means that marketers will have to strategize about the kind of presence the brand should present.
In a way, the Agent Layer means the return of the sales associate as the face of the brand. Except that now, the marketer is inside.