Cognitive Design

  Affordable, reliable and perhaps even sexy heads sets that read your brain waves and algorithms that translate them into signals that control computers, games and other devices.

When

 As mentioned in previous posts, hopefully yet this year. NeuroSky has released technology to its OEMs and has announced deals with Sega and others. Check out their latest gear and application ideas here.

Emotiv Systems released this picture of their headset:

 This headset has 16 sensors (medical grade gear has 120) and the NeuroSky’s system amazingly has only has one. They are targeting December of 2008 with a retail price of $299.

Other firms with products in the pipeline include EmSense, and Hitachi.  There is additional action outside the US. Check out the headset on the g.tech BCI research platform that won the European ICT prize in 2007:

The system (shown below) has been billed at the first commercially available brain-computer interface and has receive a good deal of buzz. Although not a mass market “consumer friendly” system (more of a research platform) it has been used to update the classic game of pong:

 Cognitive biases are strong (some say hardwired) tendencies to process information, think or interact in ways that can (but don’t always) create systematic errors.  Well known cognitive biases include the confirmation bias (I seek out new information and ideas that support my current beliefs) and the halo effect (I tend to judge all aspects of something based up one or two dominate characteristics that overshadow the rest). There are dozens of interrelated biases that have been documented and many studies that illuminate them at work in how doctors think, investors make decisions, people chose mates, teams function (or not), managers hire employees, consumers make choices and many other domains.  

   Effective decision-making requires managing cognitive biases especially in high stakes situations.  This point was made vividly by Phil Rosenzweig in his recent book the Halo Effect  and the eight other business delusions that deceive managers.   The book highlights the role that cognitive biases and other systematic errors in thinking have played in our attempt to understand and improve the performance of organizations. He outlines nine such “delusions” about high performance organization, including the halo effect: 

   “The tendency to look at a company’s overall performance and make attributions about its culture, leadership, values and more. In fact, many things we commonly claim drive a company performance are simply attributions based on prior performance.” 

   Just as we have a bias to think that good looking people must be smart, have an interesting life, or be successful, we think high performing organizations must have strong cultures, good leaders or the right values.   He quickly dashes our hope that research could settle the question (Good to Great, In Search of Excellence, etc.) by arguing the professors and consultants that do the research fall prey to other types or errors in reasoning including for example, confusing cause and effect (or correlation and cause). So does employee satisfaction generate organizational performance or does organizational success generate employee satisfaction? We are often told the former when it fact some research shows the latter. This would mean our attempts to improve organizational performance by focusing on employee satisfaction (a popular thing to do) could be wrongheaded. This does not mean we don’t want to focus on employee satisfaction but doing so under the theory that it will drive organizational performance may be more of the halo effect, or confusing cause-and-effect than anything else.  

  If  Rosenzweig is right (I have seen all the delusions he quotes in action in many different circumstances) then our inability to manage cognitive bias has wrecked (or at least limited) our ability to improve our organizations through research and managerial will.

  We need a more detailed understanding of the role that cognitive biases play in organizational improvement and then we can use that understanding to “design against” them in our research and implementation work. 

In nearly every talk I give on cognitive design someone asks about designing for trust. This includes offerings to create and maintain the trust of customers as well as management practices to create and maintain trust with employees.

This is a great design question to ask because trust is a complex cognitive relationship we establish over time with people and a wide variety of artifacts.  My answer takes the form of a simple recipe:

A.  Be sure to include the 7 features of trustworthy products in a way that customers feel 

   

B.  Give products and services a trusting “personality” when possible

 

C.  Establish a service recovery process that restores the customer’s perception of justice 

 Research shows I tend to trust products/services that are:  

• * Reliable (perform in a predicable way across circumstances)

• * Transparent or easy to understand

• * Under my control (options, personalization, easy termination)

• * Secure and safe (won’t hurt me or let me hurt myself)

• * Insures my privacy (to the degree I want it)

• * Have guarantees (a form of promised service recovery)

• * Are self maintaining or automatically  serviced by the provider

On the surface, most of these features seem to be a matter of engineering and usability. But how I establish the cognition (feeling) for each of these factors is key and very much depends on the industry context. For example, years ago Progressive Insurance started offering a free and easy to use quoting service that shared the cost of their policies and the cost of their competitor’s policies even when Progressive’s policies were more expensive. This was a bold move in establishing the cognition of trust in an industry that has had low levels of consumer trust.  It clearly telegraphed – we are not trying to hide our prices (transparency) – and gave consumers a level of control and understanding that was otherwise hard to obtain.  They establish a mutual interest with consumers – understanding the comparative price of policies even to their potential detriment. If we have a mutual interest, and you show a reasoned willingness to go may way on occasion, I can trust you.

Products with personality or those that have features that cause me attribute human-like qualities to them may play a special role in creating and maintain trust. There is some data that suggest people see simple geometrically styled products as “sensible and trustworthy”. Other more personified products that talk to us (cars) or are stylized after living or imaginary characters (e.g. cartoon characters) may in fact invoke cognitions that lead to higher levels of trust (my speculation).  The development of avartars or 2-dimensional representations of personalities for customer service (an animated figure that moves, gestures and speaks with you) over the web is one experiment along these lines.  Well designed service avatars run on a touch of artificial intelligence and do well at understanding natural language. Check out the article about Jenn a service Avatar for Alaska Airlines.

We can now develop customized avatars (and buddy icons) to express dimensions of our own personality for many online interactions (email, instant messaging, social networking, games, etc.). This is powerful cognitive design medicine.  After all, if I can’t trust myself (even as an avatar) who can I trust?

Cognitive design for trust is really put to the test during service recovery. When a product or service fails, consumer expectations and trust are betrayed. How I recover form this betrayal strongly determines how much I should be trusted in the future (assuming the breach is not consistent or acute) Organizations seek to recover that trust by “making things right again” through a service recovery process.  Some researchers argue that service recovery is mediated by emotions. According to the appraisal theory of emotions – emotional states are generated when we assess (or appraise) a situation’s fit with our goals, beliefs and values. Clearly, in the case of a service failure the essence of the situation is a negative appraisal and attending emotional states. But how can we characterize these emotions in a way that helps us design a better service recovery process? 

An interesting a recent answer to this question is that we can characterize the design problem in terms of restoring “perceived justice”. According to Chi Kin (Bennett) Yim and co-authors, consumers want three types of percieved justice in a service recovery process: procedural (a compliant process that is easy to access, fast and transparent), interactional (treated with fairness, empathy, courtesy during the process) and distributional (compensated for the service failure).  The study suggested that positive outcomes in service recovery (e.g. continued customer loyalty) were mostly driven by distributional justice or the compensation given because of the failure.  From a cognitive design perspective this is not surprising as it demonstrates that the organization is “putting its money where its mouth is”.  

Being nice to me is one thing (and expected) but showing that you are willing to give me something that reflects a shared value is a big step in re-establishing trust.

Working out a calculus for distributional justice is no easy matter. Over compensation during service recovery can lead to consumer guilt, under compensation can lead to anger.  Time-to-compensation and who delivers it (e.g. high ranking official or the clerk that made me mad) may also be important variables.  The corporate mindset may also impose the ideas that all customers must be treated identically during service recovery.

 The art of cognitive design for service recovery is found in the specifics of the what-and-how customers are compensated for a failure. It need not be complex – Domino’s Pizza is delivered in 30 minutes or it is free.

 Blunders in service recovery may lead to serious customer defection (although I have found no research to support that). Service recovery is like a second chance it the relationship. You mess that up and how can I trust you?

Traditionally designers and human factor specialists are very focused on functionality and usability. We want things that are useful and easy to use.  As we move into the next century more and more designers are looking to go beyond usability concerns to the systematic design of pleasurable or enjoyable products. The idea is to design for specific states of mind (mentality) in the user – pleasure, enjoyment and so on.   

Two excellent books that take this approach include Designing Pleasurable Products  by Patrick Jordan and Funology: From Usability to Enjoyment edited by Mark A. Blythe and others. Both books represent important efforts to design for mentality (user’s state of mind). They borrow a little from the cognitive science (or related) literature and quickly develop practical tools and methods for doing design.  Although productive, neither book provides (or purports to provide) the systematic review of what we know about how minds work (cognitive/neuro sciences) to effectively make the following shift:

Functionality à Usability à Mentality

I use the word mentality rather than pleasure or enjoyment because it denotes that the design challenge may be broader – encompassing any state of mind that is needed to drive more value from the artifact. No matter which word you use there is an important shift from a word that describes the artifact (functional and usable) to one that describes the mental states of the user. This is consistent with a more human-centric approach to design. I am not just designing to meet the “needs of users”, instead I am designing a cognitive artifact that integrates both the functional state of the product and the mental state of the user in hopefully a fully symbiotic relationship. Another advantage of defining Mentality as the goal is that it can also refer to a state of the product/artifact as functionality and usability do. Mentality in a product/artifact means I am designing smarter machines and other artificially intelligent artifacts. As described in the post below, Donald Norman sees smarts machines designed to work symbiotically with humans as the future of design.

The next step in design then is to “design for mentality” in humans by achieving a specific state of mind and in the artifact by adding functionality that makes it smarter.  

Check out Donald Norman’s latest book, The Design of Future Things.  His basic thesis is that in the future the things we design will be smart (emotionally and intellectually) permitting a more natural form of interactions (one similar to a horse and rider) and even achieving a level of human-machine symbiosis. It is people and their artifacts working together – with machines augmenting the capabilities of humans that make up the design of future things. Homes, cars, workspaces, airports, robots and other artifacts will be designed to work with us at three levels including the visceral (automatic and unconscious responses), behavioral and reflective (conscious self awareness).

 He offers six rules for designers of these smart machines of the future: 

1. Provide rich, complex and natural signals

2. Be predictable

3. Provide good conceptual models

4. Make outputs understandable

5. Provide continual awareness without annoyance

6. Exploit natural mappings.

 Being predictable and running on a good conceptual model means I can understand (to a degree) the artifact’s workings, know how it will behave and also come to trust it especially since its outputs are understandable. Rich and natural signals that support my continued awareness of (and feedback about) its operation without annoyance makes it is an active part of my environment. These are many of the same things that make my interactions with other people work.  

 All of these rules effectively lower the cognitive load associated with the artifact without sacrificing functionality. There is a higher degree of integration between the functional states of the artifact and the mental states of the people using the artifact. Indeed, in the case of a symbiotic relationship they are fused together in a dynamic feedback loop.  This represent the fifth level of cognitive fit (1= agitate, 2 = tolerate, 3= resonate, 4 = accelerate and 5 = integrate) that we have discussed elsewhere in this blog.  

 One example Professor Norman gives is that of a Cobot being developed at the Laboratory for Intelligent Mechanical Systems at Northwestern. In this example, humans can move heavy payloads (e.g. automobile engines) in complex environment smoothly and easily by sharing control and intelligence with a robot.  The robot provides not just mechanical muscle but also brains by providing natural resistance to suggested motions that are not consistent with the task or could be unsafe.  All of this takes place NOT via a controller to manipulate a robot arm but by directly interacting with the artifact (e.g. by wrapping a chain around the automobile engine to lift it). This is consistent with rule #6 – exploit natural mappings.

 In a provocative afterwards, Professor Norman describes an imaginary interview he had with an intelligent machine about a set of design rules they (intelligent machines) developed to improve their interactions with people.  We will review those rules and their implications for cognitive design in a future post.

As mentioned in a previous post, peer production is an emerging model for highly-distributed knowledge work (making predictions, writing encyclopedias, developing software, etc.) that for the most part does not involve direct compensation (pay) and results in products/outputs that rival the very best that are created using traditional proprietary production models. Linux (software operating system), Wikipedia (encyclopedia), Flickr (photographs) and Second Life (objects for a virtual world that sell for real dollars) are popular examples.

What motivates knowledge workers to produce for no pay? This question is especially interesting when you look at open source development (talented software developers).  Very good research on this topic can be found in the year old MIT press book Perspectives on Free and Open Source Software edited by Joseph Feller and others.  The entire book is online and the first two chapters deal with a profiling and understanding the motivation of open source developers.  They identify eight possible motives that could explain why open source developers participate in a project including:

 Altruistic – this is a socially important, a way I can give back

Profit – way to make money

Fame – increase my stature in my peer group and perhaps the world

Marketability – signals my skill to the market

Utility – work with others to fix a problem I have

Enjoy – doing this is fun, like going to the movies or hiking

Development – this is like a free university or mentoring program

Ideological– all software should be free, help break monopolies

  They claim all are at work and provide a “big tent” embracing many needs.  A survey revealed that the primary motivation was creative expression (working on the most creative project in their career) and social interaction including learning and developing new skills, sharing knowledge and skills, participating in a new form of cooperation.  If software is your passion working in new ways on important/creative projects and getting to build your chops while helping others would likely be too much to pass up.

  The open software development model and other modes of peer production are excellent examples of how web-based mass collaborations can satisfy deep cognitive needs in talented knowledge workers. Satisfying cognitive needs (versus economic needs) is what turns the engine of peer production. This puts the cognitive designer in the cat bird seat when it comes to creating new peer production models.

I am looking at an X-box 360 game controller. It does not resemble anything else I use and has 11 buttons and 3 mini joy sticks on it.  It takes two hands to operate. I am looking at the Wii controller. It resembles a remote control and has only 5 buttons and 1 mini joy stick on it.  It takes one hand to operate. Talk about a difference in cognitive load.

The Wii uses similarity and functional simplicity. No wonder it is such a hit.  When you use Wii to play games, you use your body just as you do in real life. This is especially evidence with sports games when you swing your arm holding the controller to hit a tennis ball or roll a bowling ball.  In terms of cognitive fit during learning, the Wii resonantes or even accelerates.  The 360 controller will likely cause mental agitation.

Of course this depends upon the user. A serious gamer that already has a mental model for operating a sophisticated game controller would likely have little problem learning the 360 controller and experience toleration rather than agitation during the learning phase. Once the 360’s advanced functionality has been mastered, the serious gamer likely experiences cognitive acceleration.

The Wii on the other hand has redefined who a gamer can be. People have no fear of picking up something like a remote control and playing simple body-based games. Check out this news story that describes Wii bowling as big hit at a retirement home in  Chicago (average age 77).  Wii seems to be increasing access to the gaming experience for potentially millions of new players.

How far will Wii go? There is now a Wii internet channel where the hope is you will use the controller to browse the web. Google has created a version of its reader for the Wii. One analyst from Merryl Lynch predicts Wii will be in 30% of US households by 2011. So the Wii may not bee a fad driven by a novelty effect or a niche device.

I know comparing Wii and the 360 is like comparing apples and oranges– different type of games, different target market. But what the Wii shows is the power of using design to satisfy unmet cognitive needs. The Wii is simple, resembles something we already understand (metaphor greases the cognitive wheels), makes use of my body (remember cognition is embodied) and lets me get to the fun faster by myself or in a group (emotional energy). All of these factors dramatically lower cognitive load. It trades off graphical quality and game play sophistication and is able to give a lower price. Some people begin to experience cognitive dissonance (holding conflicting beliefs) when they consider paying a lot for a game or game device.

Why is the Wii so popular? It is a masterstroke in cognitive design.

Peer production or the development of valuable products and content by users over the Internet for free is not a fad.  Web-based “mass collaboration” is producing quality encyclopedias, production-level software products, authoritative books, professional-quality citizen journalism (news photos, stories and even shows), stock picks that beat the market and even assisting in the search for extraterrestrial life.  

Talented people are working hard, really hard, and regularly for free (no direct economic compensation).  Collaborating, innovating and problem-solving like mad – all thing we have tried to make happen on the other side of the fire wall using knowledge management with little real success for the last 20 years.  

Why are they doing this and how can we harness it inside the firewall?  Asked another way, what is the cognition that drives peer production (user generated content) and how can we apply it in the workplace?   This is a timely question because corporations are starting to invest in employee applications of Web 2.0 technologies in the hope of stimulating productivity and innovation. Results might be disappointing if we don’t understand the motivation and cognition that is driving the behavior. Fortunately, there has been a little research.

McKinsey has published several papers on Web 2.0 and how corporations can make the most of user generated content. (You need to sign up for a guest pass to access the article.)  One key finding:

“We observed that users cite a variety of reasons for posting content online—chief among them, a hunger for fame, the urge to have fun, and a desire to share experiences with friends.”

Recognition, fun and sharing with friends is good but it is really powerful when it is driven by the modifiers “hunger”, “urge” and “desire”.   More fundamental psychology than what we normally see in the workplace. And I don’t think they are getting that because of the functionality of the tool or even the topic that is being worked on. It is more the law of large numbers. Because the task is cast over the Internet it is possible for those few people that are really highly motivated to “self select” and participate in a robust way. You will likely not be able to duplicate that effect in corporations, even very large corporations.

That does not mean that blogs, wikis, social networking and the like won’t improve communication in corporations. They will. What it does mean is that Web 2.0 technologies should not be viewed as a knowledge management silver bullet. Instead we should see them as another tool that needs to be considered very carefully from a cognitive design standpoint.