Scaling the decision process with the business
We have all experienced situations where poor outcomes occur. Often these are not due to unexpected results from well-thought choices. Instead many poor outcomes are because organisations find decision making hard. Perhaps the right people were not involved in the decision. Or the right data was not available. Maybe pet projects and personal egos won out. Or people were simply too scared to say the right thing.
I should have made a lot more noise, criticized people more directly.
How Nokia Lost the Smartphone Battle
I should have been braver about rattling people’s cages
Startups rarely have a big problem with decision making. In a small organisation, there are less people and less customers. And that means less options to juggle. There are less diverse opinions to accommodate. People are generally the best informed decision makers for their own work. And if a second opinion is needed, the founder is just a Zoom call away. Pulling the whole organisation together to discuss a problem on a Thursday afternoon is still a viable option.
In many of the scaling organisations I have worked with the decisions simply weren’t being taken and the business drifted into a default outcome which is rarely ideal. Getting decision making right becomes one of the key factors in scaling towards reliable delivery.
Decision making is a key factor in scalability.
Scaling a decision making process.
One of the key reasons for any process is to avoid busy and stressed people making mistakes. It has been repeatedly shown that skilled individuals in stressful situations will make errors and that these can be mitigated by a defined approach (such as checklists). We identify a set of “good practices”, make these available to people and evolve them as we learn.
We are built for novelty and excitement, not for careful attention to detail. Discipline is something we have to work at.
“The Checklist Manifesto” – Atul Gawande
Here on Agile Plays, the idea is to build your own best approach, so “process” in an Agile context should never mean “fixed and inflexible”. Let’s break decision making down into a set of steps and look at how we can best apply these.
In this model, we look at the following steps to good decision making:
- Identify the decision
- Understand the impact
- Agree how we decide
- Identify alternatives
- Post-decision review
Whenever you see a successful business, someone once made a courageous decision.
Peter Drucker
Identify the decision
This sounds an obvious step. Surely a decision is a decision and everyone will recognise it?
One key place decision making often fails is in failing to even recognise that there is a need to make a decision. We have all seen cases where an organisation drifts into a poor result. No-one has decided what to do and no action has been taken. With no conscious decision, a default outcome is the result. These are rarely the best option – in general steering towards a good outcome will lead to a better result than passively accepting whatever occurs.
These “failed decisions” can often be identified in retrospectives, and you can get good guidance on your decision making process here. If the team looks at the current situation and no-one can explain the steps which led to that situation, that is a clear sign that decisions weren’t being effectively taken.
Perhaps we should step back further and agree what a decision even is. A decision consists of three distinct parts which combine to form the decision.
- Variability – at least two possible future options might occur
- Choice – a selection by the decision maker of one option over the others
- Plan – an action or set of actions to make the preferred choice more likely
Those of you familiar with Risk Management will immediately see the similarity. With both decsion making and risk, we are moving away from the reductionist view of a single defined sequence of events. Instead we have to visualise multiple possible scenarios which might occur. Our decisions are used to steer which events will occur.
Understand the impact
When we talk about decisions, most people think of a single irreversable choice between one of two paths. Typically one path leads to triumph and the other to disaster. And also typically people imagine a decsion which must be taken immediately, often with too little information for the best answer to be obvious.
You are floating down a river and come to a fork. You need to choose which path to take. Paddle left towards the smooth water or right towards the rapids? And is that a waterfall you hear in the distance?
But that’s a simple model and real decision making isn’t like that.
Firstly, you may be able to see far down the river. Few decisions are sprung on you instantly (at least if you have been paying attention). So decisions may have a very varied time horizon between now and when the decision must be made.
Secondly, you have a canoe not a raft. If you made the wrong choice, once you see round the next bend in the stream you can generally paddle back and take the other path. It loses some time, it’s annoying but possible. If the current is strong you can head to the shore, pick the canoe up and carry it to the other branch. Even more disruptive and hard work.
Most decisions are far less final than you think. Jeff Bezos distinguishes between the common, reversable decisions which he calls “two way doors”, and the really final choices which he calls “one way doors”.
There are one-way doors and two-way doors.
Jeff Bezos
It’s never quite as simple as two types of decision, although Bezos’ analogy is a useful one. In any decision, there is generally some cost to paddling back and taking the other path. Time and effort has been lost. “One way doors” are at the extreme end, but there is a scale of “two way”.
- In most decisions, such as prioritising one item over another, the cost of later changing your mind is small. Some context switching, some delay in both items through parallelism, some increase in work in progress.
- In other decisions, there is a clear barrier if you later changing your mind. A financial outlay, a promotion or reorganisation.
- And a few decisions have significant cost to reverse, or may not be possible. Contract signatures might fit into this category – deciding to sell the company is very much a one-way door.
In this stage you need to decide when it is important to make the decision and how reversible the decision will be, or for how long it will be possible to change your mind. It may seem attractive to make all decisions quickly, and many “Scientific Management” style managers pride themselves on this. However, once the decision is taken, you start to incur cost if the decision is changed.
Perhaps counterintuitively, in most cases, taking the decision later allows you to amass more evidence and avoid cost of early commitment. Wherever possible, a later, well-reasoned decision will be better than an early, “gut feel” one. But that is consciously deferring, not just drifting. Understanding the timeline is key to ensure that you do not drift into a decision by default.
It is easier to change a decision that hasn’t been made.
Mary Poppendieck
Agree how we decide
Now we know there needs to be a decision, we know when it needs to be made and how final a decision it will be. Will we be able to change our mind later (“two way door”) or will the cost of change be too great (“one way door”)?
These are the key inputs into how we make the decision. We need to balance the two key factors of quality of decision and cost of decision.
- The quality of the decision is how good an outcome we achieve – the value or benefit we gain by making the decision.
- The cost of the decision is how much we invest in making the decision. No action is free and the more we plan and consider, the higher the cost.
Most decisions will be relatively minor and low impact. They will be “two way door” decisions where we can easily row back if we see we have made a lower quality decision. For these, we want to prioritise low cost over high quality. We cannot allow the decision making process to slow the team’s ability to progress.
To maintain flow, we need to take these decisions quickly and with as few people as possible. This is where lowest viable level decision making is powerful. By delegating to a self-managing team, who are empowered to make day-to-day decisions, we can make those decisions quickly based on facts which the team will typically have to hand.
A few decisions will be close to Jeff Bezos’ “one way doors”. For these, we need to prioritise high quality over low cost. The implications of the decision are high as it will be difficult to reverse. Therefore we want to make sure our choice is right. “One way” decisions are uncommon, so we can afford to invest more effort in making sure that we get the decision right.
Here we will want to identify the key stakeholders for the decision who need to be involved to make the right call. We need to invest time in finding and sharing data. We need to invest in diversity of decision making to avoid “Groupthink”. And because the decision is irreversible, we need to do all of this in advance of our choice.
Remember however that in the complex Agile world, for many decisions the facts will not all be clear without the learning from starting down the path. This is where the key part of planning is to get the answers as early as possible after the decision. If you start with the high risk areas, you know if you need to rethink the decision – the idea known as “fail fast”.
Identify alternatives
If a decision is a choice between alternatives, it is key to identify what are the alternatives between which we are choosing. This is a step which seems frequently to be missed. There may only be two or there may be several.
Each alternative is a scenario – an alternate future if you like. And like any alternate future it will have benefits and disadvantages. For a decision to be effective everyone must have the same understanding of the outcomes. Too often people agree on option “B” with different ideas of what this option entails.
A common failing when identifying alternative outcomes is not to throroughly explore the consequences. A simple example is the introduction of a new product. If you evaluate the product solely on an RoI in isolation, it is likely to seem profitable. But if you introduce it, it will have an impact on other products and will pull teams off other, possibly more profitable alternatives.
If the choice is whether to introduce the new feature, remember that introducing a new feature into an ordered backlog necessarily pushes down the lower ranked features. The scenario is a tradeoff with winners and losers and needs to be clearly thought out. Ordered backlog is a great way to make this visible – the new A is prioritised above B and below C, avoiding any ambiguity.
The outcome here is a clear set of alternatives between which we are choosing. Associated with each is the data which will let you make an informed choice. For “two way door” decisions, that data may be limited, while for the high impact “one way doors” you will want to spend enough time to be able to reliably score the different alternatives, generally with a detailed business plan.
Choose an option
If the previous stage was done well, the decision making itself may be straightforward as the scenarios will be clearly described and scored. Remember you are not saying the new choice is good in isolation, but that all the consequences are also acceptable.
Allow enough time and psychological safety for people to openly raise concerns and discuss them. As we have seen with retrospectives, much of the detail of people’s concerns only becomes clear with enough time and security. If the decision is a “two way door” then concerns need not prevent the decision being taken, but they need to be logged as areas to look at as we proceed.
The important part of decision making is that the decision must be made in a way that people can move forward. There may be multiple opinions but the decision making group must end with a single decision which can be used going forwards.
Decisions need to be communicated. A decision which no-one knows has little value, so ensure that the teams involved are aware that the decision has been taken. Clarity is again important. What was the full scenario of the decision taken and why was it chosen? This helps people understand the context and work towards that scenario.
In addition to communication, there should be a record of decisions which have been taken. A Decision Log can be a very valuable tool. For each decision you want to record:
- When was the decision taken?
- Who was involved in decision making (for later clarification, rather than for blame)?
- What were the key facts known at the time on which the decision was based?
- What was the decision taken (an outline of the scenario which was the desired outcome)?
Decision Logs are a valuable tool to keep track of what has been decided. They are important for example in keeping a record of design decisions which have been made for a system.
Even better is also to add to the Decision Log those decisions which have not been taken. Places where it is known that a decision will be needed, but by conscious decision or lack of data the decision has been deferred. This avoids those decisions being missed in the future
Nothing is more difficult, and therefore more precious, than to be able to decide.
Napoleon Bonaparte
Post-decision review
For a “one way” door decision, a decision may be irrevocable. But this is not usually the case and most decisions can be reversed.
However, decisions cannot be reversed indefinitely and the cost of reversal will increase over time. So keeping on top of whether a decision is right is key. Reviewing decisions is a valuable activity to see whether subsequent learnings have changed the viewpoint.
This can often be done in retrospectives. Topics may be raised which relate specifically to a decision. Or specific decisions may be discussed directly at the retrospective.
Often there is value when a decision is made in setting a specific review point. “We will decide X and then review the new data in a month”.
Good Practices
As a leader, building an effective decision making process is key to scaling the organisation. Remember it is not all about you – neither impulsively making snap decisions based on “gut feel” nor insisting you make the decisions but failing to make them will benefit your team.
Your goal as an Agile leader will be to maximise smooth flow of work by lowest viable level decision making and this will need you to support your team in how decisions are made. If the wrong decisions are made, you need to remember that the team made those decisions based on the context they were given by you. Your role as a leader is to enable and empower your team to make good decisions in as effective a way as possible.
When blame inevitably arises, repeat this mantra: if a mistake happens, shame on us for making it so easy to make that mistake.
“The Lean Startup” – Eric Ries
Decision making works best with a clear process, whether delegated to your team or among your peers or managers. Ensure that there is a clear process in place similar to the one suggested to ensure that decisions are clearly made, well understood and updated as needed as the facts change.
Leave a Reply