Take a good look at the screenshot below. That is a menu, an homepage. A very cool one. 20 years ago.

Chatbot design in 2017 is like web design back in 1996: In its infancy.

Even the state of the art is a little bit odd. It’s not very easy to use. It can be confusing. It has potential. It sports a lot of GIFs. Design principles that will be obvious in the future are obscure now.

Web design principles got so common we don’t even see them anymore. So if you have to decide which form sports a better design between those two:Good design vs bad design

Almost everybody will pick the one on the right, although not everybody will be able to articulate why. Moreover, if you own the form, you will see that the form on the right-hand side works better, in the sense that more people understands it, resulting in more people using your website.

Same with conversational design:Good conversational design

The one on the right has personality, sets user expectations about its scope, and, by providing examples, clarifies what the next step is.

Exactly as the web page, the right hand side chatbot works better, in the sense that more people understand it, resulting in more people using your chatbot.

The design on the left-hand side will receive answers likeAnswers on Facebook Messenger

What this means is that, if you improve your chatbot design, you will end up with a chatbot that works better, without changing technologies, code or alike.

Note that this is also true if your conversation sports only buttons, menus or other UI elements.Good vs bad conversational UI

What’s wrong with our TrainBot friend here? A lot actually:

  • It shoots three messages in a row before the user has started out (not good)
  • It starts by asking where you would like to travel from, instead of where you would like to go, which makes more sense
  • Mixing options (“Rome”, “Naples”) with a different type of action (“Share position”) makes the interaction more complicated at a glance
  • Not giving any additional, residual option (like “Other”) makes the user feel trapped

So please don’t think that using buttons will allow you to come up with a horrible design.

It is actually true the contrary: buttons and alike will hide the design flaws from you, as you will not get confused, free text answers from your answers. You will just get frustrated users who will not use your chatbot.

Ok ok, but why do you care so much? Conversational design will take care of itself, much like web design did since 1996

There’s a big difference between web design in 1996 and chatbot design in 2017.

If, like me, you were in elementary school in 1996, you will remember that the SpaceJam website was competing for your attention against the real world only:Competitive landscape

SpaceJam website competitors

The competition from the real world was substantial, but it had all the flaws all the real world: Information took more than zero to be transferred; mail had to be sealed manually, you had to walk to the mailbox to catch them; you have just one life; etc.

On the other hand, your chatbot in 2017 has a LOT of competitionChatbot competitive landscape

You are competing for attention with the web, mobile apps, and a variety of other platforms. It is fighting for relevancy.

What this means is: Conversational design has to get better a lot faster than web design did in order to remain relevant. So if you want to succeed in this space, getting better at conversational design is a must.

By the way, the SpaceJam website is still online, you can check it out here

Originally posted on Medium


Categories: Ideas

Vittorio Banfi

Co-founder at botsociety.io