Sunday, November 20, 2016

Thoughts on Groupthink

During the past couple of months I‘ve done quite some reading on cognitive biases. Wikipedia states that “[a] cognitive bias refers to a systematic pattern of deviation from norm or rationality in judgment, whereby inferences about other people and situations may be drawn in an illogical fashion.” As I read more about biases and how our brain works, I had to realize that humans are incredibly bad in making rational decisions. We often believe that we analyze a situation carefully, then evaluate the pros and cons of a decision and decide for the option with the best cost-benefit-ratio. In fact, that‘s not true. And even worse: Our brain tricks us into thinking that it is true! If we can believe the many findings of neuroscience and social psychology, our decisions are heavily influenced by factors that we are not even aware of.

What is groupthink?

One of these factors is the very strong human need of relatedness. When people feel that their belonging to a group is in danger, it‘s very threatening to them. In fact, our brain works pretty much in the same way it did 10,000 years ago. And back then, it actually was a death threat when someone was excluded from a group. The thing is: When we feel threatened, a couple of very archaic mechanisms kick in, which tend to overrule all rational reasoning.
So humans have a very strong tendency to conform with a group. This leads to groupthink, where “the desire for harmony or conformity in the group results in an irrational or dysfunctional decision-making outcome. Group members try to minimize conflict and reach a consensus decision without critical evaluation of alternative viewpoints, by actively suppressing dissenting viewpoints, and by isolating themselves from outside influences."
Groupthink is also called the Bay-of-pigs-effect, because the disaster of the invasion to Cuba is believed to be caused by groupthink: People self-censored their doubts about the plan, because they felt it was not appropriate to contradict the predominant opinion in the group. Fortunately, Kennedy learned his lesson and - less than one year later - had mechanisms in place to avoid groupthink and make much smarter decisions during the Cuban Missile Crisis. Other really severe incidents like the explosion of the Challenger Space Shuttle have also been linked to poor decision making based on groupthink.
Some people say that World War III has been avoided during the Cuban Missile Crisis, because Kennedy had learned how to avoid groupthink and embrace dissent. Image credit: doe-oakridge on flickr
Now think about a group in the context of knowledge work - could be a team at a planning meeting or retrospective; could be a management group talking about strategy; could be a committee organizing the next christmas party. When you use the same approach I used to use in my role as a moderator, chances are that the decisions of these groups are flawed. Why is that? Common moderation techniques rely on writing topics on cards, clustering them and eventually having the group dot voting the topic they want to tackle. Nothing wrong with this, but the devil‘s in the details!

A real-life example

Let‘s look at a real-life example of a workshop I was observing. The group had been talking about areas for improvement and how good they thought they already were (from 1=excellent to 5=very bad). You can clearly see big clusters and very little variation.
When the moderator saw this, he said something like: “Excellent, you are all very aligned here. Let‘s go on with the topic you‘ve rated worst.”
The problem was that people dot-voted one after the other or in small groups. So the later a person voted, the more votes were already visible. In such a situation, according to neuroscience, it becomes more and more difficult to give a completely different vote. In this context it‘s extremely interesting to look at Asch‘s Conformity Experiment, where people gave obviously wrong answers to a simple task, because they wanted to conform with the group.
When I think back to the moderations I did, I did not take this much into consideration (and probably others don‘t, as well). So what can we do about it? I think first it‘s important to acknowledge that groupthink (like all cognitive biases) is not a bad thing in itself. It serves an extremely important purpose. But in many of our modern contexts, it becomes an obstacle.

What can we do about it?

Because of their evolutionary importance, biases seem to be hard-wired into our brains. Therefore it‘s of little help to tell people not to groupthink. Here‘re some things I think are more useful to reduce the effects of groupthink:
  • Let people vote simultaneously, if possible.
  • Let people think about the topic individually or in small groups. Let them write down their opinions/votes and only after that collect all the individual votes. This gives them the space to think about the issue with very little influence from the big group. Also, if people have written down their preference on a piece of paper one minute ago, it‘s quite hard for them to deviate from this preference shortly after.
  • When you decide to do a circle, where every group member voices his/her opinion, make sure that the boss and obvious opinion leaders speak last.
  • Try to make the group as diverse as possible. Diverse groups have proved to be less vulnerable to groupthink.
  • Try to add people to the group, who are likely to have a different opinion than the majority of the group.
  • Invite an outside expert to the group. Outsiders are less vulnerable to groupthink (especially when they know they are only temporarily a member of this group). This is one big advantage of consultants, especially when you tell them they are paid, because they are less likely to conform with the group.
  • Observe closely, which members are quiet during a workshop or meeting and explicitly encourage them to voice their opinion, even if it‘s controversial. Usually we assume that the quiet ones agree with the group. But what if they are the ones who disagree but are afraid of isolating themselves from the group? Think about what would help those people to gain confidence and openly disagree.
  • Explain the role of devil‘s advocate to the group and assign this role to one person. The explicit role makes it much easier to speak up, because people know it‘s “the role” speaking, not necessarily the person (and people might dislike the role, but they won‘t dislike the person who disagrees). Actually I think the “10th Man Doctrine” was the one cool thing in the movie World War Z. It says: “whenever 9 people looking at the same information come to the same conclusion, it's the 10th's duty to disagree and actively look for evidence to the contrary.” Although this doctrine is not a real thing, I was surprised to learn that the Israeli military seems to have a devil‘s advocate office, which task it is to ensure that “intelligence assessments are creative and do not fall prey to group think.” (see this thread on Quora).
  • Think about rules and processes that help dealing better with groupthink. One such rule could be: “If everyone in the group agrees, we assume we have overlooked something and we should actively look for different angles.”
  • Be very aware of signs of pressuring dissenters. Phrases like “Are you with us or not?”, “Are you a team player or not?”, “You‘re either in or out!” should get all alarm bells ringing.
  • Brief someone to wear the clown suit. I stole this metaphor from this blog post about lonely dissent. The author states that “[l]onely dissent doesn't feel like going to school dressed in black. It feels like going to school wearing a clown suit.” Modifications of Asch‘s experiment have shown that chances to disagree with the group grow dramatically as soon as at least one other person has disagreed before. This first person is the one with the clown suit. Look for someone with enough confidence and standing to wear the clown suit. And if you consider yourself a leader, think about wearing the suit yourself.
  • As a leader, reward deviating from the group‘s opinion. I am not talking about monetary incentives here. I think it can be extremely powerful to praise someone in front of the group for wearing the clown suit.
What are your experiences with groupthink? I am happy to read your comments below!

P.S. If you find cognitive biases interesting, you might like my posts Thoughts on survivor bias and Thoughts on confirmation bias

Tuesday, April 19, 2016

An Alternative View on Company Structures

For years my thinking about company structures went like this: “The more structure a company has, the more it sucks.” So I was arguing against putting new structures in place, whenever I heard of such ideas. Of course I knew, that no company could exist without any sort of structures. So my credo was: Let‘s only have structures, where it is absolutely necessary. And in my mind it only became necessary when we were growing and the new size made it necessary to come up with new structures, because otherwise things break down..
Potential downsides of rigid org structures are well known: Less freedom for the workforce, and hence less engagement and less innovation; Dilbertesque policies that might make sense for some part of the organization but not for the rest of it; single points of failure due to hierarchical pyramid structures etc.

Back to my credo: “Let‘s only have structures, where it is absolutely necessary”. I still think it‘s valid, but here comes the catch: I‘ve realized that there are other things than sheer growth in headcount that can make it necessary to add more structure! Here are three things I think are worth taking into consideration: Diversity, Fairness and Health.

There‘s plenty of research (see this article for more resources) that shows how more diverse groups make better decisions. When people think about diversity, they mostly think about gender. While gender is very important and there is a lot to improve, especially in the tech industry, we should also think of diversity in terms of age, race, cultural and economic background, sexual orientation, political preferences, family situation, etc.
When I talk about diversity with my colleagues, we very often end up with the idea that we should have more structures. There are two reasons for this:
  1. If we want to have a more diverse workforce, we have to change the way we recruit and hire people. For a long time the way I did job interviews went something like: Let‘s have a coffee together and afterwards we do a thumb vote if we want to work with this person or not. If you use a process like this, you can be 100% sure, that your decision is affected by all sorts of cognitive biases and that you are biased against diversity. Like one of my colleagues put it nicely: “Having more diversity in a group feels like a grain of sand in the gear.” It feels uncomfortable. Humans are hard-wired to prefer being with people who are like them. And this is exactly what we want to avoid when we talk about diversity. One countermeasure for this are structured job interviews. Google does this rigidly, as Lazlo Bock (Head of People Operations at Google) presents in his book and this article. What are structures interviews? Not only are the questions for a job  interview formulated beforehand, but there‘s also a definition of the types of answers that are considered to be good/mediocre/bad. Or, as Bock puts it, a structured interview is “a consistent set of questions with clear criteria to assess the quality of responses”. And Google even goes one step further: The hiring decision is not made by the people, who did the interview, but by an separated committee. Sounds like a lot of structure, right? It certainly does to me and I was terrified, when I heard this for the first time. But if we take diversity and de-biasing seriously, this might be the (or at least one) way out.
  2. As soon as we become more diverse in our workforce, we might also need more structure. My colleague Boris just shared his thoughts on this with me: “If we were all clones of each other, we wouldn‘t need any structures. Everyone knew exactly what the others think, how things work and what the next steps are. But this would be zero diversity. If we have people with different backgrounds, we need more explicit structures, otherwise people get lost.” This totally makes sense to me. If all your employees are 30-year old, left-winged white male surfer dudes without kids, you probably don‘t need much structure, because their thinking might be very aligned. And if they face a problem, they will find a way to work around it easily, because every evening they‘re drinking beer together. So this scenario is very comfortable, and although it‘s intentionally exaggerated, I think a lot of start-ups work in a similar way. It might be okay, or even necessary for a small company to operate in such a way, but if the company is growing, and especially if it‘s critical to improve the quality of decisions, increasing diversity becomes very important. And that increased diversity comes with the need for more structures.
Every organization has implicit and explicit structures. And probably it‘s a good heuristic to say that the less explicit structures you have, the more important the implicit structures become. That can be considered unfair, because it favours those who are good in navigating through blurry structures. Often the best (or even the only) way to get things done in such a context might be having good personal relationships with the most influential people in the company. For new people it can be really hard to join this game, because the rules are...implicit. And it gets even worse, as soon as you get more diversity. Imagine that, for the first time, a single mother joins the company. She probably does not have the time nor the interest to hang out every other evening with her colleagues.
In this classic feminist article from the 1970s, Jo Freeman argues that so-called structure-less groups are undemocratic, because they tend to be dominated by elites, who are not accountable to the larger group: “For everyone to have the opportunity to be involved in a given group and to participate in its activities the structure must be explicit, not implicit.”
I am not very familiar with the feminist movement, and the groups Freeman talks about are political groups, not companies. Still the argument makes a lot of sense to me and made me think.

While organizations with little (explicit) structure provide a lot of opportunities, they also might make it easy for people to jeopardize their health. The reason for both, the good and the bad aspect of little structures, is what I would call “anything goes”. If responsibilities, decision-making-mechanisms, team structures, career paths etc. are unclear, the organization might be able to exploit the advantages of fast decision-making (“if it‘s not clear, who makes this decision, let‘s just decide in this group - right now”) and fluid teams (“let‘s team up and build this thing”). On the other hand, the same context might encourage people to work in an unhealthy way. By this I not only mean the sheer amount of hours they work, but also the effect of over-commitment and mental overload. Because it‘s unclear, where my responsibility ends and what the company expects from me, I might take on everything I find interesting or important. Of course I can only do so many things, but I am very ambitious, and nobody stops me from starting all these exciting things. So maybe I should start working a little bit longer every day and think about all the interesting problems at the weekends and during my holidays?

What now?
I think it‘s important to realize that structures in itself are neither good nor bad. And I am not making the case for excessive structures. Like with many things, it‘s a permanent trade-off decision we have to make. Instead of falling in the trap of binary thinking (“all structures are good/bad”) we should think about the pros and cons of adding more structure and then find a healthy balance for our context. My impression is that in the Agile/Lean community we have very much focused on the downsides of structures in the past. Maybe it‘s time now to take the upsides into consideration a little bit more.


Like this post? Then you should check out my previous post Radical Transparency? and one of my newer posts Seriously, what is a Pull System?