How to design a Messenger bot for E-commerce?

With the launch of Messenger bots, Facebook is reshaping the way a customer interacts with a business. In less than a year, the adoption of Messenger has grown to an extent that users have begun fully transacting on the platform. With further advancements in technology and with Facebook aggressively marketing the Messenger platform, bots will soon become pretty mainstream, just like apps or websites.

So, does your online store have a Messenger bot yet? In case you don’t, how do you go about launching one? Which functionalities should it have and how should it be designed? These questions are not easy to answer and a good amount of time consuming research may be needed before any solution can be implemented.

But, there is one way of knowing what’s right when it comes to bots. David Marcus, VP of Messaging Products at Facebook, in his speech at the F8 conference said “We really believe that in order to build a great user experience, you need a hybrid approach of structured UI with conversational elements. Many people have been talking about bots as a command line interface, but this is not how we see it. We think the combination of UI and conversation is what’s going to work”.

We really believe that in order to build a great user experience, you need a hybrid approach of structured UI with conversational elements. Many people have been talking about bots as a command line interface, but this is not how we see it. We think the combination of UI and conversation is what’s going to work.

To put it in simple words, a user will never purely use natural language (conversation elements) while interacting with a bot. “Show me a red t-shirt of size 42 US in brands like Nike” – this is an example of a pure conversation, something a user wouldn’t be likely to type. According to Marcus, a user would instead prefer a structured UI where the same search (item: T-shirt, colour: red, size: 42, brand: Nike) can be performed through a series of steps. In the first step, the user would choose “T-shirts” from available options like {Shirts, Jackets, Hoodies, Vests, Suits….}. In subsequent steps, the user would choose “color”, “size” and “brand” in a similar manner, thus completing the search effortlessly. Coming to the conversational side, Marcus believes a user may make short conversations like “Show me Jeans” or “Black shirts”. These short conversions need to be attended to with great care, especially since the user may use abbreviations and/or make spelling mistakes.

 

How can Shopbot help you?

ShopBot is a bot builder platform that helps an online business to build their own Messenger bot. A bot created using ShopBot’s platform carries a hybrid UI comprising of Structured and Conversation Elements, as suggested by Marcus. The Structured UI can be seen in the image below.

Instead of entering a complex sentence, the user at each stage simply chooses one option from the available options to build the search. At first, the user chooses Men’s T-shirts and then gets to choose between options like {“printed”, “plain”….}. These choices are stored in little bubbles which the user can select and each product the user views is displayed in a carousel format.

Coming to the Conversation Side, ShopBot allows simple text based search requests like “show me men’s pyjamas”. ┬áIn the example below, accurate pyjama results are shown, even after the user has spelt it incorrectly as “sho mi muns pujamas”.

Want to build a bot for your online store? Reach out to us at app@qgraph.io. You can also check out a live demo of Shopbot at https://m.me/shopbotqgraph