Chatbots: Cool Technology, Low Customer Retention
According to the technology experts, chatbots are the way of the future.
As companies are discovering, though, there have been a few snags on the path to tomorrow.
In theory, it seems simple:
Businesses need to go where the customers are. The customers are on messaging platforms.
Users are on messenger platforms now more than ever. In 2012, it was estimated that there were 14.7 trillion total messages sent through app platforms. In 2017, that number is expected to reach 28.2 trillion.
· As of 2015, Facebook Messenger had 87 million users on the app in the U.S. alone.
· On a worldwide scale, 72% of Facebook Messenger users purchase products online.
Trends are showing that customers are using social media less, which is undermining the work that businesses have done when it comes to advertising on social media.
What’s not getting talked about much is the drop in the users’ app downloads and usage. Statistically, the average smartphone user installs zero apps per month, and only uses an average of 5–10 of their downloaded apps on a regular basis.
This is actually a good thing for the average business, since apps are hard to build, and even harder to deliver to the right customer base. Bots, on the other hand, are relatively easy to build, and consumers can have a wide range of applications in one place. Instead of wasting resources on creating individual apps, businesses can move towards a simpler interface that uses scripted AI to interact with users.
How the Technology Works
There’s a lot of confusion over the technology used in chatbots, especially when it comes to the term “AI”.
Chatbots are actually scripted AI, meaning that they’re pre-programmed to stick to a script. In order to work, they’re created with a decision tree model — kind of like a flow chart of yes/no questions. “Have you tried restarting your computer?” No. “Okay, why don’t you try restarting your computer, and see if that fixes the problem.”
The responses of the chatbot are based on how the customer interacts with it. The scripts also use a dialog tree to have more involved conversations with users based on their responses:
So what happens when a customer goes off the script, and the programmed bot can’t handle it? Much like dealing with the automated customer service program a bank uses, customers will get handed off to an actual human who will be able to answer their questions through the chat.
Scripted AI isn’t to be confused with an autonomous AI bot, like Google, which scours the entire web in seconds to deliver the best matches to the search.
In the case of most businesses though, a scripted AI is the most that will ever be needed. If a company is selling flowers, chances are they won’t need an autonomous AI program just to help a customer send his mother a bouquet.
So…Why the Low Retention?
We know where the consumers are. We’ve built bots to reach them.
But the consumers aren’t biting. They’ll try out the chat feature once, and then never return.
No matter how intelligent bots are, a Q&A conversation with a bot is not the most effective way of getting answers. Google is.
Let’s put it this way: Even if you knew you could reach a representative in seconds, would you call your bank to ask what the branch hours were on Saturdays? Or would you just look it up on Google, and continue on with your day?
This is where bots are going wrong. Imagine you want to order food from a local delivery place, but you’re not sure what you’re in the mood for yet. You now have to go through a conversation with a bot to try and figure out if you want Indian food or Italian. It’d be much easier to look at selections offered, click, and move on.
The Future is in Bots, Not in Chat
Chat is not the best way to gain customer retention.
Users want the information fast, along with the personal assistant aspect of an AI program — but they don’t really want the chat part.
In order to be functional, the chatbots need to be streamlined.
The bot needs to be easy to access, easy to use, and easy to get answers.
Bonus points if the bot does its own thing and works in the background where the user doesn’t have to interface with it directly to get results.
Take Uber, for example. Five months before Facebook even launched the bot program in Messenger, Uber had already worked with Facebook to have company’s bot built into the Messenger program. When you tap on an address within Messenger, an option for “Request a Ride” pops up, and directs the user to an interface that looks like the Uber App… without ever leaving Messenger.
This kind of two-click-and-done system is what consumers are looking for in bots — not a scripted conversation.
That’s not to say that chat doesn’t have its place — when someone is looking for obscure information that only a conversation can accomplish, then chat is great. But companies can’t rely on obscure conversations happening with the same users more than once. This is where chat and retention go their separate ways.
Transforming the Bot Service
We may be putting too much reliance on the “chat” in chatbots. The term actually comes from a Wikipedia entry for “Chatterbot”, which got shortened and thrown around in the tech world. It’s a program that can converse like a human, but a bot itself is just a piece of software that makes something happen.
This is what businesses need to keep in mind when designing their bots: Just make something that makes something else happen.
We’re putting a lot of stock in the fact that users are flocking to messenger platforms, but what we’re not taking into account is that the 1 billion users a month on WhatsApp, or the 950 million a month users on Facebook Messenger are already there to have conversations with actual humans. Businesses don’t need to simulate conversations, and consumers have long ago given up wanting a retail experience. They want things quickly, without hassle, and preferably in a single click.
Did you learn something? Click the 👏 to help others find this article.
Lastly, I’d ❤️ your feedback via the Survey Monkey link below 👇
Many thanks for reading.
