Headless products make research the design work

More and more, people don’t use a product by opening it. They ask an assistant, or an agent does the task for them, and your product is somewhere in the background doing the work.

People call this headless. For a design leader it raises an obvious question. If there’s no interface, what are we designing?

My answer is that we’re designing how well the product understands what people want, which makes research the main work.

What a screen used to tell you

When someone uses an interface, you learn a lot for free. You can see where they clicked, where they hesitated, where they left. A bad design shows up in the data within days.

Take the screen away and most of that goes quiet. A request comes in, a result goes out. If the result wasn’t quite what the person wanted, you may never find out. They just don’t come back, or they ask somewhere else next time.

That worries me more than any design trend, because the feedback we’ve all relied on for years is disappearing.

Research has always been where my best work came from

Looking back, the projects I’m proudest of all started with time spent with people.

At Flow we did one-to-one interviews and group sessions with renters, and we shadowed residents through ordinary moments like rent day and repair requests. We ran workshops where neighbors told us what would make them proud of where they live. The principles for the whole product came out of that.

At Movoto, the homeowner product grew out of a research study into what owning a home is like after the purchase. That product went on to bring in more than a million new monthly users.

At Digs, we interviewed lenders and agents until we understood what they were afraid of, which was losing clients during the long quiet stretch before a purchase.

In every case the interface came later, and it was the easier part.

Why headless raises the stakes

With a screen, you can be a bit vague about what people want, because they’ll steer themselves. They’ll find the filter, try another tab, work around you.

Without a screen there’s nothing to steer with. The system has to understand the intent correctly the first time. If someone asks an agent to “find me homes like the last one I liked”, every word of that needs to be understood: what “like” means to them, what they liked about it, what they’d never accept.

You can only build that understanding by knowing your customers extremely well, and that takes research.

How I run it

A few things I ask of my team.

• Everyone does research. Designers, and ideally engineers and product managers too, sit in on conversations with customers. Nobody waits for a separate department to hand over a deck.
• Go and watch. What people say they do and what they do are different. Shadowing someone through their real day beats any survey.
• Study the misses. The most useful cases are the ones where the product gave an answer and it was wrong. Those are hard to find in a headless product, so we go looking for them on purpose.
• Write down intent. We document what people are trying to get done, in their words, with the edge cases. That document is the design.
• Keep it continuous. A study once a quarter isn’t enough when the product is changing every week.

What changes for a design team

The output looks different: fewer mockups, and more maps of what customers are trying to do, clear definitions of a good result, and examples of right and wrong answers that engineers can build and test against.

The skills shift too. I look for designers who are curious about people, who listen well, and who can turn a messy set of conversations into something clear. Those people have always been the best designers, and now the work depends on them.

And design gets a bigger seat. When understanding the customer is the product, the people who understand the customer best should be in the room where the decisions are made. I see it as my job to get them there.

The part I keep coming back to

I think headless products make design’s job bigger.

When the surface goes away, what’s left is the question underneath every product: do we understand the people we’re building for? I’ve built my career on taking that question seriously, and I’m glad it’s becoming the whole job.