How Facebook chatbots can revolutionise your marketing for 2020
Trust me, I’ve heard it all before, “bots don’t work”. I have to admit that on more than one occasion I have found myself jumping up and down defending chatbots. Why? Simply, I have built hundreds of chat bots and I have seen the results first hand. As a marketer, my focus is always geared towards results (rather than opinions).
Data is the key to successfully spending your money in the right channels. Chatbots have an open rate of between 80-90%, compared to email which is around 20-30%. However they do come with some rules that limit use and should be used as a bridge to other channels, an omnichannel (a marketing term to describe reaching your customers simultaneously across multiple channels) connector that can excite and engage like no other platform.
I will be the first to admit that there are many bots I have met I didn’t like, so what is the secret sauce of successful chatbot marketing? Check out my top tips below.
What is a chat bot?
A Facebook chatbot is simply a series of pre-written messages that guide a customer through a flow of copy, and answers queries based on their selections. It is not necessarily Artificial Intelligence which is what some bots use (we do cover this later in the article). The flows are created using a platform like manychat or chat fuel (the list goes on) and connected to your Facebook page and/or website or triggered when a link or QR code is clicked. They open in messenger and the conversation starts (before you say “but my customers are not on facebook”, think again, 70% of all Australians are on Facebook, so really only excluding kids).
They look a little like this:
They don’t call it a chat for nothing! I see so many bots that lack the very fundamentals of conversation. The conversation style should match your business but with a little more fun thrown in (this isn’t a static Facebook post or a letter to your local paper). This is essentially a live conversation, and while we don’t recommend pretending to be a real person (we generally start by calling out that it’s a bot), the key to a successful bot is short engaging conversational style copy. Don’t be afraid to use emojis and gifs, they give the feel of a real messenger chat, and always use delays – these give the 3 little dots that appear when someone is typing.
Get your intent right!
Figuring out exactly what the purpose of your chat bot is helps users to feel more connected to the experience. If the information is not well executed the users get stuck and easily frustrated. It can’t replace your support or customer service but it can certainly handle the initial connection and sales front. There are some incredible opportunities to seriously delight customers. Here is a list of just a few our team have built!
- Visual dining menu (great opportunity to capture restaurant patrons details while providing a unique experience)
- Lead generation (part lead page, part qualifying tool)
- Medical triage (helping parents to administer the right first aid)
- Augmented reality games (experimental techniques to excite and engage)
- Monthly or birthday subscriptions (continually engage customers with fun content)
- E-commerce shopping (the entire shopping experience in one bot!)
Using Artificial Intelligence.
Artificial intelligence is essentially machine learning that helps the chat bot to learn from queries and the data provided by you during bot training. When a query is triggered, the machine learning helps the bot to first monitor the past conversation it had with the user and give a response accordingly. This is where things become a little hairy, the responses can tend to come across a little too robot and when things go wrong the frustration starts. We recommend sticking to the method of watch and build as outlined above.
Monitor & Adjust!
This is where many bots fall down. The set and forget approach simply does not work. Ultimately your bot is a tool to communicate with your clients, it is essential to understand what they want to know rather than assume. In the first flow you will be making assumptions about what they need from you, the first few months is where the magic happens. As your customers interact with the flow of content you can (and should) adjust content to improve the experience. You can also build keyword responses to minimise your internal team input to queries outside the flow of content (which there always is).
Example: There may be customers who continually ask “do you cater to vegans?” in which case you can create an automated keyword response “thank you for your enquiry, sure Sam we do have many vegan options available on the menu – would you like to see a list?”