Skip to main content

Escaping the Predictability Trap

August 24, 2018

It has been said that the definition of insanity is doing the same thing and expecting different results, and yet we engage in an unconscious fiction of predictability every day. We work in an uncertain world, and our main goal in pursuing agility is to confront the unknown, and in doing so, to master it. Pursuing predictability causes us to lay a veneer of fiction over the real world, making it conform to a plan of what we would like to believe is true rather than what really is. 

The reality, however, is that organizations, and the people in them, hate uncertainty. They find predictability comforting, even when it is an illusion. But to gain greater insight and achieve greater results, we have to strip away our false conceptions and see the world as it is. Agility without full transparency is a sham that reduces real agility to empty rituals.

Humans are wired to avoid uncertainty; psychology studies have shown that people will choose a certain but poor outcome over an uncertain but potentially better outcome.  We have to, somehow, let that go. Predictability sounds good, and it would be nice if we could predict the future, but we have to be realistic. Those things we call “plans”? They’re just guesses with a nice-sounding wrapper around them. Forecasts? Same thing. 

Demanding predictability creates a set of predictable dysfunctions

Being forced to produce predictability warps reality and causes people to spend time and energy creating a façade that meets expectations. The common scenarios span a range of behaviors:

  • Predictable plans - Comparing actuals to plans is deeply ingrained in many organizations who somehow think that the future can be predicted with accuracy, so any deviation from a plan is evidence of poor performance, and questioning a plan is viewed as “being negative”. Unfortunately, we live in an uncertain world. False certainty does us no good; it actually prevents us from making good choices and from achieving greater goals.  And punishing people because they didn’t guess correctly is a waste of time and discourages important learning.  Instead of demanding predictable plans, focus on articulating clear goals and clearly framing experiments, including how you will evaluate them, and be open to learning new things.
  • Predictable productivity - Managers love focusing on productivity and “efficiency” but frequently fail to consider the value that is being delivered. Delivering value is what is important, not how many “units of work” (like story points, which are themselves just guesses) were delivered. What is better: driving 100 miles per hour in the wrong direction, or one mile per hour in the right? Productivity is important, but tracking does not help to improve it. Instead, focus your efforts on removing waste and impediments, and clearly articulating goals. When walking a rough and uncertain path, it’s not how fast you go that matters but whether you reach your destination. To go faster, place smaller bets, run shorter experiments, and evaluate where you are more frequently; you’ll save time not having to backtrack later.
  • Predictable careers - Each of us likes to believe that we are on the path to success. The notion of a “career” is a story we tell ourselves about how what we are doing now is leading to something better. The problem is that we are not very good at anticipating the future, and we really have no idea of the kinds of opportunities we may encounter along the way. It may seem a bit scary that we really don’t know where we are headed, and that luck plays a large part in what we end up doing. The reality is that we cannot really imagine what jobs will exist in ten years, or even five, nor can we conceive that many of today’s jobs will no longer exist. So how to do we prepare ourselves for what lies ahead? By cultivating flexibility, by trying new things, and by solving hard problems and acquiring whatever skills we need to do so. In the end, adaptability and ability to learn quickly are the keys to success, not steadily marching to the beat of someone else’s drum. Just as with Scrum, we succeed personally by trying new approaches and evaluating the results, in measured experiments. 
  • Predictable agile transformation - “Transformation” is a word that I often associate with the phrase “magical thinking”: organizations seem to believe that they can predictively plan how they are going to “become agile”.  This is usually based on the misconception that agile is a process, or is rather like a tool, that can be “installed” or “rolled out” to an organization. It doesn’t work that way. Agility, or the word I think better captures the essence of what we seek, adaptability, is a cultural quality, a way of thinking and acting that deeply changes the way that people see and act in the world. It is not a specific set of practices or behaviors that can be adopted. It is a way of thinking and acting that involves continually seeking better results and better outcomes. As such, specific practices will change as conditions and skills change. We cannot plan how this is going to proceed, and as different teams have different challenges, their path toward agility will be different. There is no “magical” set of practices, roles, or processes that makes this easier.

If predictability is bad, what’s the alternative?

As manufacturers learned, “lean” manufacturing involved a lot more than installing andon cords to enable the line to be stopped; it relied on the cultural values that let any employee pull the andon cord if they see something that is wrong. In pursuit of agile cultural values, organizations will find that agility may look messy on the surface, as different teams make different decisions in pursuit of their own continuous improvement toward their own goals. What is predictable is the empirical approach that defines our path: making observations, forming experiments on how we think we can improve, and continually seeking better performance. 

When you embrace uncertainty, you open your eyes to new possibilities. You are no longer blinded by your pre-conceptions. Once open to the facts, you can see new opportunities, new solutions. 


What did you think about this post?

Comments (20)


Blake McMillan
06:19 pm August 24, 2018

Good reminder Kurt! Thanks for posting this. When organizations use methods that are used for solving simple or complicated problems to attempt and solve complex problems they are met with disappointment, frustration, and failure. We see this mirrored in the Cynefin Framework where categorization and analysis do not illuminate the solution to a problem. That is where experimentation (inspect and adapt) excels in navigating through the complexity one bit at a time.


Louis-Philippe Carignan
12:20 pm August 28, 2018

Have you seen the Need for Closure test?
http://nymag.com/scienceofu...

In my classes, if uncertainty becomes a topic, I ask the participants to take it as it takes a few minutes to fill. We then reflect on the different scores because as leaders, we need to know how we ourselves can handle uncertainty.

I would also argue that not all human being are wired to avoid uncertainty. I would say that depending on the level of consciousness, some people are better than others to handle uncertainty.


gazza8
09:26 am August 30, 2018

Thanks for the link to the online (and free, quite rare) quizz.


Sen Sei
03:07 pm October 1, 2018

Great post. What is strange though is that many so called Agile Coaches thrive on misconceptions and weasle their way into companies by offering full boxes of silver bullets. I just happen to wonder how many people are reading this blog actively.


Basia Jurak
07:57 am January 13, 2019

Thanks for this article.

Avoidance of uncertainty can also have a cultural background. You can check it here for your country https://www.hofstede-insigh...

The fight may be more difficult when you are surrounded by the people who celebrate predictability, like in Poland :)


Robbie
06:05 pm February 14, 2019

I'm a product owner and my scrum team is based in Poland. This is definitely an uphill battle. Any thoughts on how to make it easier?


skjoldburger
11:09 am February 19, 2019

I am a bit torn. I see the advantages of small increments as small experiments. On the other hand I worry about how to make these increments add up to something. How does one avoid adding features and making adjustments that ultimately do not add up to anything, like ship that is making constant rudder changes that lead it to go in circles? Is it that the experiments, though at a micro-level, provide macro-level or "big picture" insight and feedback as well? How is conflict between the two levels handled? In other words, if the clients love the incremental features but also love an overall direction that is different from what the little features entail, what should one do? Any clues would be appreciated.


Robin
06:33 pm February 20, 2019

So, I started with this huuuge reply, only to discover that I can bring it back to this.

1. The Product Owner prioritizes everything that needs to happen based on business value. If you do this properly and if circumstances do not change (see main article how predictability is a trap), this means that you will have little feature-waste. How to do this is a topic on it's own, but for instance it could be a formula taking into account the factors described here: https://medium.com/the-libe... .

2. In Scrum we regularly inspect and adapt both the product and the process. As long as we monitor our progress towards the goal, huge deviations will be rare. Remember however: there a valid reasons for making deviations and with the small increments approach at least you won't put too much effort in not-needed work.

3. Agile allows for multiple levels of planning (if needed!) to make sure you also take the overall (product and company) direction into account. We just keep it small and light on a 'need to have now'-basis. You can share this with customers to show them the overall direction. Look up 'Planning Onion' on google or read the rest of the explanation below the lines. Note: i purposely said 'Agile' and not 'Scrum', as Scrum does not prescribe this as mandatory.

Good luck with your endeavors and safe voyage!

------

Original (extensive) answer:

I understand your dilemma and would like to offer you some of my insights.

First of all: Agile has multiple levels of planning. Next to your sprint planning (iterations) you will hopefully have a product vision that has been co-written or at least verified with your stakeholders, a limited roadmap where your product will be going in the semi-long term and a release planning painting the bigger picture for the next couple of iterations. We call this the Planning Onion. The point with Scrum/Agile is more that we accept that we cannot predict the future. Therefore we make the information on these planning levels as condensed and as little time-consuming as possible, while still attaining the value that there is in these items. We add information to these items when we reach the threshold where we will be actually working on them soon, so that they have been thought out enough to work on them.

Second: we prioritize in such a way that we always start working on what has the most value at that moment. Determining the value of features is a whole challenge in itself, but you should try to back this up with actual data. What is the current most requested feature among your customers? If we build this, then surely the chance of this adding up to nothing will be slim.

Third: to keep with your ship analogy, if the ship is going in circles, then the captain and ships officers are not doing their jobs. All crew mates should have been informed on the intended destination, so that all aboard could have warned if they felt like the ship was drifting too much from this goal. Both the Sprint Demo (for all the passengers that want to get somewhere) and the Sprint Retrospective (for the crew mates) can be used to inspect the progress towards the destination and adapt the route.

Much like the point made in the article above, there are however times that we purposely need to reposition the destination. Maybe to go around a storm or because we have learned that our original destination has been closed off to foreign ships. In business terms: because maybe a competitor throws a curve ball which means we need to realign our purposes or because your government enforces new laws that you need to incorporate in your product asap. It might make your route longer and feel like going in circles, but your problem would be even bigger if you decide to keep to your original route to what now is an unreachable destination.

Does that make Scrum the silver bullet? By no means. I've had cases where we rebuild features either because the product owner after gathering customer feedback wanted things to work another way or where due to the incremental nature we ultimately chose for a different technical solution. However the same would probably have happened if we had designed more upfront. In this way at least we were able to react quickly and due to the incremental nature with relatively little waste of time/budget.


HK
06:46 am March 21, 2019

Great article.
Very true that Agile is more of an attitude to react rather than plan the predicted.

Harshal;


Darren Smith
11:20 am June 3, 2019

I find myself agreeing with your sentiments and thinking about how we can use tools such as probabilistic forecasting to give us some level of certainty (or confidence, if you will) in delivering within a certain time frame. I agree that it is unlikely that you can forecast a specific date that you can deliver something with a high degree of certainty but surely it is possible to provide some valuable forecast of a date range and probability? Some things in life are predictable. For example, I can say for a roll of a 6 sided die, you will never score a 0 or 7 and above. Of that I am certain. However there is a 1 in 6 probability of rolling a 1, 2,3,4,5 or 6.

I believe we need to get comfortable with probabilities and monte carlo simulations much like our friends in the financial industry have. Then we will see realistic predictions based on system performance data. As long as we all remember that a prediction does not necessarily come true.


Masoud Bagheri
06:30 am April 7, 2020

It's all about complex word! Prediction doesn't make sense in the complex problems. I recommend you to read the Cynefin Framework.


Shaukat Shah
05:04 pm May 8, 2020

Escaping the Predictability Trap (Kurt Bittner)
Summary
•Challenging plans shall be encouraged instead of predictable average plans.
•Concentration shall be on goals and methods of measuring the performance not the plan.
•Foster courage, experiment and learning rather than acceptable performance.
•Set short goals and evaluate more frequently.
•Trying new things and experiments are beneficial for the personal development of the developers and
their survival in future.
•Change of processes and tools is easy rather than change of thinking, attitude and culture which are
more important in Agile.
•Alternate to predictability is observations, experimenting, and continuous improvement.
•Risk brings challenges, opportunities, achievement, learning and growth.


Shaukat Shah
05:06 pm May 8, 2020

Exactly, needs courage and depends upon the organizational support


aditya pandey
05:03 pm May 16, 2020

Nice One !! Thanks .


Unsung Patriot
10:49 am August 26, 2020

India's story is no different . Being a hub for mass employment of software engineers, there is a lot at stake for managers . They simply cannot digest unpredictability . Senior Management are obsessed with numbers and execution level folks simply have to work around those numbers .Somewhere down the line the key objective of delivering value to customer gets compromised . Hopefully more and more companies should become empathetic towards employees and instill faith in them . Good customer experience will follow .


Celso Recchioni
09:00 pm October 5, 2020

Thanks for the good article and insights. Setting goals and pursuing them also requires planning. Actually plans might be worthless, but planning is essential to set and accomplish achievable goals. Seting expectations and fulfilling them buids the required credibility for the team, so, in spite of working in an uncertain world, some predicability is required. The point is what should be predicted. I fully agree that the delivery of value should be planned, predicted and pursued instead of the delivery of story points or any other unit that is not related to value.


agilemonk
03:31 pm March 21, 2021

Thanks for this article. It gave me the boost that I needed.


DR Reboots
03:36 pm April 14, 2022

Ask your leadership how many times a large waterfall/SDLC project has come-in on-schedule and/or within budget. (and then duck, roll and cover!) :)


Luis Enrique Sanchez
11:04 am August 5, 2022

Very good article! It helps me a lot. The only gap i see is to adapt this cultural swift in the mind (heart) of the people in a pace where they can see the results quickly, why? because we are living (somehow brainwashed) in the now or immidiate results. A cultural swift is like an animal adapting to a new enviroment, its takes time and effort.

This is what we face (one of the challenges) as Agile leaders.


Amr Ashraf
01:58 pm March 18, 2023

Thanks too much for the article.

When I read it, I enjoyed a lot with such words and facts mentioned, and I didn't take care about before and I found a new way of thinking about what personally and professionally doing.