Team of teams by General Stanley McChrystal – leading teams to work effectively together

February 23, 2019 Keith GuthrieLeave a comment

The book has a few essential ideas which are worth while but it takes quite a lot of background to get to them. Below are my key takeaways.

The context for McChrystal was trying to get specialist units in very different parts of the military, who each worked incredibly effectively in their specific area, to form a cohesive whole to adapt to rapidly changing situations in Iraq Eg. Getting Army Rangers, working with Navy Seals, with airforce, with the NSA and with the CIA. Each branch tended to create its own cohesion creating tightly knit teams but resulting in territorial behaviour and collectively failing to complete their missions.

The basic message is that in the the 20th century progress was made through industrial efficiency with perfectly planned production processes around complicated problems but with perfectly predictable outcomes that engineer can solve. In these structures vertical command and control management worked effectively with each team operating efficiently but limited need for close interaction between teams.

In the 21st century, in modern organisations, we face problems of complexity, networked systems where small perturbations can lead to unpredictable outcomes. To operate in complex problems we need to be able to function with much greater flexibility and adaptability, connecting disparate information, and making quick decisions with dynamic and changing plans. To do this requires a very different management style for our organisations.

His prescription is three fold

1. A need for complete information sharing across all teams to create contextual awareness across teams and a “shared consciousness”

2. A need for strong trust between teams with multiple connection points, to create a team-of-teams type operating mentality

3. The need for the right type of leadership creating an environment of “empower execution” , where the leader is focused on culture and prioritisation to drive the team dynamic

Taking each of those in turn

1. The need for information sharing across teams

“In a domain characterised by interdependencies, what ever efficiency is gained through silos is outweighed by the costs of “interface failures””
Emergent intelligence between teams can be achieved in larger organisations willing to commit to the disciplined deliberate sharing of information
Fuse generalised awareness, “shared consciousness” with specialised expertise
To achieve this there needs to be common purpose.
Emphasis on group success to spur trust, cooperation and common purpose.
To do this they created a daily common forum, using technology, like a global video conference where everyone called in from all of the world. Anyone from any team could participate, everyone had access to all the information with almost total transparency.
The success of this depended on it being quality useful information rather than beautifully dressed up rehearsed message sending.
The update piece from a team outlining their facts would be short eg 60 seconds, then there would be 2 to 3 minutes of open questioning and conversation from leadership. Key is active listening and real exploration, potentially followed by some perspective or framing from the senior team, but then letting the individual team decide how they would proceed. Allowed all teams to see problems being solved real time and the perspectives of senior leadership team. This gave teams confidence and permission to solve their own problems, rather than having to have decisions come from the top.
Think about the physical space and the way you go about doing this information sharing carefully, but also about your decision making procedures.
Information was shared widely without constraint. As information was shared, it encouraged others to share.

2. Creating real trust and collaboration between teams

The key issue is that good collaboration between teams requires sacrifice (of resources or achievement in one area) on behalf of each team for the greater good. This happens any time there are scare resources, eg engineering resources working for something good for one team or something else for another team.
In Game theory the prisoners dilemma type problem illustrates a situation where the individually dominant strategy (betrayal, taking the resource to further your own ends) is suboptimal to the collectively dominant strategy (cooperation but sacrifice of the resource to the greater good). Even with wholistic awareness of the situation the prisoner still has to take a leap of faith in trusting the other party.
The dominant strategy in a multi round game is to start with cooperation and then to always follow what the other person did in the previous round. If they betrayed you, you betray them in the next round as punishment. If they cooperate you continue to cooperate. The punishment only lasts as long as the bad behaviour continues and stops as soon as there is cooperation. A track record of cooperation at a certain point then becomes the norm and trust builds.
Leaps of faith are only possible when there are real relationships of trust between individuals on the different teams.
To build trust they encouraged individuals from one unit to spend a secondment with another unit, to be a liaison officer with that unit. And they encouraged the teams to send their best people on these assignments. People capable of building relationships even in an initially hostile environment on another team, people with low ego. They encouraged the units “if giving up this person does not cause you pain, you are sending the wrong person”
They supplied the liaison officer with continued intelligence and information that would be useful to the unit they were in, and gave them access to the senior team so that when a liaison officer called in a favour, they could deliver value to that team.
This built a system where teams got more out of accepting these liaisons and were then willing to commit their own best people to do the same in reciprocation.
When it comes to sharing scarce resources, if teams can understand why and how their resources will make a difference somewhere else they are much more willing to make the sacrifice of giving up that resource.

Together, the strong sharing of information around a common shared purpose, and a strong bond of trust and mutual cooperation at multiple levels between teams create the ground for “shared consciousness” across teams. Hence the books title team of teams.

3. The role of leadership

So their aim is coordinated operations that exhibit an emergent adaptive intelligence, decentralised control with empowered decision making built around a shared consciousness and information. The role of leadership is to enable all of that.

The role of a leader is to build, lead and maintain a culture that is flexible and durable.
Don’t misinterpret empowerment. Simply taking off constraints can be dangerous
It should only be done if the recipients of new found authority have the necessary sense of perspective to act on it wisely.
Team leaders and members can be free to make decisions as long as they provide full visibility under the “shared consciousness” model. They have to provide sufficient clear information to leadership and other teams about what they are doing.
It’s an “eyes on – hands off” model of leadership.
The objective is “smart autonomy”, not total autonomy, because everyone is tightly linked in a shared consciousness with the same purpose.
The role of the senior leader is “empathetic crafter of culture, rather than the puppet master”. It’s a gardner creating the right environment rather than the heroic leader or chess master taking all the big decisions.
The leader should be taking fewer decisions, but should be keeping the organisation focused on clearly articulated priorities.
This leadership comes from consistently explicitly talking about what the priorities are but also demonstrating the way the team should operate, leading by example,
Less is more, focus on only a few key messages and repeat them consistently. Nothing is learned until it’s been heard multiple times, and it’s only sunk in when it’s echoed back in the words of others.
Your strongest form of communication is your own behaviour.
Eg. Information sharing sessions never cancelled and attendance mandatory
The rules for any meeting are established more by precedent and demonstrated behaviour than by written guidance.
Be clear on your central role as a leader. To lead, to inspire, to understand, to guide, to prioritise
Watch the small behaviours. If you look bored, if you are unprepared you send a message. Interest and enthusiasm are your most powerful behaviours. Prepare, ask questions, demonstrate you have really listened, compliment work publicly, suggest improvement privately, and say thank you often.

Get the balance of reporting information vs active interaction right for the meeting. Get the right level of candour through the way you interact.

Think out loud, summarise what you have heard, how you process the information, outline your thoughts on how we might proceed, ask the team members what would be an appropriate response and what they plan to do. Ask for opinions and advice. Admit when you don’t know. Empower them to take the decisions.
Develop the art of asking good questions. Questions that help people arrive at the answers and see errors for themselves.

Be careful of overcommitment on your schedule, when you cancel people get disappointed, work done preparing for meeting with you is wasted.

Avoid a reductionist approach, no matter how tempting micromanaging a situation may be. The leaders first responsibility is to to the whole, to the big picture, no matter how good they may be at the particular situation.

Business · Investment · Psychology

Human behavioural motivations and insights with Rory Sutherland

January 6, 2018January 6, 2018 Keith GuthrieLeave a comment

Here is a facinating conversation between Shane Parish of Farnam Street and Rory Sutherland, the Vice Chairman of Ogilvy & Mather Group, one of the largest advertising companies in the world.

Rory started the behavioral insights team at OM and spends his days applying behavioral economics and evolutionary psychology to solve problems for their clients.

The conversation is broad ranging and long (Rory certainly likes the sound of his own voice) but it is littered with hundreds of nuggets of mental models and insights that are applicable to both running a business, investing and understanding trends in the world today including political trends and technological trends like Artificial Intelligence. So well worth a listen.

A few of my takeaways (my own thoughts in italics)

“The problem with (traditional) economics isn’t only that it’s wrong, it’s that it is very creatively limiting, it posits a very one dimensional view of human nature, entirely driven by monetary utility.”

Different consumers respond differently to different distribution channels

Did you know that advertisers have exploited double-blind testing of what works in adverts since the 1800’s, often running different printing presses with different variants of adverts to test which were more successful! The medical world only caught onto this is the mid 1900’s.

In an example he gives an advert in the pre internet era is run in three different forms: an option of responding by mail only, an option responding by telephone only, or an option where both channels were available. In their example the first mail option had a 2.5% response rate, the telephone advert had a 3.5% response rate. The interesting thing is that the combined option had a response rate of around 5.8%, ie. the two different channels for the same thing accessed almost completely different non overlapping groups of consumers. So how you sell something is as important as what you sell.

The value of unpredictability, emotion and irrationality

“It’s impossible for anything completely rational to successfully evolve, because a byproduct of being optimally rational is being completely predictable and if you were completely predictable you would be dead.” In evolutionary terms other aninals would take advantage of you due to your predictability. So there is a danger in looking at everything as an optimisation problem. Our psychology has evolved to be a bit random and unpredictable.

Examples:

People will take advantage of driverless cars because they are programmed to be safe, pedestrians will walk in front of them or mess around with them to have some fun. Driverless cars will not only have to cope with the normal unpredictability of human drivers but with also how humans will evolve their driving because of how they know driverless cars will behave!

Anger is a useful emotion for humans because the potential to provoke it prevents someone from taking advantage you. They know at some point they will get a reaction and that basically keeps most of us well behaved.

A connected though from me is: Will general artificial intelligence have to develop the same evolutionary mechanisms to operate in our world or survive in a world with other GAIs? Like genetic algorithms will it have to have random mutations or like humans have to evolve emotion to create unpredictability to be able to survive and evolve? What are the implications of that?

Some societies have a cultural tilt towards being tolerant of ambiguity, and have more of an attitude to “give the benefit of the doubt” to someone else compared to those who are more rules driven “this is my right so I will assume I can do this and not make for any allowance for others” – the former (Sweden, Ireland, Netherlands) have lower road accident rates than the latter (Germany, the US). People who are less reliant on rules operate more carefully to avoid costly mistakes.

Contextual biases and value signalling

He explores how the value we place on either a good or service is incredibly contextual. For example we will have a bias towards paying more for the same service or good from one provider or another depending on our sense of the brand of the provider. eg we will pay more for a beer from a boutique hotel than for the same beer sold by a pop-up stand even if we take away all of the “other benefits” like the location and atmosphere. In a financial context think of this as a “boutique hedge fund” versus a lowley long only manager with the same alpha! How does this effect how you market your business?

Other insights into human behaviour from Robert Trivers: his work on Reciprocal altruism (another behavioral trait, that we feel obliged to reciprocate towards someone who is altruistic to us) and work on self deception: Self deception is evolutionarily advantageous, as the best way to deceive others is to believe your own deceits.

Our desire for artificial certainty justified by rational models

Beauracrats, business men (and many investors) love a formula because of the artificial sense of certainty that it creates for then when making a decision. That prevents them from having to exercise judgement for which they might be blamed. Avoidance of blame is a key driver in many businesses.

And that creates herding behaviours because of the asymmetry of reward to getting something right (often a small bonus) versus consequence of getting something wrong (getting fired). The “rational boring norm” is the safe place for most businesses and most peopke working for a big business. Most businesses and people are focused on avoiding bad outcomes rather than maximising the opportunity.

This creates bogus rationality where we try to use simple, understandable algorithns to prove that the decision was logical rather than take a risk of complex judgement demanded in understanding a more complex system: like the financial markets. Does this have potential consequences for the current trend towards “smart beta” algorithms in investing?

This leads on to the role of imagination vs reasoning in decision making. It usually takes takes imagination to formulate an insight or hypothesis and then logic to prove whether it is true. The role of imagination is often downplayed because afterwards we present the thinking as if the linear sequence of the logic led us to the conclusion. Science and insights are very seldom arrived the linear way.

He highlights three different forms of reasoning

Forward reasoning: what will happen next because I know of what has happened so far, based on our heuristics of how the world works. We then can test out model or heuristic by seeing if the outcome matched our expectation. If not we need imagination to hypothesise how to change our models or heuristics.

Reverse reasoning: like a detective, I observe an outcome and then I hypothesise what preconditions could lead to it and like a detective I sift through possibilities. This requires a lot of imagination to hypothesise what could lead to the outcome.

Post-rational reasoning: what we usually do when we come to a conclusion: our emotional limbic system reacts very quickly resulting in a conclusion or action using its heuristic systems in a situation, afterwards we justify the outcome by coming up with a narrative or logic that justifies the decision (fitting the facts to what we want the answer to be). We are post-rational creatures rather than rational creatures. Most of our reasoning happens after we feel an emotion, to justify why we feel the emotion.

The psychology of choices and value.

On a menu we can use the price of different items to signal worth or quality. The menu context pushes people towards selecting certain options, for example seldom choosing the most expensive or cheapest option on the menu, or the more expensive options being indicative of higher quality.

Game theory and the difference between one-off games vs repeat games.

In a repeat game you don’t just optimise your outcome for the next move but for many moves into the future. For example in selling to someone, you don’t just want this sale but future sales to the same customer, so you had better make sure that this experience is a good one for the customer even if that does not maximise your short term profit. In long term games you have to be tustworthy and reputation is incredibly important. Consumers make many choices based on their intuitive understanding of service providers trustworthiness and alignment. The bigger the upfront cost of committing to a decision that takes a long time to see results the more the long term alignment will matter.

There are many practical applications in every day business. For example the focus by many large businesses incentivised by stock market behaviour towards short term earnings results is a bias toward “one time games” to make themselves rich. Think of the contrast between hedge funds with annual performance fees, private equity with 5 or 7 year incentivisation and Amazon with a very long horizon strategy for growing the business through reinvestment.

The value of experimentation.

In business you don’t have to be right all the time, you can be experimental and see what works and what doesnt (provided the cost to testing is low). It is important to test some counterintuitive or imaginative ideas because as its much more valuable when they pay off as your competitors won’t or are unlikely to have tested them.

In the case of the financial markets the payoff of a contrarian bet may be large as it is counter to expectations with little downside if the crowd is already pricing in the consensus scenario.

So have a listen here

https://www.farnamstreetblog.com/2017/06/rory-sutherland-podcast/

Business Culture · Investment · Learning · Psychology

Improving decision making in committees

September 2, 2017September 7, 2017 Keith GuthrieLeave a comment

A review of decision making literature from UCL on what makes for good decisions in committees is particularly relevant for investment committees:

1. Diversity of participants (functional diversity) with different independent expertise or knowledge makes optimal decision making more likely (finding global rather than local maxima more likely with different starting points for the mathematical optimisation geeks)

2. Slower decision making is more accurate. (And conversely sometimes you need to trade speed for accuracy).

3. Don’t vote, prefer discussion. Getting to a consensus answer or leave it to an expert final choice, if you do vote it needs to expertise weighted (ie some votes count more than others) based on objective expertise measurement, not one man one vote.

4. Over confidence bias: The worse you are at something the more delusional you are that you know the right answer! More expertise often results in less certainty from expert individuals: the problem is they know how much they don’t know and they can take other less expert people’s opinion into account too much. Go with the experts opinion after having the discussion and as the expert balance your opinion carefully with the input from others.

Podcast: The Naked Scientist

Episode 22/08/15

Dan Bang University College London 5:40 mins up to 10:40 mins

https://itunes.apple.com/gb/podcast/the-naked-scientists-podcast/id74171648?mt=2&i=1000391340611