Prologue
April 1899, 7 Hawthorn Street, Dayton, Ohio
It was the first warm day of spring. The two brothers sat on the sprawling front porch, excited over their new and innovative business endeavor.
“Well, Orville, what do you reckon? Should we start with flight dynamics or propulsion, or do you think landing gear could be more important?” the older brother said with a thoughtful gaze.
“That’s a tricky question, Wilbur! Shouldn’t we work on the entire flying machine from the start?”
“Nope, that is too risky with our tiny budget. If we could make one of these things work properly, I think there is a business for us. With a too large scope done by halves, we will just be wasting our time.”
“Ok, but how could we possibly know?”
Wilbur looked patiently at his little brother. “You know, we cannot know, and we can only make assumptions. I propose we do a WSJF with those relative numbers. At least it will, in a short time, help us to become more aligned. Besides, I think it is rather fun.”
“Yep, agree. WSJF is much better than the fighting we used to do when we were kids. But you need to help me out a bit; the model is still not clear to me.”
“Ok, I think User and Time value is pretty straightforward, and it will certainly take some time before customers are queueing up for our services. So, let us focus on Uncertainty reduction and Job size, which we are after right now. For us, nothing is more important than the value of learning.”
“Ah, ok, but that is the one I don’t understand. Is uncertainty a value? I think uncertainty goes into an estimate of job size? I mean, it will take a longer time to do the work with a lot of uncertainties.”
“You have to listen, Orville,” Wilbur said using his harsh voice. “I said, “Uncertainty reduction!” It means we will give a higher score to work that provides the most knowledge. The knowledge that will reduce any important uncertainty for our business. The job size is the time it will take to deliver whatever provides that knowledge.”
“Alright, so what you are saying is, do not estimate the time to success, instead estimate the time to success or failure. Like the first kite we built. It didn’t take much time putting it together, and even if it did not fly very well, we learned a lot.”
“Yes, brother.” Wilbur smiled. “And the best outcome is if we understand both success and failure, then we can decide about the tradeoffs. The sooner we have this understanding, the better it is.”
Wilbur was happy to have his brother on board, and he truly sought their mutual understanding. After some reasoning, their estimates looked like as follows:
Note: WSJF = Uncertainty reduction / Job size. User and Time value is not included since the focus is entirely on learning.
Orville summarized, “Ok, it seems as we have the best chances within Flight Dynamics. So far, the world has not gained much knowledge in this area. Also, we are already quite skilled in building kites, and even with our small budget, we should be able to, rather swiftly, do the needed experiments and get the value of learning.”
The most valuable asset
In many of the training classes I do, there is this exercise where the participants are asked to read an article and then reflect. The article is about how the Wright brothers worked to develop the first successful piloted flying machine. Compared to their competitors, who focused on designing a flying machine, the Wright brothers conducted numerous diverse experiments to create the knowledge necessary to develop a flying machine.
The class exercise explores the importance of learning and how any product development may take advantage of the principles rooted in Lean Product Development. It’s clear that without deliberate investment in knowledge, the risk of failure is much higher.
At the end of the exercise, I usually ask the participants about the assets of their business. I ask, “Would you say that your collective knowledge on how to develop, produce and sell your products are more valuable than all your current products?”
So far, the answer has always been, “Yes!”
How to quantify the value of learning?
Even if intuition tells us that a company’s, sometimes lifelong, collective knowledge is precious, it is seldom prioritized. Money usually talks louder. Larger companies try to report knowledge as an intangible asset in their annual reports, where the value may be found as patents or capital expenditures. But these are, except for maybe in pharmaceuticals, not reflecting a very high value. Nor do they provide a reliable understanding of any future value.
In the Lean Startup, Eric Ries has defined “Innovation Accounting” as a concept to assess gained knowledge. Today, this is an established approach for working with knowledge as an important value, and it creates support to select the best path to the future. Even if most widely used in new businesses, the concept is gaining popularity also in established companies.
I’m sure the Wright brothers would have liked the Lean Startup. It would have helped them to define and assess their experiments. But “Innovation Accounting” can be time-consuming and does not quickly give the oversight of different alternatives. This is where Weighted Shortest Job First (WSJF), as defined by Dean Leffingwell, comes into play as another proactive model that includes the value of knowledge.
In his book Agile Software Requirements, Dean defines Cost of Delay (CoD) as an aggregation of “User value,” “Time value,” and “Risk reduction/Opportunity enablement.” The last of these three attributes, which I rather call “Uncertainty reduction,” opens for a focused discussion of where knowledge is most valuable.
WSJF is widely used in organizations that are improving their ways of working with the Scaled Agile Framework (SAFe) support. The model may be used on any level, from initiatives involving many teams to features developed by a single team. WSJF is, together with techniques, like Kanban, optimizing the entire product development flow.
Naturally, CoD, in the best of worlds, should be calculated using absolute numbers, and the approach to use relative estimates, as suggested by Dean Leffingwell, has also got some criticism. Black swan farming is one of the critics who want to bake the value of knowledge into User Business Value. But with absolute numbers, the estimates will take a longer time, and the swift discussion that creates a common understanding is likely to disappear into calculations made by experts in isolation. Dean writes, “We could fall into the analysis trap … overinvestment in calculating size estimates … plus overinvestment in calculating CoD could lead to too much overhead plus a potential bias by those doing the work.”
Take the opportunity to align
Maybe the Wright brothers had their way of quantifying knowledge, or perhaps it was pure intuition that guided them to prioritize the learning. Perhaps one factor was that they were brothers, thus having a deep understanding and acceptance for each other and being totally aligned. Of course, speculations here, but the point is; common understanding and alignment is key success factor.
I hear people complain about how difficult it is to do estimates and that they do not understand the WSJF model properly. But as an independent facilitator, I always see great and appreciated discussions when WSJF is used. And that is what it’s all about, estimates are never the truth, but discussions are.
In my experience, most organizations start their Agile journey with some kind of “iterative waterfall.” There are backlog items in the shape of activities like analysis, design, coding, or testing. Estimating activities with WSJF is tricky since there could be questions like “Is the analysis more valuable than the design?” But even without a perfect backlog, WSJF can be a great tool to help improve the content of the backlog towards completed deliverables.
Don’t skip WSJF “just” because your prerequisites are not perfect. Instead, take the opportunity to align. If you involve developers in the Job size estimate, they will have the chance to get early insights and feel included. As you go, your WSJF technique will improve. It will be easier and easier to understand the different attributes. Find good examples, working patterns, and guidelines. A good start is to use a model, but when you know it well, you can adapt according to your needs. Dean Leffingwell writes, “Teams may apply a rating scale to their specific context.”
Dealing with uncertainty
Dealing with future uncertainties is not intuitive for us smart human beings. We like to have facts and think we can get them. But that is obviously not true, and we need to accept that we cannot know. As uncertainties grow, it becomes more important to use models that can help us assess future paths deliberately.
The background of Uncertainty reduction is explained by Dean Leffingwell, “Our world is laden with both risk and opportunity. Some features [or any initiative] are more or less valuable to us based on how they help us unlock mysteries, mitigate risk, and help us exploit new opportunities.”
Most people I have seen estimating backlog items like Epics, Features, or Enablers with WSJF, as defined by Dean Leffingwell, are struggling with Uncertainty reduction. To get it straight, Uncertainty reduction needs to be separated from the likelihood of getting value and the likelihood of being able to deliver something.
Uncertainty reduction
The guiding questions in SAFe:
- “Does it reduce the risk of this or a future delivery?”
- “Is there value in the information we will receive?”
- “Will this feature open up new business opportunities?”
The questions are all about finding knowledge that may be valuable in our business. Knowledge with high value naturally gives a high score.
Estimating the Uncertainty reduction is an excellent opportunity to discuss different horizons and find the proper balance between investments in your current or future business. In general, User-business value is higher for the existing business, and Uncertainty reduction is higher for possible future business. Consequently, putting Value on Learning is about motivating investments for the future.
Time criticality can be based on how the user value decays over time but can also be based on how the value of knowledge decays over time. It is sometimes crucial to own a piece of knowledge before competitors get it. When everyone knows, the same knowledge may be hardly noticeable. By this logic, I propose to estimate Time criticality after the estimates of User Business value and Uncertainty reduction.
Likelihood
Likelihood applies to all attributes of WSJF. The higher likelihood, the higher score on the value attributes, and the lower the score on Job size according to the following logic:
- If you have two items with the same estimated value (User, Time, or Uncertainty reduction), the one with the highest likelihood of getting the value is preferable.
- If you have two items with the same estimated Job size, the one with the highest likelihood of finding the technical solution will require less effort.
The separation of Uncertainty reduction from Likelihood is the same thinking as is often used in risk analysis, where the two components are normally called Impact and Probability.
Summary
When prioritizing investments, it is essential to include the value of knowledge. Objective quantification of business value and knowledge is difficult territory, and there are no exact measures. Instead, agreement and common understanding may be created by the use of models like WSJF.
With some patience, WSJF is a straightforward tool that will amplify early involvement, learning and local decision making.
Epilogue
November 2002, Merritt Island, Florida
The two men sat tightly strapped in the huge tilted vehicle. They had been sitting there for several hours while the countdown proceeded. Their mission was extremely advanced and even dangerous. However, as well-trained professionals, they were chit-chatting while routinely walking through the checklists.
“Do you know, it is almost one hundred years since the Wright brothers made their first controlled flight. Just 260 meters in 59 seconds, but that was the start, and we would probably not have been sitting here if it weren’t for them.”
“Yeah, the ones who invented the first flying machine.”
“Well, they invented a bit more than “just” the flying machine. I think it is fair to say, ‘they invented flying.’ Also, they came up with a development process that is today widely used by successful companies and by our own engineers.”
“What, here at NASA? Still? With all these new and advanced computers and so much more knowledge?”
“Yes, indeed. You can read about it on our web page. But now, we must focus on getting this little thing off the ground. We will have plenty of time to talk about the history of flying once we get into orbit. I will tell you more then.”
“I look forward to it. It is hard to grasp all knowledge that human beings have created in the last 100 years. I wonder what we will learn in yet another century.”
References
“The process which leads to the first successful airplane is the same process used by NASA engineers today to solve problems.” wright.nasa.gov/overview“
“The brothers’ fundamental breakthrough was their invention of three-axis control, which enabled the pilot to steer the aircraft effectively and to maintain its equilibrium. This method became and remained standard on fixed-wing aircraft of all kinds” wikipedia.org/wiki/Wright_brothers“
“Their first U.S. patent, 821,393, did not claim invention of the flying machine, but rather, the invention of a system of aerodynamic control that manipulated the flying machine’s surfaces” wikipedia.org/wiki/Wright_brothers
Scaled Agile Framework (SAFe), scaledagileframework.com © Scaled Agile, Inc.
Assessing risk probability, Project Management Institute (PMI) pmi.org/learning/library/assessing-risk-probability-impact-alternative-approaches-8444
Agile Software Requirements: Lean Requirements Practices for Teams, Programs, and the Enterprise by Dean Leffingwell amazon.com
The Lean Startup: HowToday’ss Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses by Eric Ries amazon.com
About the article
By writing this article, I wanted to think through the arguments and explain the thinking behind Weighted Shortest Job First (WSJF). I hope to clarify and make the usage of WSJF easier, especially to assess the value of learning. Also, I really like to read about the Wright brothers and other brave inventors.
The dialogues in this article are a work of fiction. Names, characters, businesses, places, events, locales, and incidents are the products of the author’s imagination or fictitious. If you disagree with something, I will welcome a discussion. It would be my privilege to learn from what you have to teach me.
Thanks to Paulina Janemalm, Kim Lindberg, and Andreas Johansson for proofreading, editing, and advice.
