Designing for Interaction: An Interview with Dan Saffer

If you've been delighted by your iPod, intrigued with your TiVo, or frustrated by your mobile phone, then you have encountered the work of an interaction designer. And an interaction designer, most likely, has crafted the experience we have with many of the products and services we encounter every day. Dan Saffer, a senior interaction designer at Adaptive Path, leads us through an exploration of this emerging discipline. Published this month, Saffer's new book, Designing for Interaction, is a much-needed primer on the topic, helping us understand the design of interactive systems. Voice talked with Saffer just prior to his book being published in July.

Liz Danzico: How would you describe interaction design? And why is it important to write this book now?

Dan Saffer: I have a fairly expansive view of what interaction design is, which is that interaction design is about people: how people connect through products and services. Now, what does that mean?

Interaction design is about behavior, how things work. I push a button on my mobile phone and something happens. Or I enter a fast food restaurant, walk up to the counter, and something happens. Defining what happens when a person uses a product or service is what interaction designers do.

The reason we do it is to enable connections—interactions—between people. People want email and instant messaging and their mobile phones to be easy and fun to use. They want their trips to the DMV to be pleasant and efficient. They want the check-in kiosks in airports to work smoothly and well. All of these things—and many, many more—are about connecting people and helping them communicate better between themselves and the world.

This book is important now because we need new interaction designers and people who understand what interaction design encompasses. Technology is spreading into all corners of our lives, whether we want it to or not. Political, social, and economic forces are making it so. In order to make all this new technology useful and usable by humans, it needs to be designed with humans in mind. That's where interaction designers come in.

Danzico: In the book, you point out that Bill Moggridge (a principle at IDEO) was the first to call the practice “interaction design.” Haven't we always been designing for interaction? Why is interaction design, as you (and he) describe it, new?

Saffer: Bill Moggridge and his colleague Bill Verplank at IDEO realized in the late 1980s that they had been doing a different kind of design than what was traditionally called “graphic design” or “industrial design,” so they gave it this name (which is much better than their alternate choice: “SoftFace”). But in my opinion, it's something we've been doing since before recorded history. Aboriginal peoples made cairns to mark trails—that is, to communicate through time via a product. Native Americans used smoke signals to communicate over long distances.

The only thing new about it is that now, thanks to microprocessors being embedded into all sorts of objects that can now exhibit all sorts of different behaviors, it's been recognized as a discipline. Somebody needed to figure out how these newly empowered objects should behave, and the tools of design were well-suited for it. Now, you can study it in school, and get paid to practice it. Whereas before, like other types of design, it was simply done without much reflection.

Danzico: In a recent interview with Brian Oberkirch at Weblogs Worknotes, you describe interaction design by saying: “The discipline that makes technology useful, usable, and fun to use. Good engineering is what makes it happen. But interaction design is what makes it approachable for people to use.” Is interaction design just about technology, or can it involve other types of products?

Saffer: I was giving the easy answer. It's not only about technology, but these days it often is. Most interaction designers work on software, websites, and other technology like mobile devices. But interaction designers can also design services which have little to no technology in them. By services, I mean processes and ways of doing activities. So you see interaction designers working in retail environments, figuring out flows of the store. Interaction designers work for the Mayo Clinic, changing how health care services are delivered. You even find interaction designers working with government agencies, making the system of paying taxes, say, better for people.

Of course, services can be a combination of technology and non-technology. Netflix, for example, has its website, but it also has the envelopes that the DVDs get mailed and returned in. Someone designed that service.

Danzico: What was your first experience with interaction design? In other words, was there a time where you saw interaction design emerging as a thing separate from other design disciplines?

Saffer: My first experience with interaction design took place when I was a teenager in the mid-1980s, about 15 years before I ever heard the term “interaction design.” I designed and ran a game “online,” meaning users dialed in to my Apple IIe using their 1200 baud modems. Of course, I had no idea what I was doing at the time.

But around the mid-1990s, others certainly knew what was happening. Carnegie Mellon established its interaction design program in 1994. Agencies started offering it as service (albeit often mislabeled as “information architecture”), and software companies started hiring people for these roles. Right before the internet bubble burst, interaction design started to come into its own, and it began to get known. In 2003, Alan Cooper changed the subtitle of his seminal book from The Essentials of User Interface Design to The Essentials of Interaction Design. Also in 2003, the Interaction Design Group (now Association) was formed as a professional organization for interaction designers. So it has some traction now.

Obviously, “interaction design” is still not a term you hear often, and probably never will be. But thankfully, “Design” with a big D covers it pretty well.

Danzico: Can you give a good example of a typical interaction design that we're all familiar with?

Saffer: The Automatic Teller Machine (ATM) is something that most would be familiar with—at least most people who might be reading this. An ATM facilitates interactions between banks and their customers. It has an interface—both the digital screen and the physical structure—that has been designed for privacy and rapid transactions by a wide variety of people with a broad range of familiarity with technology. My grandfather—deaf, in his 80s, never owned a computer—used his first ATM only a few years ago. ATMs do a remarkable job of turning the complexities of banking into some clear choices, usable by large segments of the population.

Danzico: I was really surprised by your pointing out that “user-centered design” is only one of four approaches an interaction designer can take. Can you talk about one of these four approaches: what you call “genius design?” At first, it might seem counter to the things we were taught as good researchers and designers, where it was important to do diligent user research.

Saffer: It is counter to what we're told today is good design practice, but I deliberately tried not to judge any of the approaches to interaction design, to include all the ways you can practice interaction design whether I agreed with those methods or not. I find myself moving through most of them frequently, often on the same project. Each of the approaches has produced great products over the years, and perhaps none more so (because it is used the most often) than what I call “genius design.” Genius design is when the designer relies on his or her own experience and skill to design, without any input from users. It's done by designers who either don't have the resources or the inclination or temperament to do research. Too often, it is practiced by inexperienced designers with little skill, but it can and has been used by many designers to create impressive things. Reportedly, the iPod was made with no user research, for example.

Danzico: When have you used the genius-design approach successfully?

Saffer: More often than I care to admit. In the past especially, I've worked on projects where there was no time or money or willpower to do any of the other approaches. I just finished designing Soundflavor, a music application and accompanying playlist-sharing website with the genius design approach, and I'm pleased with the results thus far. Of course, even if you do have the resources and inclination for one of the other approaches, I find there are always moments on every project when I employ genius design. I have hunches and make educated guesses based on previous experience. One could argue (and many have) that this is why people hire designers: for this sort of genius.

Danzico: Why is it important to design hackable products?

Saffer: That's a good question: I'm not sure it is important. People will hack your products anyway! That being said, leaving “seams” in your product for people to customize it to suit their needs is a very interesting practice.

Danzico: Seams?

Saffer: As designers, we're traditionally taught to get out of the way of the product, to leave no trace of ourselves or how the product was made. Think of the iPod in its hermetically sealed case, for instance. But Matthew Chalmers had this idea of “seamful systems (with beautiful seams)” where, for those so inclined, you could see and take advantage of how the system was created and adapt (hack) it for your own use. Seams afford hacking, in other words.

Companies can get new ideas for new products through exposing the seams and affording hacking, and could even repurpose their existing product to take advantage of the modifications people are doing to it. Of course, it's also a dangerous practice. People can hack things in dangerous ways that could open up the companies to serious liability issues. If they are going to build in seams for hackers to rip open, designers need to make sure just what it is exactly they are exposing. On a financial website, of example, it's one thing to expose the CSS so that someone could change the colors of their version of your site. It would be quite another thing to expose users' financial data!

“The idea that we as designers control any product is a myth.”

Danzico: For some time, people have been able to hack their TiVos to view their flickr streams on their televisions. Next, you might imagine a similar hack for YouTube videos, streaming on our TV as well. With users having this much control over the design of their environment, where does the interaction designer's role start and end? Are interaction designers in danger of losing control?

Saffer: The idea that we as designers control any product is a myth. It's a useful myth, to be sure, since it allows us to actually make the product. But once it is out of our hands and out into the world, we can no longer control what people do with it. Sure, we can design how we hope people will use it, but there's no guarantee they will use it that way.

The interaction designer's role is one of facilitating particular uses for a thing, and possibly dissuading other uses. I will design X so it can be used for Y. If someone uses it for Z, well, that is his business. The problem comes when Z is something harmful. If I design a hammer, and someone uses the hammer to bludgeon someone, how responsible am I? Think of email: we want to design email clients so that they are easy to send and receive emails. But you don't want to design them to enable spammers to easily send out tens of thousands of messages. Not that spammers use email clients, but you get the idea.

Danzico: In your book, you build a nice definition of interaction design by saying, “It's about making connections between people through these products, not connecting to the product itself.” What do you mean by “making connections between people?”

Saffer: Traditional industrial design is about making a connection to an artifact: This is a great chair. Traditional communication design is about making a connection to information: Yes, I will attend the event this poster is advertising. Human-Computer Interaction is about connecting with the computer: I enjoy using my Mac OS X operating system. But interaction design, although it draws on all these fields (and many more), is subtly different in its purpose: to connect people via our products and services: I know you better because I read your blog.

As I think about it, an interaction is really a communication. It can either be one-to-one, like a telephone call. It can be one to many, like a podcast or a blog post. Or it can be many-to-many, like a giant system like the stock market. All these things are surrounded by tools that make the communication possible, and those tools, for the best experience, should be designed.

Danzico: Is good interaction design visible? In other words, is the success (or failure) of interaction design something we talk about and point to? How can we recognize good interaction design?

Saffer: The visible part of interaction design is the interface, which is usually the controls for manipulating the features and functionality that make up the interaction design. Interface design is only the physical expression of interaction design. The interaction design part of a product or service is usually invisible. However, it can be felt. The iPod would just be a beautiful object if it also didn't work well. And certainly the failure of interaction design can cause anger, frustration, lost time, and, in the worst case scenarios, injury and death.

In the book, I list the characteristics of good interaction design, things like trustworthy, appropriate, and smart. Things that are hard to visualize, although there are certainly visual cues for these things. And users certainly notice, usually unconsciously, both their absence and their inclusion. My mobile phone, for instance, is a beautiful piece of industrial design. But the interaction design is terrible. I simply can't use it easily and well to make phone calls and do all the other things a mobile phone does these days. It annoys me and causes me angst and embarrassment. It is the opposite of another trait I mention: clever. It doesn't anticipate any of my needs and tailor itself to help me accomplish them.

Danzico: Do interaction designers need to be good graphic designers? How much cross-over is there between the visual and the functional?

Saffer: No, although it certainly helps, as it would to be a good industrial designer. On small teams, often the visual designer and the interaction designer will be the same person. And even when each role is played by a different person, there is a constant back and forth. I was recently on a project where my interaction design called for four buttons on an application's interface. The visual designer came back to me and said, “Due to X, Y, and Z, I've only got room for two buttons.” So then I had to tailor my design to fit his. And of course, since my work was done first, he had already had to tailor his design to work with mine.

What visual and interaction designers have to collaborate most on are the affordances of the interface: those things that indicate how the product could be used. The visual cues users rely on to understand what they can do with a product: push a button, turn a dial, and so on.

Danzico: What are the ways that we might train interaction designers differently from a non-interaction designer?

Saffer: For the most part, I think interaction designers should be trained the same way most designers should be: taught to draw and model and prototype, about typography and visual principles. And, most importantly, to problem solve.

But also in the same way that industrial designers need to understand the properties of, say, metal and plastic (their materials), I think it helps interaction designers to know how the technologies they use work. An interaction designer working on the web should know about how web pages are made, for example. Not that they should be programmers necessarily, but knowing what the medium you are working in can do is immensely valuable. The difficulty in teaching this is that those things change rapidly and it is hard to keep up for even people working with it every day. Plus, in school, you aren't certain what medium you might be working in afterwards. Tricky dilemma.

I also think more experience with writing is helpful. Both creative and technical writing are valuable, and interaction designers use both very frequently, for scenarios, storyboards, documentation, and so on.

“Services are the new frontier of interaction design.”

Danzico: You talk about the field of “service design.” Can you describe how service design is becoming more important to designers?

Saffer: We are coming to a time, if we aren't already there, when most products aren't stand-alone. They are part of a broader service. My mobile phone has a service plan. My television has a cable service and TiVo hooked up to it. Even the tea I buy (Peets: delicious!) can be ordered online. The point is that most products have to be viewed as part of a broader context: a service. Designers have to pay attention to the environment, the processes around the product, and a new set of users: the employees providing the service. Services aren't only about end-users: they are co-created by service providers (employees) and customers.

Services are the new frontier of interaction design.

Danzico: How are new technologies influencing the sphere of influence that designers have?

Saffer: The history of design can be thought of as the history of materials. Now that we've gotten this new material—the digital—wherever it goes, hopefully we'll go as well. Bill Moggridge says, for instance, that the reason interaction designers are getting involved in services now is because technology is involved in them. There are opportunities everywhere for interaction designers, in all areas of life. RFID and similar technologies are about to change the way we shop, cross international borders, and find objects. Robots are in our homes now, vacuuming floors. People are wearing devices on their arms to monitor their bodies. And the internet...well, don't get me started.

With all this technology, we really can't help but have an influence on people's lives, on public discourse, on the future of the planet. Certainly, I don't want to overstate the power of designers (that's been done enough lately), but I don't want to understate it either. We're almost an invisible force, shaping the tools that shape us as human beings, to paraphrase Marshall McLuhan.

Danzico: You write, “To design is to make ethical choices. In other words, design is ethics in action.” Is it really the designer's responsibility to make ethical choices, or should those come from the client? How can a designer know the “right” thing to do?

Saffer: Ideally, ethical choices come from both the client and the designer, but I don't think designers can rely on the client for those. Since many of our clients are for-profit companies, the lure of filthy lucre can sway even well-meaning companies from doing good. And certainly, designers aren't immune to money either.

Design, being mostly subjective, isn't usually filled with clear-cut “right” answers unfortunately. There's just the scale of better and worse. And this is why it is so difficult. Having professional codes of ethics, like those promoted by AIGA and IDSA is a starting point. They provide a baseline, to which the designer can bring his or her own ethical beliefs for further enrichment.

For interaction design, I've proposed that the quality of the interaction be our ethical baseline, for both the instigator of the interaction and the receiver. It should be easy and pleasurable for users to both send and receive email or instant messages. If it isn't (aside from technical issues), then something is probably wrong, ethically. If I design a service that only benefits one side (the business, say, or even only the users), then I've probably failed in my duty as a designer.

Danzico: You've developed an interesting model with your book's site. Not only is it available as a Rough Cut through Peachpit, but you've posted the full interviews and interview excerpts on your site. Was your intention to demonstrate interaction design as well as talk about it, or was that just a nice coincidence?

Saffer: It was by design. I deliberately wanted to show the seams of the book as I was working on it, both for marketing and to get interaction designers involved. Since this is a new field, I certainly don't have all the answers. I wanted feedback as I went along on the book. I also knew that in the final book I only had a limited amount of space for my interviews, and I wanted every part of the interviews to appear somewhere. So the website became a place where I could do that: share ideas the interviewees had generously given to me, but that were probably for space reasons not going to make it into the book. All my interviewees are amazing and had really great insights, and they really took the book to a higher level. So in a sense, the website is just a way to extend the book, stretching it out to create (hopefully) more interactions.