Adaptive Path MX 2007: Interview with Irene Au, Director of User Experience, Google

Jeff Veen interviews Irene Au, both of Google

A: Background was in EE. She went to U of I for grad school and got into human factors. Worked on Mosaic.

Then went to Netscape to work on Communicator 4. Then to Yahoo. Netscape was interesting but really they were just working on a viewfinder but the interesting stuff was inside the viewfinder.

Rocketmail had become Yahoo Mail and My Yahoo was brand new. They were moving beyond the directory of web sites that started. There were a small number of designers who were called on to quickly turn around graphics at the end. The challenge was to get designers involved earlier.

Now Google for the last four months and the idea is to do the same thing.

Google is like Disneyland. But apart from the food and the fun stuff, there is a tremendous opportunity to bring design thinking to the organization to radically change the organization. It's so innovative and now the challenge is to take that to the next level.

V: The design culture came out of the search product. That's an easy interface challenge. All the difficult interaction decisions are in other products like mail clients.

A: Strategies deployed in the past work for search but not for more complex experiences where you are trying to tie it together.

They used to look for a generalist with a background in HCI with an eye to aesthetics who would code his own stuff.

That works for search but not for experience design. You need a cross-functional set of disciplines and people who can work across these to move users from one thing to another.

V: The challenge then is to maintain quality. All the resumes and hiring must make it hard. How do you maintain quality?

A: Need to set expectations around different skills. People who are anthropologists and psychologists. Best visual designers. But avoid siloing. Hiring T shaped people with deep experience in one are but some experience in lots of areas.

V: So prioritization. There are so many initiatives under development. How do you prioritize.

A: The laws of physics don't allow everything at Google to get covered. You have all these engineers and PMs with ideas. People are empowered to bring those to life but there isn't enough bandwidth to work with designers. Need to be more thoughtful about how engagement happens. The solution is to do fewer things really well. Do you want the design team to be a shop or a high-end strategic asset to the organization? This requires careful choice of projects with meaningful impact on the business. Must also take into account how amenable the project team is to working with the design team. Must get involved early and up front.

Also look for opportunities to offer goodwill to the project teams. Have a chance to upsell on more high impact activities. Getting the foot in the door to introduce new ideas.

V: So that happens at product and project level. What about the org?

A: Managing a portfolio. Part is projects, part is infrastructure, part is blue-sky thinking. Speaking to the trojan horse thing, the wedge to use to engage with the teams is very interesting. At Yahoo people had never watched users using products. The usability studies were a great voyeuristic view into how people were using the products and the developers ate it up. This turned into what are you trying to do, what is the opportunity. Over time people start to understand and learn. Just keep asking and trying.

V: So what about at Google?

A: Sometimes its the mockup. Teams feel that they can't launch with out the UI. Looking at mock-ups of a project you ask about use-cases and sometimes nobody has stepped back to ask those questions. If you just ask people how they are dealing with this issue a design can emerge out of that.

V: At AP, always talk about speaking the language of higher-ups. At Google that language is engineering. Hard to communicate a social network to engineers.

So a little bit about selling design. What about when a team is really struggling?

A: Miscommunication is often the cause. A designer is left out and a PM does mock-ups. First step is to explore the problem. Understand the politics, the stakeholders. Maybe its just really opinionated people who believe that designers have no value to add. If they believe that design is easy it is hard to communicate design rationale to them. But that's where prioritization comes into play. Is it better to staff a project with no opportunity for success that has a high priority or a low-priority project with a high opportunity for success. Spreading the team thin on projects where they aren't set up to succeed doesn't serve anyone well.

V: Pretty flexible about how that happens within the organization. Some people are aligned to metrics and others are more personality oriented.

A: Important to be flexible and adaptable to the culture of the organization.

Culture of advertising related projects is very different than other products. They are much more top down. If they are more structured around driving priorities, the design team can use that to their advantage. If other project teams are more free-flowing the design team must use that factor in context.

Q from audience: How do you bring design into this kind of organization?

A: The view was that this was an important role but they don't necessarily have a clear vision. Evangelize up and down.

V: There is also a next step about the integration of design.

A: There isn't an understanding of how to use ethnography, for example, so the question becomes about how that space can be created.

In a previous life she had an experience where someone just wanted to copy the competitor. They convinced them that this was a bad idea by doing rapid ethnography. Analysis, brainstorming, etc. can lead to the transformation where people can see a new vision appropriate to the situation.

At Google often focus on rigor but also need to focus more on the softer skills. How do you influence teams, etc.

Q: What we all need is meeting facilitation skills. But to the question, many people find themselves at a point of success, they are managing teams, but should that group be centralized or out in the business units?

A: Now had experience with two or three rapidly growing companies. Early in the life, centralization is important. The team is smaller, it is cross-functional requiring different kinds of people with different skills who can share best practices and standards and methods over time. This prevents reinventing the wheel and leads to higher-end activities. At the same time, user experience is always better when tightly integrated.

V: So Google is really centralized. There is one way to right code. In the design group they are now working on style guides. How did this play out at Yahoo?

A: At Yahoo the matrix structure invested people in their products. Yahoo Sports tried to compete with ESPN, etc. and this fractured the organization. It was hard to focus central resources on a style guide. Getting 20% time on a style guide was really hard and the idea of a centralized tax was hard to push for. One style guide was the universal look and feel. Consistency without uniformity. Message Yahoo as a company without making the products the same. The second component was a design pattern library, a catalog of best practices for interaction design. Best way to do pagination or table design, etc. Parts of this are publicly available. The third part is the UI library, now YUI (and very useful), this allowed a more coherent experience.

Q: Changes in the hiring practices at Google. Are they connected to the presence of designers? Is that required for change?

A: The changes are very subtle. In UE team the point is clarifying expectations for people. And an interview process that gives the interviewer more certainty about their experience with the applicant. In the past interview scores were inconsistent based on whatever they happened to talk about. Now seeing more consistency .

V: Do people use design exercises? Engaging people in design exercises can be really handy if done well. Not just leaving them in a room but working with them to see what happens.

A: Also had a design exercise at Yahoo. Needed something that would allow people to express what they are about. It's hard to find where people are in the spectrum. If you can't get people who can do everything get people who compliment eachother and give them flexibility.

The design exercise isn't to see if they are good but to see what they are good at.

V: What about mentoring? There are lots of young people at Google. How do you approach that?

A: It is important to set aside time for this. In programming people talk about pair programming. What about pair designing? When this happens it is often in pockets. How do you get the spread around, to facilitate communication within the team? Need to be more open and forthcoming.

V: Office hours is great. Managers have office hours to get one on one time. Put designers in a room and tell people come in and get design questions answered.

A: Office hours is part of Google's culture. It's hard to get used to a long line of people but its nice to have time blocked out.

V: Design office hours, accessibility office hours, etc. gives time for answering question.

A: Fix-its are great, where everyone drops everything to fix things. Also great is testing-on-the-toilet. When you go into bathrooms at Google you have toilets with warm seats and a screen with a video about unit-testing.

Q: What about cross-platform consistency? As you go to mobile, etc. how important is consistency.

A: There is consistency in look-and-feel and consistency in interaction. Not everything needs to look the same everywhere but it is important to find the elements that communicate things as a family.

V: That's a point about the difference between mobile and web. A spreadsheet on the phone is different. What am I doing with a spreadsheet on my phone?