Website Chatbots: The New Alternative to Facebook Messenger Bots

How chatbots are changing the user experience on non-messaging channels such as the Website.

Jiaqi Pan
Landbot.io
8 min readOct 30, 2017

--

Chatbot, a term popularized by Mark Zuckerberg when he launched the Facebook messenger platform. Though the concept of chatbot has been around for more than 60 years, it was Zuckerberg who claimed that chatbot is new UI and declared the death of mobile app. Soon after, most major tech companies like Microsoft, Google, Apple, and Amazon quickly followed the trend of Chatbot.

However one year after the initial hype, chatbot seems to be at trough of disillusionment in the Gartner Hype Cycle

Gartner Hype Cycle

The problem is mainly due to the approach most companies took. They were focusing too much on the so-called intelligent assistant using AI technologies like Natural Language Processing to make chatbot understand every user request. However, the state of art of NLP is nowhere near to get enough accuracy which resulted in massive user deception.

Today, most chatbots are built on top of messaging platforms like the Facebook Messenger who claims to have more than 130k bots available. There are also many other messaging channels available like: Wechat (Tencent), Skype (Microsoft), iMessage (Apple), etc.

For a complete list of different messaging platform check out this Platform Overview slide from Andy Mauro.

While the industry is still looking for the killer bot, many people realize that Chatbots are not necessarily attached to messaging channels. What happens if we take the concept of conversational interfaces to another digital platform like website or mobile app? The answer to this question leads to the rise of the so-called Native Chatbot Development.

Today I would like to explore in detail about Website Chatbot . First a quick definition:

“A Native Chatbot which interacts with users in the website browser.”

Types of Website Chatbots

We can differentiate three types of chatbot for website:

Live Chat

It´s one of the most frequent types of bot website due to the similarities between live chat and messaging app. The idea here is to use chat windows you typically see when visiting a website as the communication channel. So instead of a human agent, the conversation with the visitor could be automated using a chatbot. Many live chat software companies are already experimenting with chatbot; Some great examples could be Intercom with Operator or Drift with Driftbot.

Driftbot by Drift & Operator by Intercom

Banners

This is another variant of the previous one. Instead of showing the chat window at the bottom of a website, you could put it as a banner in any part of your interface. You can use it as a pop-up window to offer a special discount or use it as an interactive guide to do product onboarding. A very visual example is the conversational article of Typeform, where they combined the post with small fragments of chatbot to offer user extra info.

Conversational article of Typeform

Full Website

Imagine when entering a new website instead of throwing all information at you, the site starts having a conversation with you. Literally! This is by far the most ambitious type of website chatbots with many UX challenges. There not many people working with this approach yet, but you can see some interesting cases like the Upcoming Pages from Product Hunt or Landbot.io turning landing pages into a chatbot.

Landbot.io “Home” screenshot

What are the benefits of a Web Chatbot?

Customization of User Experience.

One fundamental difference between the website and messaging channel is the control companies can have over the complete customer journey. They can customize every element of the UI which can benefit hugely in the user experience. Instead of NLP, you can use interface-based interaction to solve many UX issue. For example: in a registration process if a user wants to edit some info, instead of asking the bot to change it, you can design an interface to modify the data directly.

Less dependency and risk.

When it comes to working with 3rd party platforms, there are always risks of generating too much dependency of the platform and jeopardize future sustainability. What will happen if Facebook decides in the future to limit the reach rate of each bot, just like what they did with news feed? Or Twitter shutting down their API for bot developers (again)? That´s why Media companies didn’t want to work with Instant News.

User Privacy Issue.

For companies in highly regulated industries like healthcare, insurance, banking, etc. the privacy issue is critical. You can have more engagement or more open rate on messaging channels, but at the end, the users belong to the platform owner. So companies will lose control of data exchanged in a messaging channel. That’s one of the main reasons many Banks are moving away from Messenger and building their own Chatbot platform like Erica from Bank of America.

Benefit existing channels.

The Website for many companies is one of their main customer interaction channel. It makes a lot of sense to build better experiences on that native environment to benefit their existing users and to avoid adding unnecessary friction. Think, for example, a new user is going to your mobile website and wants to know about your product, and instead of giving him the information right away you invite him to chat in Messenger. In the mobile browser, users have to log into their Facebook account and give permission to start talking to the brand.

Some cons of Chatbots for Website

Nothing is perfect, so web bots also come their limitations:

Lower user retention.

Unlike the messaging channel where users are active on a daily basis, on a website the retention rate is lower. Naturally, it will depend on the type of website: a saas product will have better engagement than an e-commerce site.

Discoverability and distribution.

Messaging channels can offer a unique context for delivering product/service: i.E., the Uber chatbot for ride-hailing in Messenger. Furthermore, as a “virgin” channel, messaging bots enjoy of 40% average CTR with push notifications. The good news is, with website chatbot we can use tools like retargeting and browser notifications to regain user attention.

Offline 2 Online interactions (O2O).

The main advantage Wechat has in China is that it has standardized the use of QR for people to interact with the offline world: movie center, restaurant, supermarket, retail store, etc. Facebook is using the same strategy in Messenger, and this kind of O2O interactions is more accessible with a commonly used messaging app than a website.

Ideal use cases for a Website Bot.

Lead generation:

We can use chatbot as a landing page to capture leads of potential customers. I have argued here that this might be the killer use case for a chatbot. In short, Conversation interfaces by nature focus better the user attention, have better personalization capabilities and can be used to do real-time qualification.

Why better than messaging?

Landing pages are the natural destination for most marketing campaigns. Unlike messaging channel, it´s platform-agnostic, you can use the same landing page for AdWords, LinkedIn ads, or Twitter ads.

With proper SEO configurations landing pages can position well to get organic traffic in search engines. Companies could also leverage existing traffic and use web chatbot to improve conversion rate.

Feedback survey:

Another compelling use case is to ask user feedbacks, but instead of annoying forms, we could humanize it with conversations. Furthermore conventional surveys don’t usually adapt well in mobile, but with a chatbot, we could use alternative input systems like voice to improve the UX.

Why better than messaging?

With conversational UI and proper survey design, you can get more reliable, actionable data, because the polling experience is interactive, which gives you the freedom to prevent survey fatigue, speeding and other types of “bad” respondent behavior.

Unlike messaging channel, user interactions are usually anonymised by default on a website, which will help to reduce possible friction from the user about revealing his ID.

Customer Service:

Answer customers question using a chatbot to replace FAQ and knowledge base with a more interactive experience. Customer Service is one of the most common use cases businesses have in mind due to the potential cost reduction in human support.

Shep Hyken wrote an excellent piece to cover this topic here

Why better than messaging?

For companies with a digital product like SaaS platform or marketplace site, a web bot can gather all user data thus get better context and offer a more personalized customer experience.

A website chatbot can reduce friction by solving user request immediately without changing the communication channel.

The challenge of cross-channel communications

A final thought about the future of chatbot: in my opinion messaging and website chatbots are not exclusive, rather than compatible between them.

For example, in a marketing campaign, a user can start the journey at a website based chatbot. Through the conversation, he ended up subscribing with his Messenger account. After getting the lead, the business could leverage the messaging channel to deliver relevant contents to the subscriber according to the data collected from the web chatbot.

Furthermore, Website and messaging are not the only channels available; we should also take into account emails or voice channels like Echo. Customer journeys from today are much more complicated and rarely rely on one single channel. The real challenge for companies is to offer frictionless experience with cross-channel communications.

Thanks for reading the whole piece, I'd love to hear your thought about website chatbot!

For more content from me please subscribe to my WhatsApp Newsletter 🙌

--

--

Jiaqi Pan
Landbot.io

CEO at @landbot_io | Humanizing the internet | Love talking about #chatbot #conversationalUI #business #techs #messaging