Who is smarter – a diverse group or the best expert? Can a boss, a project manager or the most experienced specialist make better decisions than a whole team? What are the problems with decision making in traditional corporations?
James Surowiecki in his book The Wisdom of Crowds argues that given the right conditions a group consistently outperforms best experts in decision making. Few people can do better than the group, but each time they will be different people. A group answer will often be at least as good as an answer of the smartest member.
The author provides interesting examples:
- 800 different people (mostly non-experts) tried to guess weight of an ox at the International Exhibition of 1884 in London. The average guess of the crowd was only one pound off (ox weighed 1,198 pounds) and was better than any estimate of the cattle experts.
- Stock market knew what company was responsible for the Challenger disaster within a half hour of the shuttle blowing up. Mostly uninformed investors dumped shares of Thiokol almost simultaneously.
- Team of the men with wide range of knowledge (e.g. mathematician, submarine specialists, salvage men) correctly estimated location of disappeared U.S. submarine Scorpion. The submarine was found 220 yards from collectively estimated spot (original search area was 20 miles circle) and it was better than any individual guess.
Crowd is wise when these criteria are present:
- Diversity of opinions (each person have some private information)
- Independence (people’s opinions are not determined by the opinions of others)
- Decentralization (people are able to specialize and draw on local knowledge)
- Aggregation (some mechanism exists for turning private judgments into a collective decision)
Groups vs. Experts
- Groups are better at deciding between possible solutions to a problem than coming with alternatives.
- Crowd intelligence is most useful for broad skills as decision making, problem solving or forecasting unknown future.
- Experts judgment is most useful in specialized areas as programming, design, and even management, where experience, skills and talent are important.
- Experts lack “judgment calibration” – they often overestimate their knowledge and performance (except maybe weather forecasters).
- Advices and predictions even from the best experts should be pooled with others to get the best answer.
Diverse groups vs. Homogeneous groups
- Diverse groups add new perspectives, wider set of possible solution and approach problems in novel ways.
- Homogeneous groups, particularly small ones, are often victims of “groupthink”:
- Information that might represent challenge to the conventional wisdom is either excluded or considered as mistaken
- Homogeneity fosters pressure toward conformity. A member changes own opinion as it is easier than to challenge the group.
Independence vs. Imitation
- Independence is important to intelligent decision making for 2 reasons:
- It keeps the mistakes that people make from being correlated.
- Independent individuals are more likely to have new information.
- Lack of independence brings following problems:
- “Social proof” – tendency to assume that if lots of people are doing something or believe something, there must be good reason why. The crowd becomes more influential as it becomes bigger. The governing assumption is when things are uncertain, the best thing to do is just to follow along.
- “Information cascade” – to supplement their information people will look what others are doing. If the first couple of people get bad information they will make wrong choice. Everyone who follows, even if they have better information, will make the wrong decision too.
- “Herding” – sticking with the crowd and failing small, rather than trying to innovate and run the risk of failing big.
- However, imitation works much of the time. Each person can’t know everything. With imitations, people, can specialize and the benefits of their investment in uncovering information can be spread widely when others mimic them.
- Imitation also requires little top-down direction. It is a powerful tool for spreading good ideas fast.
- Imitation work when there are:
- Initially wide array of options and information
- Willingness of at least some people to put their judgment ahead of group’s.
Decentralization vs. Top Down Authority
- In terms of decision making and problem solving these things matter:
- Specialization – increases the scope and the diversity of the opinions and information in the system.
- Tacit knowledge – can’t be easily summarized and conveyed to others, because it is specific to a particular place or job or experience, but it is very valuable.
- Decentralization great strength – encourages independence and specialization while still allowing people to coordinate their activities and solve difficult problems on the other.
- Great weakness – no guarantee that valuable information will find its way through the rest of the system.
- Decentralized system can only produce genuinely intelligent results if there is a means of aggregating the information of everyone in the system (e.g. price for free market).
In the traditional corporate model, top management hires the best employees it can, pays them and gives some direction about the problems to work on and hopes for the best. It is great for mobilizing people and get very good at doing things they know how to do. But it is also necessarily limits of the number of possible solutions (company has limited number of workers and their time) and because of reality of organizational and bureaucratic politics.
Coordination and decision making problems when power is concentrated on top
- Costly in terms of time.
- Requires too much information
- Saps initiative of workers and mangers
- Discourages free flow of information
- Incentives skewed against dissent and independent analysis
- Lack of diversity among top managers, insulation of top managers from competition
- Elimination of rigid managerial hierarchies
- Wider distribution of real decision making power
The virtues of decentralization:
- More responsibility people have for their own environments, the more engaged they will be
- Coordination – instead of orders and threats, companies rely on workers to find new, more efficient ways of getting things done. That reduces need for supervision, cuts transaction cost and allow managers to concentrate on other things. Toyota Production System is a supreme example.
Tacit knowledge (emerged only from experience) is crucial to efficiency of markets and as just as important for corporations. Companies should start with the assumption that, just as in the market place, people with local knowledge are often best positioned to come up with a workable and efficient solutions. The benefits of specialization and local knowledge often outweigh managerial experience in decision making.
How these ideas could be applied to the Software Development? See the next post.