The Six Principles of Conversational Transformation 

Everyone wants to innovate. No one wants to change.

The good news is that technology is advancing to the point we can interact with one another and with our systems in a way that reflects our nature as social, conversational beings, with instant access to boundless information and freed from the constraints of physical presence. The bad news is that we are too often trying to use technology to avoid dealing with some of the more problematic aspects of human interactions, the ones that technology doesn’t fix and often exacerbates. And we are missing out on real opportunities.

Step away from the machine

“Computer literacy…is really a euphemism for forcing human beings to stretch their thinking to understand the inner workings of application logic, rather than having software-enabled products stretch to meet people’s usual ways of thinking.”
— Alan Cooper, About Face

Technology is just a tool. It’s easy to lose sight of this, even though that’s literally the definition—knowledge applied to some task or purpose. The growth of the internet and advances in processing power over the last couple of decades have changed so many things so fast, it is legitimately difficult to tell when you have a technology problem and when you don’t. And technology problems attract more attention because they are newer and seem solvable.

If the telephone isn’t working—that’s a technology problem. If someone is hitting you over the head with a telephone, that’s a people problem. If dealing with your bank’s interactive voice response system makes you want to throw your phone at the wall, that’s a people problem, too. Scaling and automating systems and processes that make people want to throw their phones at the wall doesn’t help anyone. Except maybe phone manufacturers.

Using conversation as a model for design is exciting because it gives us more ways to adapt technology to a wider range of human needs. Instead, too many companies are lining up under a banner reading “Put a bot on that!”

Whether a system that interprets natural language on the fly is an improvement over other ways to do things depends on what people need and what value that interaction might create. Just because a technology is cool and tricky doesn’t make it superior in a given human context.

Solving the right business problems in the right way requires the perspective to figure out which parts are technology problems with technology solutions and which are people problems with…well, you get where I’m going here.

Everyone loves a makeover

“The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency.” —Bill Gates

“Digital transformation” is the consulting term for overhauling a business operation to stay afloat as everything is being digitized and interconnected. This term became popular because “Oh shit! What the hell is even happening?” doesn’t fit on a Powerpoint slide as neatly.

The tangle of problems called digital transformation may be the thickest and thorniest for those established enterprises who make up the core client base of management consultancies, but absolutely every business or not-for-profit organization of any size has to confront continuous change.

Consider the cognitive load of the average videoconference, or even of choosing the best method to communicate in a given instance. Do you email, Hangout, Zoom (whose Zoom?), Slack, stop by, just…phone? (“Oh shit! What the hell is even happening?”)

Attempts to radically change processes and product offerings to incorporate new technology tend to go sideways. And the first misstep is trying to fix people problems with new technology solutions. It turns into Godzilla vs Megalon, battle of the personal agendas, and the organization emerges feeling trampled with nothing to show for it but incinerated budgets and crushed dreams.

In reality, we still have a limited awareness of the full implications of digital technologies. Organizations of every size have to move deliberately, overhauling their thinking and approach based on the underlying values. And because organizations are—for the time being—made of people, those values should be conversational.

Going conversational

“Every complex system starts out as a simple system that works—and the simplest coherent system is a feedback loop.”—Amy Jo Kim, game designer and entrepreneur

Adopting technology and managing change must start with the simplest system that works for people: the conversation.

Being conversational means applying the deeper principles of how humans interact with one another to create systems that succeed on human terms. It does not mean—necessarily— talking to a computer. It does not mean ditching all documentation. A good conversation is an exchange that begins with a fundamental level of agreement and proceeds towards a shared outcome.

Any two strangers who speak the same language can strike up a spontaneous conversation because it’s the basic interface between people. Conversation tends to follow a set of basic principles and works best when it is:

• Cooperative
• Goal-oriented
• Context-Aware
• Error-tolerant
• Turn-based
• Polite

People who engage in conversation tacitly agree to cooperate towards the implicit purpose of the exchange by taking turns and helping each other through rough patches. An awareness of the surrounding context (what we call “reading the room” even when there is no room) is essential. We converse much differently with a stranger in distress than with an old friend over coffee.

If you have had the experience of interacting with a voice or messaging interface (or a human) only to become extremely frustrated, that means it wasn’t truly a conversational interaction at anything but the most superficial level.

You know what system has been embodying these principles for the entire 21st century? Google Search. That is a humdinger of a turn-based feedback loop that delivers meaningful value and aligns business and user goals. Machine learning has only enhanced the initial core interaction.

In order to get under the surface, here is a starter set of six principles to consider.

1. Face humanity

We all have to live in primate bodies on this one planet in an interconnected ecosystem and economy. That isn’t hippy-dippy bullshit. That’s level zero for all human endeavor. If being conversational is being context-aware, there is your context. Keep it in mind at all times, because it’s the one constant.

Optimizing solely for shareholder value gets you a pretty minimal local maximum. Solving the technology problem is meaningless unless we connect it to human needs. Because no people, no business. (And take the rest of the ecosystem into account. I don’t mean to be completely speciesist.) Ray Bradbury knew what was up.

Individuals and organizations make bad decisions when they refuse to acknowledge how irrational and interdependent we are, usually out of fear or ego. Technologists and business leaders need to make the same leap economists did decades ago with behavioral economics and acknowledge that “rational human decision-making” is a lie.

The most rational thing to do is take how irrational we all are as table stakes.

2. Designing with humans for humans means qualitative understanding is critical

On the Internet, nobody knows that your Fitbit is attached to a dog.

Paradoxically, the fact that we are slapping sensors onto every activity and taking measurements means that it will be ever more difficult to determine which measurements matter. Companies and their consultants tend to figure out what is easiest to measure and ascribe some fabricated meaning after the fact. Decisionmakers use these numbers to justify (or “validate”) particular courses of action.

There is no substitute for observing the world around you and talking to people where they are, to get a deep holistic understanding. Only then will you understand the why behind the how much, see what is meaningful, and identify the opportunities to add value. Otherwise it’s too easy to interpret the numbers using one’s own biases and then be surprised when things don’t add up. All the quantitative data in the world won’t suddenly generate meaning. Unexamined, it will just introduce bias at scale.

3. Orient business and design around an explicit exchange of value

“Empower” is the trendy business word for giving employees the tools to do the thing that serves business goals. “Delight” is the trendy design word for giving users the good experience you hope also serves business goals. Value is what we should really be talking about.

The most avoidable reason ambitious initiatives fail inside and products fail outside is the failure to ask the question “What’s in it for them?” about every single individual whose participation you need for a successful exchange within a system. And of course to listen and act according to the answer.

Is it status, belonging, pleasure, survival? See above about everyone being human.

Conversational design is value-oriented design. Too many business and design decisions are still authority-oriented. Someone with power says a thing and everyone else acts as though it’s true (for status, belonging, pleasure, survival within that system) and everything goes up in flames, and not the good campfire kind.

This isn’t hard to figure out conceptually. I’ve gone through the process with dozens of organizations. Talk to the people inside to see what’s in it for them. Talk to representative users and customers to see what’s in it for them. Look at the overall context, available resources, and relevant technology, and then balance the equation.

It is hecking hard to put into practice operationally. The truth is that business goals, individual employee goals, and customer goals rarely line up perfectly or well, with sufficient resources to serve all. So you need to decide what to optimize for. But you can only make rational trade-offs when you look at the whole system using the best available evidence and a realistic assessment of your role.

And as Nathan Shedroff pointed out:

“Value can only be exchanged within a relationship — therefore: no relationship, no value.”