How to make decisions as a team

Some of the clearest evidence for writing first comes from research on brainstorming. Brian Mullen, Craig Johnson, and Eduardo Salas combined the studies in 1991 and found that groups brainstorming together produced fewer ideas, and fewer good ideas, than the same number of people working alone whose ideas were pooled afterwards⁵5. Mullen, B., Johnson, C., and Salas, E. (1991). Productivity Loss in Brainstorming Groups: A Meta-Analytic Integration. Basic and Applied Social Psychology, 12(1), 3–23. doi.org. The loss was larger in larger groups, and when ideas were spoken aloud instead of written down.

Writing first does not make the discussion less useful. Joaquin Navajas and colleagues asked a live audience of 5,180 people general-knowledge questions, such as the height of the Eiffel Tower⁶6. Navajas, J., Niella, T., Garbulsky, G., Bahrami, B., and Sigman, M. (2018). Aggregated knowledge from a small number of debates outperforms the wisdom of large crowds. Nature Human Behaviour, 2(2), 126–132. doi.org. Each person answered alone first, then discussed the question in a group of five and agreed an answer, and then answered alone again. The groups’ agreed answers were substantially more accurate than the average of everyone’s first answers, and the average of the answers from just four groups was more accurate than the average of thousands of individual answers. In that study, each person had formed a view before hearing anyone else’s.

We recommend the same order for a team decision. State the question, and give everyone two minutes to write down their view and the main reason for it. Read the views out before any discussion, with the most senior person reading theirs last. Discuss the question only when every view has been heard, and let people change their minds. It takes about five minutes.

In 1985 Garold Stasser and William Titus gave groups of students information about the candidates in a simulated election⁷7. Stasser, G., and Titus, W. (1985). Pooling of unshared information in group decision making: Biased information sampling during discussion. Journal of Personality and Social Psychology, 48(6), 1467–1478. doi.org. Taken together, the information showed which candidate was best. But it was split up so that each student’s share made a different candidate look better. The groups could have put the full picture together by talking. Instead, their discussions were taken up by the information they all had already, and by information that supported what they already thought. The groups chose the candidate with the most support at the start, and discussion reinforced the students’ wrong impressions instead of correcting them.

Researchers call this a hidden profile, because the best choice stays hidden until the group pools what its members know. Li Lu, Y. Connie Yuan, and Poppy McLeod combined 65 studies of hidden profiles, with 3,189 groups⁸8. Lu, L., Yuan, Y. C., and McLeod, P. L. (2012). Twenty-Five Years of Hidden Profiles in Group Decision Making: A Meta-Analysis. Personality and Social Psychology Review, 16(1), 54–75. doi.org. Groups mentioned far more of the information they all shared than the information only one member had. Groups facing a hidden profile were eight times less likely to find the best answer than groups in which everyone had all of the information. The more of the unique information a group brought out, the better its decision.

Jessica Mesmer-Magnus and Leslie DeChurch combined 72 studies of information sharing, with 4,795 teams, and found that sharing information predicted better team performance in every kind of study they looked at⁹9. Mesmer-Magnus, J. R., and DeChurch, L. A. (2009). Information sharing and team performance: A meta-analysis. Journal of Applied Psychology, 94(2), 535–546. doi.org. A structured discussion and a cooperative team both made sharing more likely.

In a team, the information only one person has is usually about their own work, such as a customer’s complaint, a supplier’s warning, or what happened the last time the team tried something similar. Ask for it directly. Before the team decides, ask each person what they know about the question that the others might not, and start with the person closest to the work.

Managers are often told that conflict about the task is good for a team and that conflict between people is bad. In 2003 Carsten De Dreu and Laurie Weingart tested it with a meta-analysis¹⁰10. De Dreu, C. K. W., and Weingart, L. R. (2003). Task versus relationship conflict, team performance, and team member satisfaction: A meta-analysis. Journal of Applied Psychology, 88(4), 741–749. doi.org. They found that both kinds of conflict had strong negative correlations with team performance and with how satisfied members were. The negative link was stronger for complex work, such as decisions and projects, than for production work.

The picture has changed since. Frank de Wit, Lindred Greer, and Karen Jehn combined 116 studies with 8,880 groups in 2012¹¹11. de Wit, F. R. C., Greer, L. L., and Jehn, K. A. (2012). The paradox of intragroup conflict: A meta-analysis. Journal of Applied Psychology, 97(2), 360–390. doi.org. They found stable negative links between group outcomes and conflict about relationships or about how the work is organised. They did not find a strong negative link between conflict about the task and performance. Disagreement about the task went with better results in top management teams, when performance was measured as financial results or the quality of decisions, and when disagreement about the task did not turn into disagreement between people.

The disagreement also has to be real. A devil’s advocate is someone asked to argue against a proposal whatever they think of it. Charlan Nemeth, Keith Brown, and John Rogers compared a devil’s advocate with a person who genuinely disagreed, and found that genuine dissent produced more and better solutions than any form of devil’s advocate¹²12. Nemeth, C., Brown, K., and Rogers, J. (2001). Devil’s advocate versus authentic dissent: Stimulating quantity and quality. European Journal of Social Psychology, 31(6), 707–720. doi.org. In an earlier study, a devil’s advocate had mainly made people defend their first view.

For a team, this means asking the people who disagree to explain why, as early as possible, and keeping the discussion about the options. Say what you disagree with and why, and not who is wrong. When a discussion turns personal, stop and go back to the question.

Paul Nutt studied 358 decisions made in medium and large organisations in the United States and Canada, and found that half of them failed¹³13. Nutt, P. C. (1999). Surprising but true: Half the decisions in organizations fail. Academy of Management Executive, 13(4), 75–90. doi.org. The failures could be traced to managers who imposed a solution, limited the search for alternatives, and used their power to push their plan through. Decisions were more likely to succeed when the manager made the need for action clear at the start, set objectives, searched widely for solutions, and got the key people to take part. The tactics most likely to fail were used in two decisions out of every three.

A team decision can be made in several ways. One person can decide alone, one person can decide after hearing the team, the team can vote, or the team can decide by consensus, which means everyone has to agree. We recommend the second for most decisions. One named person decides, after hearing everyone’s view. Usually this is the person who owns the problem, or the person who leads the team when the decision affects all of it. Say who it is before the discussion starts, so everyone knows whether they are being consulted or asked to decide. The guide to deciding who owns what covers how to give each area of work one owner.

Consensus gives every person a veto, which makes hard decisions slow. A vote turns the discussion into a count of two sides. Both have their uses, and neither suits most decisions about a team’s work.

Hearing everyone before deciding is also a matter of fairness. Jason Colquitt and colleagues combined 183 studies of fairness at work¹⁴14. Colquitt, J. A., Conlon, D. E., Wesson, M. J., Porter, C. O. L. H., and Ng, K. Y. (2001). Justice at the millennium: A meta-analytic review of 25 years of organizational justice research. Journal of Applied Psychology, 86(3), 425–445. doi.org. How fair people thought decisions were, including the way they were made, was related to their job satisfaction, their commitment to the organisation, and how they rated the people in authority.

Once the decision is made, everyone acts on it, including the people who argued against it. If new information comes to light, anyone can ask for the decision to be reopened. Having lost the argument is not a reason to reopen it.

Once people know how something turned out, they remember it as more predictable than it was. In 1975 Baruch Fischhoff gave students a short account of an event with four possible outcomes and asked how likely each had been¹⁵15. Fischhoff, B. (1975). Hindsight ≠ foresight: The effect of outcome knowledge on judgment under uncertainty. Journal of Experimental Psychology: Human Perception and Performance, 1(3), 288–299. doi.org. Some were told nothing about the ending. Others were told that one of the outcomes had happened. In all 24 comparisons in the experiment, being told an outcome made people rate it as more likely, by 10.8 percentage points on average. They were largely unaware of the effect, and in a second experiment they overestimated what they would have known without being told.

This is the main reason to write a decision down on the day it is made. If the decision turns out badly, the team is likely to remember the risks as obvious and the reasons as weak. If it turns out well, they are likely to remember it as the only sensible choice. A written record keeps what the team knew, what it chose, and why.

A decision record needs six things.

• What was decided, in one or two sentences.
• Why, including what the team knew at the time.
• What else was considered, and why it was not chosen.
• Who decided.
• Who was asked.
• When the decision will be reviewed.

Record whether the decision is a one-off, which covers this case only, or a standing rule, which applies from now on. Keep standing rules together where anyone can find them, so the team can check whether a question has already been decided.

Gary Klein’s premortem is a way to find the risks in a decision before it is final. The team imagines that the decision has been made and has failed, and each person writes down why. Klein recommends it because many people are reluctant to raise their doubts about a plan while it is being made, and imagining that it has already failed makes those doubts easier to say¹⁶16. Klein, G. (2007). Performing a Project Premortem. Harvard Business Review, September 2007. hbr.org.

Set the review date when the decision is made. At the review, compare what happened with the record, and not with what people now remember. Judge the decision by what the team knew at the time and how carefully it decided, because a good decision can turn out badly and a poor one can turn out well. A review is a kind of debrief, which the guide to running a weekly team meeting covers.

For a small decision that is easy to reverse, some of these steps can be skipped. For a decision that is hard to reverse, use all of them.

1. Write the question as one sentence, and say who will decide.
2. Ask everyone to write down their view and their main reason, in two minutes, before anyone speaks.
3. Read the views out. The person who will decide reads theirs last.
4. Ask each person what they know about the question that the others might not, starting with the person closest to the work.
5. Discuss the options. Ask the people who disagree to explain why, and keep the discussion about the options.
6. Before deciding, imagine that the decision has failed, and list the reasons.
7. Decide, and write the record on the day, with the reason, what else was considered, and a review date.
8. Act on the decision, including the people who argued against it. Reopen it only if something new comes to light.

Headwater is built for companies that want to run on one plan with a weekly rhythm, and a team makes most of its decisions while it works through the issues on its list. In Headwater, the outcome of an issue can be recorded as a decision, and standing rules are kept in a rulebook that everyone can search.