Organisational Decision-Making

When talking about decision-making biases, (ref. last week’s article), it is one thing to recognize them and then another for an individual to adapt to mitigate their effects; however what about at the organisational level? A recent McKinsey study of more than 1,000 major business investments showed that “when organisations worked at reducing the effect of bias in their decision-making processes, they achieved returns up to seven percentage points higher” (ref. “The Case for Behavioral Strategy,” McKinsey Quarterly, March 2010). How to ensure better decision-making at the organizational level is the main focus of Daniel Kahneman (the Nobel prize-winning economist) along with Dan Lovallo, and Olivier Sibony (two McKinsey consultants) in their June 2011 HBR article “Before You Make That Big Decision...”

Here’s what to do before you make that big decision, followed by further considerations (“et alors”)

Organisational Decsion-Making

At the organizational level, a review process should be put in place so that executives can assure the quality of decisions by thinking through not just the content of the proposals but the biases that may have distorted the reasoning of the people who created them. That review will comprise of the following 12 questions which check for:

Self-interested bisases

·         Is there any reason to suspect errors motivated by self-interest?

·         Review the proposal with extra care, especially for over-optimism

The “Affect Heuristic”

·         Has the team fallen in love with its proposal?

·         Rigorously apply all the organisation’s quality controls

Groupthink

·         Were there dissenting opinions within the team? Were they adequately explored?

·         Solicit dissenting views, discreetly if necessary

Saliency Bias

·         Could the diagnosis be overly influenced by an analogy to memorable success?

·         Ask for more analogies, and rigorously analyse their similarity to the current situation

Confirmation Bias

·         Are credible alternatives included along with the recommendation?

·         Request additional options

Availability Bias

·         If you had to make this decision again in a year’s time, what information would you want, and can you get more of it now?

·         Use checklists of the data needed for each kind of decision

Anchoring

·         Do you know where the numbers came from? Can there be unsubstantiated numbers, extrapolations from history and/or a motivation to use a certain anchor?

·         “Reanchor” with figures generated by other models or benchmarks and request new analysis

“Halo Effect”

·         Is the team assuming that a person, organisation, or approach that is successful in one area will be just as successful in another?

·         Eliminate false inferences, and ask the team to seek additional comparable examples

Sunk Cost Fallacy

·         Are the recommenders overly attached to a history of past decisions?

·         Consider the issue as if you were a new CEO

Overconfidence

·         Is the base case overly optimistic (eg. “planning fallacy” and “competitor neglect”)

·         Have the team build a case taking an outside view. Use scenarios

Disaster Neglect

·         Is the worst case bad enough?

·         Have the team conduct a pre-mortem. Imagine the worst has happened and develop a story about the causes

Loss Aversion

·         Is the recommending team overly cautious?

·         Realign incentives to share responsibility for the risk or to remove the risk.

Note: the goal is not to create bureaucratic procedures; rather it is to stimulate discussion and debate. For sundry reasons, the review is best performed by “ad hoc” critique teams assembled from a diverse pool of talent within the organisation.

Et alors

To accomplish any of the above, as the authors state, “organizations must tolerate and even encourage disagreements”: hence the need for the persons conducting the review to be independent of the executives responsible for the decision. It is not only independence which is required but diversity. If everyone thinks, acts and behaves the same way then the review process is going to be limited in effectiveness. To implement all of the above checks is incredibly challenging but also very rewarding when done “properly”: organizations that have achieved such discipline have not only yielded better returns for shareholders but also consider their decision-making process to be a competitive advantage in itself!

Minimising decision-making biases

Managers like to think that the decisions they make are objective and rational and have been executed taking into consideration all the pertinent facts. In reality, many other forces come into play when making a decision such that most decisions are usually subjective, irrational and have been made with only limited information. As Wolf explains in his HBR blog post of September 2012, “How to minimise your biases when making decisions”, it is mainly inherent biases which cause decisions to be sub-optimal. Accordingly, in order to optimise decisions, biases have to be recognised and then minimised.
 

Here’s a summary of decision-making biases and how to overcome them followed by further considerations (et alors).

Minimising decision-making biases

Drawing on studies from Rosenzweig, Kahneman and Schoemaker, there are six principle decision-making biases noted as follows:
 

Anchoring

If expecting the same outcome as before, contradictory data tends to be ignored or omitted. This becomes more acute with numbers which are difficult to “adjust away” from.

·         To overcome anchoring, “seek diverse outside opinion to counter overconfidence”.

Framing

How a situation is presented affects the decision. In addition and generally, the pain of losing is perceived to be more powerful than the pleasure of winning.

·         To overcome framing, inverse the perspective from positive to negative or vice versa.

Availability Heuristic

Vivid and easily imaginable events (even if uncommon) along with recent events are weighted disproportionately higher than unimaginable or past events.

·         To overcome availability heuristic, “search relentlessly for potentially relevant or disconfirming evidence”.

Confirmation Bias

Initial decisions become self-fulfilling prophecies. Data is collected after the event to justify the decision (similar to anchoring).

·         To overcome confirmation bias, accept a “chief contrarian” as part of the team.

Commitment Escalation

It is difficult to accept “failure” and start again from scratch; instead, previous commitments tend to influence current decisions.

·         To overcome commitment escalation, ignore the old problem by clearly redefining the new problem.

Hindsight Bias

Once something is known, it is difficult to remember when it was not known. This can make it more difficult to learn from failures (as it is now “obvious”).

·         To overcome hindsight bias, reward due process (and in particular “lessons learnt”) rather than penalising failure.

Et alors

How many organisations do not penalise failure in one way or another? The above decision-making biases are compiled at the individual level but despite the best efforts of any individual, decision-making biases can easily become “institutionalised” in an organisation. All of the biases noted above can reinforce organisational culture: for example, it would be a very brave individual who could go to the board of Kodak in the early 1990s and tell them that photographic film was doomed and they should start preparing for an exit (yet look what happened)! On a collective basis, a lot of the biases could be compiled into one key tenet, namely “groupthink”. On the other side, the one concept which seems to appear as the solution on a collective level is “diversity”! To optimise the decision-making process and overcome decision-making biases, make sure the team making the decision is diverse! The process might seem to be more "painful" but the end-result should be a higher quality decision!

Why Leaders Need to Think Slowly

One of the best-selling business books of 2012 was “Thinking, Fast and Slow” by Daniel Kahneman, the winner of the 2002 Nobel Memorial Prize in Economic Sciences.  The author asserts that there are two cognitive systems: “system 1” which does not take much effort and makes quick judgements based on familiar patterns; and “system 2” which takes more effort requiring intense focus and operates methodically. The first is fast and the latter is slow with both systems working continually but not necessarily fluidly: people prefer to make simple stories out of complex reality. Seeking causes in random events and giving too much weight to their experience, there are some “traps” for the unaware, particularly leaders.

Here’s a summary of the traps followed by further considerations (“et alors”).

Why Leaders Need to Think Slowly

When leading a team or making decisions, the prevalence of the “system 1” can cloud judgement in the following ways:

1.       Halo Effect

System 1 prefers links between discrete facts: if only one of the facts is available but it has been seen before associated with another fact, the link is made automatically. When a positive fact relates to a person, this can result in the “halo effect” and this can be a trap when it comes to talent selection.

2.       Anchoring

System 1 is responsible for unconsciously tying your thinking on a topic to information you have recently encountered (even if the two have nothing to do with each other). This can be a trap when it comes to (what is believed to be) “rational” decision-making.

3.       Hindsight Bias

System 1 simplifies the “narrative” such that success is ascribed to virtue and skill whereas failure is attributed to bad luck. By being overly optimistic and overvaluing your talents, the “hindsight bias” can be a trap when making investment decisions.

4.       False Expertise

System 1 promotes expertise in relatively simple environments; however where challenges vary, luck influences success and gaps exist between action and feedback, the trap is to rely on “expertise” which may give the wrong counsel (“quick answers to difficult questions”).

5.       Risk Bias

System 1 promotes “loss aversion” (rather than gain seeking) and individuals suffer from the “endowment effect” where ownership results in an overestimation of value. This can be a trap when it comes to divestment decisions.

System 2 promotes slow reflection and intense focus: the leaders can then help the organisation to “operate with more methodical rationality than can the separate individuals within it”. Achieving slower reflections can help leaders make better judgements and achieve a rationality that is otherwise only “fictional” when “system 1” is prevalent.

Et alors?

Most leaders will wonder when they will ever find time to think slowly! On the other hand, does any leader have the time to lead their company, department or team in the wrong direction due to having made a decision too quickly under the “influence” of “system 1”? In today’s instant-everything fast-paced environment it seems that this tension between careful reflection and getting things done is non-existent: getting things done always prevails! Whilst not slowing down permanently, most leaders might benefit from being able to at least recognise the “signs that you are in a cognitive minefield, slow down and ask for reinforcement from system 2”!

One definition of a group’s culture is the group’s collection of behavioural short-cuts. The group (be it a team, a company or a nation) has devised organisational and inter-personal “norms” that are essentially adhering to a collective “system 1”. So what happens when an outsider joins that group? Can all the members of that group be asked to convert to “system 2” to stop and reflect on the new member’s behaviours so as to welcome and possibly adapt? Whilst not everyone could, perhaps leaders should, especially when decisions relating to people might be clouded by a cultural “halo effect” bias. Diverse organisations need to be particularly aware and leaders should ensure that talent selection is performed in an objective and transparent manner!

Groupthink and the Need for Diversity

Groupthink is a psychological phenomenon that occurs when the desire for consensus in a decision making group overrides the members’ motivation to realistically appraise alternatives. From the smallest group up to large organisations, groupthink can reduce creativity, inhibit innovation and result in sub-optimal performance, possibly even to the point of self-destruction (cf. Enron etc.) Why does this happen and how can it be avoided?

In an article by Powell, “Stop Groupthink Damaging Your Business”, Finance and Management, ICAEW, November 2011, the author draws on the original work by Janis and cites three “antecedent” conditions to groupthink: 1/ High group cohesiveness; 2/ Structural faults (such as insulation, lack of impartial leadership and sociological and/or ideological homogeneity); and 3/ Situational context (such as highly stressful external threats, recent failures and difficult decision-making processes).

Here are the symptoms and possible solutions followed by further implications (“et alors”).

Groupthink and the Need for Diversity

According to Janis, there are eight symptoms of groupthink:

1. Illusions of invulnerability creating excessive optimism.
2. Unquestioning belief in the morality of the group.
3. Rationalising warnings that might challenge the group’s assumptions.
4. Stereotyping those who are opposed to the group as weak, spiteful, ignorant…
5. Self-censorship of ideas that deviate from the group consensus.
6. Illusions of unanimity where silence is viewed as agreement.
7. Direct pressure to conform placed on any “disloyal” members of the group.
8. Self-appointed members, “mind guards”, shield the group from dissenting information.

Taking these into account and reviewing current research, Powell suggests various solutions:

1. Encourage diversity
2. Avoid directive leadership
3. Embrace the culture of enquiry
4. Don’t chase consensus; rather explore disagreements
5. Do not isolate yourself from the ideas of others
6. Discuss issues in the spirit of openness
7. Avoid converging on an answer too quickly
8. Learn to deal with ambiguity

Et alors?

When seeing that sociological and/or ideological homogeneity is an antecedent for groupthink, it seems almost obvious that diversity is a solution; however is an “injection” of diversity an immediate remedy? On a small group basis: probably yes – diversify the members and groupthink might evaporate. On a large organisation basis: perhaps not – diversity might be prevalent in the organisation but it still might not be immediately included in the decision making groups.

When the decision of who can join the decision making group is subject to groupthink itself then the organisation is probably never going to benefit from increased diversity; instead the group will self-replicate and groupthink will continue.  This is evidenced in many organisations where the diversity at entry level is very high but then steadily diminishes in direct inverse relation to seniority in the organisation.

Diversity is not just about nationality and gender: it is also about beliefs, values and behaviours. An organisation can therefore have lots of visible sociological diversity but very little ideological diversity. The organisation believes it is diverse but is only recruiting into its upper ranks new members who share the same ideology as the former members despite their nationality or gender.

Another form of diversity is to be found in different leadership styles. If there is only one leadership style and that is principally directive then dissent, discussion and open review are unlikely to be encouraged; instead the risks of groupthink and diversity exclusion are both increased. Diversity is needed to remedy groupthink; but for it to have an effect in large organisations, diverse leadership is also needed.