UX design & research at Microsoft

Last week, CHI Nederland organized a presentation by two dutch members of Microsoft's user experience team in Redmont: Sander Viegers (design) and Arne de Booij (user research). Microsoft also recently featured in a series on managing design by the British Design Council:

"In 1993 design was a luxury.
It is now generally accepted that design is critical to our success"
(Brad Weed, UX Director at Microsoft)

This emphasizes the big changes since the early days of MS Office, when most developments were technology driven. However, with ongoing progress, user expectations keep growing, too.

At Microsoft, rather than employing a corporate design department, individual designers are embedded in the engineering teams. According to Sander, this allows for a stronger influence of the designers as part of the team, and a better feel for, and communication with, development. However, the challenge is to provide consistency across products. This is facilitated by several central resources, such as the UX excellence group, the central UX support team, Microsoft research, etc.

The design process at Microsoft consists of three phases:

• Understanding
  (what users want/need and which steps are required)
  
• Prototyping
  ("show what you mean")
  
• Production
  (work with developers to ensure final products conform to specifications)
  

All members of the team (not just the designers!) are involved in brainstorming and creating a large amount of iterative prototypes. This allows for cross-pollination and stimulates team work; "The more contributers with different backgrounds, the higher the chance of coming up with new approaches and great solutions."

Sander suggested the following strategy to overcome the perception that designers make things look better, whereas in fact they make things work well:

• Embrace
  (go along with current perception and process, in order to "get your foot in the door")
• Prototype
  (show different options, demonstrating the value of good design)
• Finish
  (leave your audience enthusiastic and eager for more :-)
  

As designer you need to "choose your battles". The challenge is to have a clear vision of the entire user experience, and steer towards that aim throughout smaller detail sessions. Sander called on designers not to end up with a "mediocre average" based on a multitude of input, complexity, and opinions.

"Technology can master complexity, but design must master simplicity."
(Surya Vanka, UX Manager at Microsoft)

Arne provided a short overview of user research at Microsoft. Amongst other means, Microsoft often uses survey data to identify the greatest frustrations users experience with their products. A general "Dynamic Customer Model" has been defined (and is continuously refined), based on numerous interviews, site visits, and historical user research sources. This global model comprises ca. 50 persona's, their relationships amongst each other and a task model of their most important work.

When usability testing a new design, 10-15 users of the same profile as each main persona are invited to perform the most crucial tasks defined for this persona. On a scorecard the measure of success, time, bugs, etc for each task are recorded and compared against the goals.

Arne pointed out that in order to set measurable goals, it is important to set specific aims (e.g. "80% of all users can accomplish this task within 1 minute" instead of "the task is easy to accomplish "). He also warned of showing film fragments of usability tests, as short scenes can often be misinterpreted or given too much weight by developers or managers not trained in the subject domain and unable to see the "total picture".

Microsoft has recently been involved in some interesting research regarding the three aspects proposed by Don Norman:

• self image (marketing)
• appearance (visual design)
• pleasure & effectiveness (interaction design, user research & usability)
  

For each product, this may mean very different things for different people in different circumstances. Sometimes effectiveness is essential, while other times image or emotional value (e.g. memories) might be more important. A good balance needs to be found for the intended audience. In order to achieve this, we need to know what the intended audience is and understand them well. Particularly the aspect of 'pleasure' is very difficult to measure and design for.

In order to measure visual design, and to establish effect, impact and consequence of visual appearance, Microsoft is currently looking to develop new models together with several universities. The aim is to measure what users think about current products, what they find generally appealing, and what effect the overall design has on them.

One approach relies on images and associations; When users are asked about products, generally the answer focuses on smaller details rather than the 'overall picture'. To overcome this problem, a set of 25 mood images were developed with Copenhagen IT University. The test subjects are shown screenshots of a specific design and are asked to choose 5 mood pictures that best match this design. Next they are asked to explain these choices. This helps users to disassociate from details, allows them to 'think differently', and results in more emotional answers (e.g. "peaceful", "boring", etc). The designs are then adjusted according to the feedback. Through iterative prototyping and testing, the overall visual design can be improved towards targets such as "powerful", "innovative", etc. However it still needs to be established, what the (commercial) importance of such research will be for (business) users.

Currently the common pyramid of useful -> usable -> desirable is still seen as a sequential or incremental process. Hopefully in the future this should be a concurrent and complimentary process. The aim is to provide a continuum across all touch points (home, office, web, devices, etc)

On the question what it is like to work at Microsoft, Sander pointed out that it is a very large company and "the people you work with can have a bigger influence than the problem you're working on". According to Sander, most employees at Microsoft are very passionate and driven. To get a better impression, he recommended the following resources:

• View my World
• UX Design at Microsoft (incl. product design)
  
• MSDN Channel 9 (a look behind the scenes)

It seems that Microsoft has a similar setup as I noticed at SAP: designers embedded in the product teams, supported by central resources and centers of excellence. Both companies are dealing with complex software and a large user base. They also both have been working with Alan Cooper in the past. Of course Microsoft bears the additional burden of having many novice users and often being a trend setter for third parties.

Their recent move on Vista and Office to deviate from established standards and risk initial decrease in productivity and client satisfaction to ultimately hopefully achieve a much better user experience were quite brave. I'm certainly very interested to see how their research into appearance and emotion pans out! An opportunity to catch up to more aesthetic and design-driven companies, such as Apple?

Considering the call for traditional design disciplines at the recent CHI symposium in Amsterdam, and the work by groups such as Haring institUte of Happiness and Partners-InC, there certainly seems to be a strong trend towards aesthetically pleasing and emotionally engaging solutions. Should be fun :-D

[This report was also published by CHI Nederland]