The conversational design equilibrium

Maybe chatbots could learn a thing or two from Goldilocks.

Remember the fairy tale about Goldilocks 👧🏻 and the 3 bears 🐻🐻🐻 out on a walk until their porridge 🥣🥣🥣 cools down?

Over the years that little girl and the 3 bears have inspired many analogies.

The Goldilocks Principle states that things must fall within certain margins, as opposed to reaching extremes. For example, the Goldilocks Zone is the habitable zone around a star where the temperature is just right — not too hot and not too cold — for liquid water to exist on a planet.

This habitable zone for chatbots needs to be a bit more specific than just the right temperature. It needs the right level of humanness, automation, conversation, interaction, visuals, and strategic intent to survive the hype.

Product, design, technology, and marketing teams must be fully aligned to make this happen. If not, it’s easy to be too much of this or too much of that.

Kind of human, kind of machine

Just enough human to relate to, but not so much that users are fooled into thinking there is a real person talking to them.

Human interactions with machines have always been about input and output. With the increasing sophistication of machine learning, machines have been catching up on human speak.

Conversational design is truly human-centered design, every step of the way. There is no next big thing, only the next step in an unfolding story of how people use technology to be more themselves. —

Erika Hall

Give you chatbot all the personality it deserves, but let’s not elevate them so much that they, even for a second, are able to trick us into entrusting them with our emotionally loaded communication.

Just enough machine to do all the heavy lifting, but not so much that users feel a lack of control. Machines like being machines- searching, calculating, analysing, predicting, and presenting.

No one wants to do more than they have to, thats why we build systems that are invisible to the eye as well as to the brain. Our tolerance for friction depends on the context, what is important to not go beyond that threshold.

Somewhat like messaging, somewhat like an app

Just enough messaging environment to make it feel familiar, but not so much that users are required to type a lot.

Being able to exchange text and voice messages with a device doesn’t really make it easier for people to get things done. We are no more impressed by machines understanding our unique slangs and accents.

Just enough app to facilitate smooth information entry and optimal information display, but not so much that users need a navigation menu to get things done.

The good old Graphical User Interface best practices still play a big role in providing a smooth experience. Great apps are multi layered. Chatbots need to be even more multi layered while keeping the complexity under the hood.

A little conversational, a little transactional

Just enough conversational to make it feel natural, but not so much that users lose their patience.

How we talk to each other is not how we want to talk to machines.

It seems so obvious that conversation is the most human centered way to create device-agnostic interfaces. Conversations in real life have an underlying protocol, and so do conversational systems. What we say, and how we say it is based on many quick real time decisions.

When to stop talking and start doing is critical to the overall experience.

A nice waiter 👨‍🍳 may engage in some small talk but doesn’t continue talking once the order has been placed.

When I ask Google Home to turn on the lamp, it executes the task and utters a long sentence that it has actually done what I wanted it to do. Amazon’s Alexa on the other hand does it without making a ruckus about it.

Just enough transactional to be functionally valuable, but not so much that users develop no connection with the chatbot.

“When designed well, conversational interfaces are able to amplify the human-computer bond — and create a relationship grounded in communication. When they work poorly, they produce mistrust. When they work effectively, they promote trust.” —

John Maeda

A bit of a push, a bit of a pull

Just enough push to nudge towards an on-boarded state, but not so much that they get annoyed and abandon the chatbot.

Some amount of external push is certainly needed to introduce the chatbot to its audience and then slowly onboard them so they can use it well. The problem arises when this permission to introduce turns into permission to push out of context advertisements. It’s the worst place to be sneaky about promotional content.

Just enough pull (engagement) to be fun to use, but not so much that users start to feel trapped. Over engagement is actually a good problem to have, as long as people don’t feel too zucked. (there’s a new book by this name, if it piqued your interest)

Part customer success, part business success

Finally, chatbots exist to get a job done for the user as well as for the business. It is hard to focus too much on the customer while forgetting the business objectives. But it is very easy to rush through the user’s problem space, and do it just because everyone else is doing it.

Let’s not try to use the chatbot hammer to nail random problems. Some specific problems worth solving using chatbots but not all. I have been a believer of hybrid systems and digital products. The goal should be to retain what’s already working and fix only whats broken.

Most of the chatbots I have used make the mistake of picking the inhabitable zone. The only way to survive in that zone is through life support 💰. A case in point is this misuse and overuse of “chatbot technology” for helping people find content on the website. It’s not a good experience to type a specific request only to be recommended links to check out. There are other better ways to solve that problem.

A dash of hype, a dash of reality

And then there’s this one more Goldilocks Zone to consider— a bit of hype and a bit of being grounded in reality. For every hundreds of failures, there are only a few success stories.

The hype may be over but we’re just getting started with conversational shopping, banking, marketing, support, and what not.

Conversational is a truly native experience across platforms. Chatbots will remain a great way to uniquely serve everyone without overwhelming anyone. These are very strong reasons that the hype that started in 2016 is still trudging along.

Do not try to replace apps or people via chatbots yet. Just focus on the elements of functional and emotional value. People are just trying to be better at something, have your chatbots play a role there.

Also read: Your Chatbot Strategy Needs These 4 Levels of Interactions