How A Chatbot Named Otis Got People Excited About Oatmeal
Oatmeal. Not only does it make for a hearty breakfast, but it also turned out to be a great reason for Quaker Oats to engage with inquisitive customers, while also influencing consumer behavior and brand awareness.
Quaker Oats decided to add a Facebook Messenger bot, named “Otis,” to its marketing mix. Otis is meant to guide customers through the customer journey, from finding products online and in-store, to discovering new recipes, to setting reminders to make overnight oats, to offering responsive customer support.
Otis went to work in August 2019. Since then, Quaker Oats has engaged with over 6,000 consumers, most of whom have been interested in finding recipes or connecting with customer support. How did the brand get such great results so quickly?
CMO by Adobe’s Giselle Abramovich sat down with Elena Parlatore, global director of consumer experience at PepsiCo, who spearheaded the bot’s launch, for the details.
Why was a chatbot the appropriate channel for what Quaker Oats was trying to achieve?
Knowing the information consumers are looking for, coupled with the fact that today people lead busy lives and are stretched for time, the bot allows consumers to receive direct and instant answers when compared to exploring our website on their own.
Can you walk me through a hypothetical engagement between the chatbot and a consumer?
A consumer may either engage the bot through the QuakerOats.com recipe pages or direct message the bot on the Quaker Facebook page. After they engage our bot, named Otis, he will introduce himself and what he can do. Since our highest engagement comes from recipes, a consumer likely would select to see recipes. We present them with several categories and options from which they can select. Chances are they will select a seasonal recipe, like one of our new smoothie recipes, or type in a food emoji if they know what they are looking for.
Are you personalizing the bot’s engagements with consumers?
The bot is a great way to learn about customer intent, preferences, and characteristics. We already know the products people are searching for, where they intend to buy products, types of ingredients they are looking for in recipes, top recipes, and other engagement metrics. This data allows us to tweak the experience and content to serve what customers are seeking and is a way to personalize at a macro-level. Further, the bot can look to remember recipe and ingredient choices that the customer previously made and suggest products and recipes matching that on subsequent visits. Additionally, the bot can push a notification to customers who are opted in when a recipe matching their needs is introduced.
How are you measuring success, and can you share some early results?
Our initial intent with the bot was to engage consumers around new usage occasions, so we’re definitely looking at engagement over time, and we know year-over-year we’ve increased conversations by 13% without any dedicated media or marketing support. As we’ve evolved our approach to the bot, we’re also looking at goals around efficiency for our consumers. Specifically, we’re looking at the percent of consumers that were able to get answers through Otis without needing a live representative to assist them. We’ve just started to measure efficiency so there will be more to come.
Now that you’ve got some data on how people interact with your bot, are there any trends or learnings you can share with our readers?
Half of the bot usage is around recipe engagement, so we have the richest data based on that intent. Through this data we know the most popular recipes are the vanishing oatmeal raisin cookies, prize-winning meatloaf, best banana bread, favorite oatmeal pancakes, and a handful of overnight oat recipes–which we think speaks to consumers’ curiosity and interest in experimenting with using oats in unique ways.
What has been truly interesting are the ingredients that consumers search for. We’ve seen everything from the expected oatmeal and cookies to the unexpected egg and okra.
Can you offer some hints of what’s to come with version 2.0 of the bot?
We’re always looking to improve the bot experience for our consumers. At the moment, we are adding new content and Q&A based on the data we receive from user interactions.
Generally speaking, what do you see as the future of chatbot technology?
The first generation bots are out there. Our observation is many of them are scripted and missing the mark when it comes to understanding human intent. As technology advances, bots are going to get a lot smarter in understanding what humans are typing. Personalization–including remembering customer’s preferences, proactively notifying them, and making content relevant–will make bot experiences a lot more engaging and meaningful. Bots will employ more AI technology to invade in-store experiences, such as taking a picture of a product and getting more product details, offers, or recipes around that product within the bot. It’s exciting times as conversational experiences mature.