six steps to distributed decision making

Here is where you start if you want autonomous distributed teams

I was introduced to David Marquet’s ‘Intent- Based Leadership’ in around 2014. If you haven’t read the book ‘Turn the Ship Around!’ or seen the short video ’Greatness’, I thoroughly recommend it. David teaches leaders how to distribute authority throughout their team to deliver extraordinary results using military leadership principles.

I think it was approximately 2015 or 2016 when my business partner Paul shared Stephen Bungay’s wonderful ‘The Art of Action’ and we decided that we’d give leadership based on mission command a shot in our consultancy.

Our business had been growing at a fairly hectic pace. For six years running, we were recognised as one of Australia’s fastest growing businesses in the ‘Australian Financial Review’s Fast 100’ and scaling at speed brings with it lots of challenges and learnings. Distributing decision making seemed like a great way to defend against bureaucracy and group-think and it also provided an opportunity for our people to grow.

We made Stephen’s book compulsory reading for our next strategy day and started the process of leading with intent. Frankly, this wasn’t as smooth as it sounded when David presented it at Agile Australia, but the net result was really positive. I’ve worked with a number of organisations on the same challenge since then and what I’ve attempted to capture here is some guidance from those experiences.

  1. Acknowledge the problem

The first step of implementing distributed authority is to acknowledge that there is a problem. Without a clear driver for change you’re unlikely to create any. 

Leaders often can’t see the costs of centralised decision making whereas they are viscerally obvious to frustrated team members. Mistakes due to bad decisions are highly visible whereas the cost of indecision is often hidden:

  • Wait time added to processes;
  • Wasted effort due to late decisions or changed decisions;
  • Reduced team motivation.

Often organisations will prioritise the availability of executives, and will have monthly or quarterly decision forums, without querying the cost of the decisions they are delaying.

There are a number of ways that you can surface the problem to leaders. In remote:af’s operating model design approach we set autonomous decision making as an explicit design principle. We then encourage the team to discuss and acknowledge organisational constraints that prevent us from designing to that principle.

  1. Understand the culture

One of the biggest challenges to implementing distributed decision making is culture. People who are highly driven by autonomy tend to leave bureaucratic organisations with centralised decision making, while people who don’t mind taking orders tend to stick around. With enough time, the only people left are order givers and order takers: there’s nobody willing to challenge the status quo and people will respect management line / hierarchy over common sense.

As Marquet says: “People who are treated as followers treat others as followers when it’s their turn to lead.”

I once worked with an organisation that had deliberately selected people who demonstrated ‘compliant’ qualities in psychometric testing as part of recruitment. For a time it made managing teams easier, because they were less vocal about issues, but eventually it led to serious customer challenges as unchallenged problems surfaced in a very visible fashion. I found that teams in that environment required significantly more time and support before they would even start to take responsibility for decisions.

I’ve also worked with many organisations and teams where a dominant or micro-managing leader has left behind a team that simply won’t make decisions out of fear of the consequences. It takes sensitivity and care to bring those teams out of their shells, they need to see autonomy role modelled and more importantly they need to see that people making decisions will no longer be punished for it.

The local culture will impact how ambitious you can be with your decentralised decision making goals, and also the pace at which you can expect change. In bureaucratic environments or in teams that have been working for a dominant leader expect a much longer journey requiring significant support.

  1. You’re not the military

I think a key takeaway from my experience was that while you can learn lessons from the military, you also have to appreciate the nuances.

On David Marquet’s submarine

In your civilian organisation

Civilian Org: You operate in a dynamic competitive landscape and threat landscape

Submarine: His people work for the US Navy, and hence for him.

Civilian Org: People work for the organisation, for consultancies providing services to the organisation, or for themselves as independent contractors

Submarine: Leaving to the competition is considered to be treason

Civilian Org: People leave and join the competition all the time

Submarine: People were taught leadership doctrine as part of onboarding.

Civilian Org: People often come straight off the street with cursory onboarding programs.

Submarine: His people have been through training and certification programs that allow him to trust that everyone knows their role.

Civilian Org: There is huge variance in the quality of training and certification programs are often based on attendance rather than competence.

Submarine: His people can be court martialed and punished with imprisonment

Civilian Org: People need to be performance managed within a complex policy environment

Civilian organisations are different to military organisations, and you’ll need to be prepared to adjust to your context. People need care and support during the change process and you’ll also need to make sure that new people on-boarded into your organisation understand what you’re trying to do.

  1. Be explicit

To get the best out of distributed authority you need to be really explicit about decision making boundaries.

In remote:af’s governance, leadership and team launch patterns we encourage people to spend time on thinking about the decisions that they need to make, and how they need to be made.

As per Don Reinhertsen in ‘Principles of Product Development Flow’, leaders should be encouraged to control the economic logic behind decisions, not the entire decisions themselves.

We encourage leaders to push decisions down to teams and individuals, but we also encourage them to think about frameworks that align decision logic to central considerations and information flows that are required for good decisions to be made.

Good examples of decision support frameworks are escalation policies for risks and issues, value limits on contracts and expenses, frameworks for supporting prioritisation decisions, etc.

You also need to be explicit about how decisions are to be made. In our leadership launch pattern we ask leaders to think about some of the key decisions that need to be made and to be explicit about the decision making approach that they’ll use for them. Below is an example of the remote:af team's older decision design;

We also love delegation poker from Management 3.0.

  1. Tolerate mistakes

If I had an instagram follower for every dumb decision I made in the early years of starting the consultancy and looked half decent in a bikini I’d be an influencer by now.

Starting a business is a crazy learning journey. In the early days before we could afford support I needed to pretend to be a business manager, marketing manager, sales manager, people leader, HR specialist, commercial lawyer, accountant, bookkeeper, finance manager, risk and compliance manager, technology manager and IT security specialist. When I look back at some of the decisions I made in those early days I often cringe at how naïve I was.

I’m also really good at forgetting the journey. I have a tendency to unreasonably expect that people who haven’t got all the scars that I have are going to be able to make decisions just as well.

The plain fact is that nobody is born a good decision maker. You become a good decision maker by making decisions and learning from the bad ones. You need to encourage people to make decisions, but also critically tolerate the inevitable lessons.

The biggest challenge I’ve found is thinking about how you can give people autonomy but also minimise the risks and make sure that the blast radius for errors is small.

  1. Reflect regularly

The quickest way to help people grow is through retrospection. Regular reflection on how decisions are being made and how the group can get better at it is key to success.

The team should regularly reflect on what decisions were made, who made them, how they were made, and the results of the decision. You should challenge them to build better approaches for decision making and to create time so that they can surface management information for decision support.

Often you’ll find that the biggest gap is context. People at senior levels of an organisation or with strong networks / lots of experience will have context that team members who are closer to the work are missing. Prioritising the flow of information and valuing transparency over secrecy will help your team to make good decisions more often.

You also need to be very careful that you don’t just snap back into centralised decision making when failure occurs. In a complex environment some failure modes are unavoidable, even with the best information we can still make mistakes. Focus on exploring the decision making process and how the decision was reached, not the outcome.

Despite all the research on the power of autonomy as a motivator, implementing distributed decision making is really, really hard. You can’t tell people that they get to be autonomous now and expect things to happen overnight. The process will likely take years, not months to implement, and you’ll likely hit a few speed bumps on the way.

People need time to get comfortable with the idea of making decisions, and they need to learn how to make good decisions, often with some missteps on the way. You’ll need to fight your natural instincts to take over when things are getting messy and balance the need for expedience / direction against the future value of a distributed decision making culture.

That said, the power of a distributed decision making culture is truly amazing. The Ukrainian war has shown the power of autonomous teams over centralised control, even in the face of overwhelming asymmetry in firepower. This tale by Dmytro Yarmak is really inspiring:

If you’re the one person making decisions, you’re at best a bottleneck, at worst a roadblock. You’re limiting your team to the pace that you can make decisions and you’re also limiting their potential. As Marquet says: “People who are treated as followers have the expectations of followers and act like followers. As followers, they have limited decision-making authority and little incentive to give the utmost of their intellect, energy, and passion.”

I thoroughly recommend that you give it a try, and hopefully my learnings plus some of the tools in remote:af can set you up for success.