Why Facebook Messenger Chatbots get me excited as an App Developer
From Apps to Chatbots — learning about Facebook Messenger Bots, ManyChat, ChatFuel and more.
When the App Store crossed the 100.000 apps mark at the end of 2009 it became clear to me that I had to jump on the platform. I am glad I did and focused on building apps ever since. I am trying to figure out now if I should jump on chatbot development.
The transition from bloated desktop applications to mobile, single-purpose apps enabled entire new industries and being a developer suddenly became cool. According to App Annie’s latest report the global app economy has crossed a trillion USD in 2016 already.
Ease of use and instant accessibility of software reached a new level with apps. Apps don’t need extensive training programs to learn how to use, they don’t require an IT technician to install and they are available wherever you go.
At the same time creating and distributing quality software has never been easier, cheaper and faster. It lowered the bar of technical skills and experience required to build useful tools for users. Instead of having to spend budgets mostly on mastering the technical challenges and software distribution, teams have more room to focus on delighting the user. Time to market for software used to be measured in years, with apps it is measured in weeks.
In the last days I have been trying to learn if chatbots are the beginning of a transition at similar scale. People love chat, it has replaced email in our personal lives in the form of Facebook, WhatsApp, WeChat or iMessage and by now even in many corporations thanks to tools like Slack. Chat is often even more convenient than calling the other side, especially when dealing with service agents.
At least in the western world the Facebook Messenger is ubiquitous — more than a billion user are active every month. Messaging is an even more humane way of interacting than apps and even simpler to learn. Installing a messenger bot is instant as the Messenger app is already there. Creating a bot requires even less technical skills than building apps. Getting to market with a quality messenger bot is a matter of hours.
High engagement, simple onboarding, less commitment with Messenger Chat Bots.
I’m always careful with generic figures like this but it seems that engagement and conversion rates in Facebook Messenger bots are through the roof compared to traditional email channels. Josh Fechter reports message open rates of 90% and click-through rates of 60% compared to a similar email campaign which was at 30% open rate and only 4% click through rate. These figures are probably above average but apparently not unheard of by other early adopters.
You need to try signing up to a Messenger Bot yourself — the experience is mind blowing compared to installing an app or signing up to a newsletter online. Signup is instant. Try the onboarding experience of Bot Academy to get an idea. Literally the moment you tap the button no matter if on desktop or mobile, you receive a message on your phone with the next steps.
It feels much less of a commitment to sign up for a service than if you have to hand out your email. In Messenger you are in control of the interaction and can opt out of notifications or block a bot entirely at any time. Compare this to the stream of spam you receive by most email based signups even if you long stopped using the service.
Messenger Chatbots for ebook signups, travel and e-commerce re-engagement.
A basic example is the signup flow for the BAMF bible — I still have to finish reading it to be honest but I loved how its delivered. You just press the Messenger button and Josh’s chatbot sends you a download link directly in Facebook. Way easier than having to type my email and gives Josh the benefit to engage with his audience directly via Messenger.
A more app like example is Expedia’s messenger bot that actually lets you book flights, hotels or manage your reservations via chat. It feels a bit like some of the human-powered travel assistant apps out there but a lot dumber… For me personally everything felt too slow and I didn’t feel like its possible to find the best deal with this bot. It could easily be improved by remembering basics like the context of the current search while changing filters and parameters. I had to start over every time even for just filtering for direct flights or comparing one-way versus return flights.
Wish.com has an interesting approach to engage with customers. Upon completion of your purchase they will ask you if you would like to receive your receipt via Messenger. This gives them a direct channel to you and they will send weekly messages with the latest deals. I learned about this use case in Andrew Warner’s great intro to Messenger Bots at Converted 2016. It wouldn’t be hard to combine this with infrequent questions about your interests and would allow them to build great customer profiles over time.
Create Messenger Chatbots with ManyChat, ChatFuel or BotMock — no coding required.
I’ve only played with a few tools so far so will just document my first impression here.
ManyChat is probably the most established tool at this point. But the space seems to change fast, other players could catch up any moment. Andrew Warner, the founder of Mixergy and Bot Academy, invested in the product and recommends it in his courses. It allows you to configure and deploy a complete bot with no development knowledge required. The use cases you could tackle with this actually seem really broad with features to design complex sequences of interactions. You further get a dashboard with basic metrics, an audience tool for segmentation and tools to directly engage with your users.
ChatFuel appears very similar, it seems like they support more complex interaction flows but maybe the interface was just easier to grasp for me. As a developer I’m most excited about their JSON API plugin which allows you to integrate with APIs in your bot. I couldn’t figure out a way to view the interactions of users with my chatbot other than in the Facebook Page Manager directly. I would probably have to upgrade to Pro to access their People tab — to me that makes the free version a bit pointless to use.
BotMock is another tool I discovered late at night but was too tired to play with. It looks really promising with a very visual chat flow designer and support for a wide range of messenger platforms.
Great for marketers. Replacing apps? Not sure yet.
Most of the successful use cases I read about so far revolved more around customer engagement for sales and marketing. If the metrics are representative then Messenger Chatbots are a gold mine for digital marketers.
I haven’t seen a really convincing example of a more app-like experience yet. But we are still at an early stage and I believe that just how apps had to rethink user interaction compared to desktop applications, chatbots will have to come up with new, radical approaches too that leverage the medium.
Tying myself to a channel that is controlled by a single platform is a bit worrying. But what counts in the end is convenience for users and conversions for businesses. When engaging with a conservative enterprise B2B audience I believe Messenger Bots will be hard to sell though. Employees can feel uncomfortable using their private Facebook accounts on corporate devices or the organization might even block Facebook entirely.
Next up: my own chatbot, use cases with value, WeChat and Alexa.
I’m going to replace the Drift chat on blackbeltlabs.com with a Messenger Chat. My target audience for Blackbelt Labs is clearly present on Facebook, with most clients I engage directly in chat as its more efficient and less formal than mail. So making that connection easier is only a logical step. I will definitely over-engineer the Blackbelt Labs bot and document my experience here.
The next goal is to find industries and use cases where chatbots can deliver actual benefits. There are a lot of gimmicky bots out there that are fun for show but deliver little business value. The lowest hanging fruits are probably businesses who already engage heavily in manual chat interactions with customers on Facebook or other platforms. Automating some of the qualifying steps with bots should be a quick win.
I heard lots of stories about WeChat in China and the wide adoption of a range of bots in the platform. Researching and documenting those will be another item in the next days. Also the iMessage for Business rollout will be soon and WhatsApp just introduced its business app.
The whole voice based bot space around Alexa, Siri or Google Home is closely related to chatbots. Alexa Skills looks like an exciting space to get into and many of the learnings from chatbots can probably be applied to voice bots and vice versa. I want to focus on chatbots initially but Alexa will be up next.
Hit me up or comment if any of this resonates with you. I’m also always happy about links to useful resources to dive into.
