How Business Leaders Leverage Chatbot Personalization
See how major brands are outsourcing customer service work to chatbots while creating a more personalized online experience for their users.
June 21, 2018
Chatbots, or bots, can help brands realize personalization and interaction goals with their customers better than a website alone. In this article, we explore how brands are successfully using bot personalization for customer engagement.
Why marketing personalization?
Marketing personalization means tailoring messages and interactions to individual consumers, based on their interests and needs. Until now, personalization has been primarily driven by the availability of big data that allows brands to construct profiles of individual customers.
Bots, however, present another avenue to personalization—one that doesn’t require the availability of big data to provide customers with tailored messages, because the bots can respond in real time to the data as it’s presented. Not only do bots reduce the need for on-hand, live customer support, but they also tap into online users’ need to interact rather than to passively receive information. They also reward users’ desire for immediate gratification.
How bots break open personalization
Imagine some possibilities:
Your favorite restaurant has a bot you can chat within Facebook Messenger, and this bot remembers your most frequent orders. When you start a conversation to order takeout, it asks whether you want one of those same menu items.
A shoe store bot greets you and asks your shoe size, along with the style you’re looking for. You browse the suggestions it gives you and comment on particular products so the bot can give you increasingly relevant product options.
You start shopping on a clothing brand's site by telling a bot what your body type is and whether you need outfits for a specific occasion (such as a wedding or job interview), and then you receive suggestions based on what is currently in stock and in fashion.
A grocery store bot asks you what you want to cook this week and then directs you to recipe ingredients that are on sale. Or the bot sees what’s in your cart, knows what you usually buy, and suggests recipes, based on your ingredients.
These possibilities are not yet realized, but the technology and data access are available to create these types of bots. Here are some of the ways that brands are currently using bots to deliver personalized experiences on their websites and on other platforms.
Alaska Airlines: Jenn
Goal: Help users find information without searching or contacting support.
Jenn is a customer support bot, prepared to answer certain questions but not equipped to help with all requests. To see what the Alaska Airlines bot can do, we asked, "Where can I fly direct from Seattle?"
Jenn had Alaska’s website redirect to a route map page and a couple of other links she deemed useful. When we asked about the baggage policy, we had a similar experience. We were redirected to the relevant site page and given an answer in the chat window.
Goal: Give users makeup tips, recommendations, and product reviews, based on their personal preferences and face shape.
In 2016, Sephora launched a bot on the Kik platform, which has been a popular messaging app for its US teen audience since 2010. A hybrid salesperson/personal assistant, the bot starts off with a short quiz to learn a customer’s age and makeup preferences. From there, it provides product recommendations and educational content (such as a tutorial on face contouring) in a blend of sponsored and non-sponsored content.
Much of personalization is about being where your customers are. For Sephora, that meant using Kik and launching a bot shop in 2016. The brand understands that bots bring value because they give customers a way to reach out to brands on their platform of choice and to immediately engage.
Goal: Gather opinions from people in countries around the world on domestic concerns.
Marketing is not only for direct-to-consumer brands. UNICEF is creatively using the technology with its U-Report bot, which gives residents of many countries and regions an opportunity to tell UNICEF what’s on their minds. The organization then uses that data to increase citizen awareness of local matters and to inform policymakers on the issues about which their citizens care most.
In the US, UNICEF recently used a bot to poll citizens about matters related to internet safety for Safer Internet Day 2018. As responses come in, you can see the results and statistics on the website.
The U-Report bot uses Facebook Messenger and Twitter to communicate, and it boasts close to 5 million users worldwide. It’s a stellar example of an outbound or data-gathering bot. Users opt into the communication and then periodically receive informational requests.
U-Report was explicitly designed to give young people a voice in civic and political concerns. As with Sephora, UNICEF understood that in creating this tool, a brand has to go where its audience is and it needs to invite them to engage on platforms of their choosing.
Goal: Make it easy for users to find insurance based on their needs and location.
Progressive’s home page features a conversational interface (CI) to engage site visitors. Conversational interfaces and bots share a common goal—better user experience through real-time interaction. Progressive’s website bot demonstrates a sales use case.
Many bots rely on natural language processing (NLP) or programmed scripts and are designed for more open-ended interactions. CIs, on the other hand, ask questions and provide a few options to help get users what they want, with less effort.
The upside of a CI is that it quickly gives users the answers they’re looking for, with little room for the misinterpretation that might come with a bot. The downside is that there’s less flexibility. In a conversational interface, users are given predetermined options from which they can choose, and these options have associated answers or next steps. They can help with clarity, but sometimes a user has a concern that isn’t addressed in the presented options. When this is the case, the CI isn’t useful unless it’s programmed to give responses that explain where to find additional help. Most current conversational interfaces are not equipped to provide further direction for inputs other than those in a finite set.
The future: Bots use big data and interactions
Historical data is not a one-way street. Bots can harness big data to provide personalized interactions and collect data to give companies more information to use in other marketing activities. Right now, bots acquire data in a noninvasive way—the user initiates the interaction and is therefore primed to provide personal information.
In the other direction, bots using big data can provide product or educational suggestions to existing customers without first asking for personal information, thus creating a new kind of recommendation system. Brands can use social media sentiment analysis to adjust the tone of bots. Data about common queries and pain points in interacting with a website can inform bot developers as they create advanced responses to meet user needs.
Advances in big data analysis and NLP will go hand-in-hand in realizing these opportunities for customer engagement and personalization.
Where will bots show up next?
There are many benefits to giving users the opportunity to interact directly with brand websites rather than having them rely on navigation and internal search, which can be wildly variable in quality. Brands can also retain more users with the “personal touch” and immediate responses that bots can offer. And there are many interface options for integrating the feature on websites, from live chat to banners to full web pages.
The advantage of using existing platforms, such as Kik, Facebook Messenger, Twitter, WeChat, or Slack, is undeniable. And the audience is already there, choosing to interact when the option is available.
We’re curious to see how brands will integrate personalized bots with their websites or with messaging apps in the next couple of years.