Rigid performance targets often derail adaptability in a turbulent marketplace. The Guiding Star OKR model offers a framework for creating shared strategic direction, ensuring employee actions across the organization are aligned for success.

Our CTO George’s team meeting is a fortnightly gathering of the heads of the Project Management Office (PMO), product, and strategy, as well as our agile coach. The CEO’s eighteen-year-old daughter, Stella, is also present at her father’s request. George suspects his boss has big dreams for his daughter. However, for now, she spends most of her time as a university student.

This time, there was something expectant in George’s eyes. It was as if he was about to unveil his plans for a major transformation, and he was curious about the team’s feedback.

CTO shares his proposal

Good morning, team.

I’m here today to talk to you about a proposal for a new standardized process for our development department. As you know, we recently crossed the 500-employee mark, and with our continued growth, it’s become increasingly important that we have a consistent and efficient way of working.

I’m proposing a process that is designed to do just that. The process will ensure that everyone on the development team works in the same way and that we’re all on the same page. This will reduce the cost of our work, increase clarity, and make it easier to hold people accountable.

The process is based on agile methodology, but it’s been scaled up to create a feature factory. This means that all of our work will be broken down into four levels: themes, initiatives, features, and stories.

We will implement follow-up, what agilists call standups, at all four levels of the organization, with more frequent meetings at the lower levels. This allows us to identify and address any deviations from our already busy schedule early on.

It’s important to note that this process is not intended to replace the work that our teams are already doing. In fact, it is designed to support and enhance that work. By providing a clear framework for how we work, we can free up our teams to focus only on our top-priority features.

I believe that this process will have a number of benefits for our team, including:

• Reduced costs
• Increased clarity
• Improved accountability
• Increased motivation
• Scalability

I am confident that this process will help us to continue to grow and succeed as a company. I’d like to hear your thoughts and feedback on this proposal.

Head of PMO shares his concerns

As the head of PMO, I care deeply about the success of our projects and our company’s compliance with regulatory requirements. Our project managers are focused on delivery and compliance. Give us a scope, a deadline, and resources, and we’re prepared to go through fire to deliver on those terms.

Of course, our company will continue to manage projects even if they are not visible in the feature factory picture. We must honor the contracts we have signed with customers who have paid large sums of money for customized solutions. If we fail, we risk our brand and reputation.

If we don’t meet the regulatory requirements of the authorities, we can be fined and, in the worst case, banned from doing business. Nothing is more important than avoiding this.

While the proposal is good, I see some risks.

• It is difficult for us, as project managers, to prioritize the work of our resources, since the teams plan and track their own backlogs.
• Projects become more difficult to monitor and evaluate. Teams break down all work into small pieces and mix work from multiple projects into a common queue.
• By breaking down the work based on organizational departments, it becomes difficult to coordinate across these silo boundaries. This leads not only to ambiguities about integration points but also to unnecessary misunderstandings and conflicts.
• The absence of overall transparency in the projects makes them less traceable. What decisions have been made and by whom? Without that information, we will not be able to hold anyone accountable when deliveries fail. Even worse, a lack of transparency unfortunately opens up risks for corruption and irregularities.

I want to remind you all that project management is essential for the success of any organization. Project managers play a critical role in ensuring that projects are delivered on time, on budget, and within scope. We also play a vital role in ensuring that organizations comply with regulatory requirements.

Head of the PMO’s key message

• Progress monitoring is aggravated by fragmented and distributed work breakdowns.
• Misunderstandings and conflicts follow from a decrease in overall transparency.
• Project precision matters.
• Meeting regulatory requirements ensures avoidance of fines and bans.
• The company’s reputation is at stake when tier-one delivery is late.

Agile coach shares his concerns

As an agile coach, I care deeply about our teams and their intrinsic motivations. These 500 people are highly educated and experienced in their respective fields. We leverage their expertise by organizing them into cross-functional, semi-autonomous teams. This is what it takes for us to be successful as a company.

So, I’m glad your suggestion starts with agile teams. However, I also have some concerns.

• We know that intrinsic motivation comes from things like autonomy, mastery, purpose, growth, and social interactions. These factors trump carrots and sticks.
• Our teams need to own their own processes, of course, within certain guardrails. Otherwise, they’ll be prevented from continuously improving the way they work.
• The technical solutions we deliver are often impossible to describe in detail in advance. Teams must learn by experimenting with different approaches.
• Prioritization can’t just come from the top in the form of large feature breakdowns. Only the teams can warn us about and address technical debt. They, not the managers, are also experts on the optimal technical solutions.
• We also need to listen to the teams when it comes to estimating work. No one can better guess how long it will take and how complex the tasks are than the people who will be doing the work.

I’m not saying that a feature factory is a bad idea in principle. However, I’m concerned that it would be a step backward for our teams and our company.

I urge you to reconsider this proposal and work with the teams to find a solution that preserves their autonomy, mastery, purpose, growth, and social interactions.

Agile coach’s key message

• Agile teams are key to success.
• Autonomy, mastery, purpose, growth, and social interaction drive teams’ intrinsic motivation.
• Teams should own their processes to improve continuously.
• Technical solutions may emerge through experimentation.
• Teams are experts in estimating their work effort and technical solutions.

Head of product shares her concerns

My product managers always think about the customer first. Our products should be easy to use and solve the problems that customers face. It’s a sign of the times that companies can’t keep delivering specific solutions to the highest bidder. With popular products, we can reach many more people. Together, they may pay more than a single tier-one customer.

• The static slicing of teams easily results in a bunch of component and platform teams that have difficulty relating to the customer’s reality. I would like the slicing to be based on the business — that is, customer types or product types.
• Just because we improve our ability to churn out features doesn’t mean we’ll get a consistent customer experience. Customers will get lost, and longer flows in their work will require manual steps.
• All the roles, levels, and sync meetings described in the feature factory picture might lead to longer lead times. We want a hyper-fast loop based on build-measure-learn. Our product people are awesome at doing market research and learning from our experiments.
• Feature factories tend to focus on specs and forget about the big picture. They churn out features without considering what customers want or what competitors are doing. This can be a recipe for disaster. Just look at Nokia. They were once the world’s leading mobile phone manufacturer, but they fell behind when they failed to keep up with the latest trends. Now they’re a shadow of their former selves.
• I can’t see that we’ll be measuring anything other than the amount of output in a feature factory. For us, it’s more important to measure customer satisfaction and business value. We want to see a paradigm shift from tech-focused proof-of-concept to proof-of-value.
• Tech-driven development is a solution that looks for a problem. It may impress some people that we can intercalate bleeding-edge technology, but it doesn’t help our customers. And it’s actually the customers that we should keep closest to our hearts if we want to be successful.

Several of these problems stem from the fact that a feature factory creates a gap between the development department and the product managers.

Head of product’s key message

• Products should be easy to use and solve customer problems.
• Teams should be sliced by customer or product type.
• Feature factories focus on specs and forget about the big picture.
• We should shift from tech-focused proof-of-concept to proof-of-value.
• Market research and fast experimentation feedback are needed.

Head of strategy shares her concerns

As the head of strategy, I care deeply about our company having a long-term and coherent strategy. I also care that everyone knows the strategy and that it has an impact on the work done in the company. Another of my passions is that the company should be a learning organization.

I am concerned that the feature factory would lead to:

• Incoherent product portfolio. The scattered focus on individual features is organizational multitasking. Successful strategies require us to believe in them persistently.
• Strategies being sidelined. The strategies we have developed are sidelined by the execution in the feature factory. The dual command, execution versus strategy, confuses the staff. The result is usually that the implementation of the strategies is sacrificed, which puts the whole company in a risky position.
• Lack of risk analysis and management. As you surely remember from examples from our company’s history, a lack of risk analysis and management has led to catastrophic product launches. Let’s never let that happen again.
• Lack of learning. Our governing documents state that we should be a learning organization. We will never be able to “sharpen the saw” if everyone works in their own little silo.
• Lack of broad competence development. Our employees are asking for broader competence. This is also a clear industry trend. What is the use of developing broad competence if the work still revolves around a small, recurring problem area?

I see a general risk with feature factories that they prioritize urgent over important. Yet, we know from our own experience that the more time we spend on firefighting now, the more time we will need to spend on firefighting in the future. Without proactive work, what was important yesterday becomes urgent today.

Head of strategy’s key message

• The company needs a long-term and coherent strategy.
• Everyone should know the strategy, and it should impact the work.
• The feature factory may result in a lack of risk analysis.
• The company should be a learning organization.
• Employees are asking for broader competence development.

CTO is disappointed

I’m disappointed. I was hoping for more passion, more willingness to change, and more of a winning instinct. While I can understand some of your concerns, I find it hard to accept the one-sided whining. We are a team working for each other towards a fantastic goal, aren’t we?

With the attitude you’re showing today, I might as well ask Stella for feedback. She’s only seventeen and has no work experience at all. But with those genes, I’m sure she’ll recognize a good idea when she hears it.

Stella, what do you think about the idea of modernizing the development department into a state-of-the-art feature factory?

Stella tells about the duck and the ants

Eighteen. I’m eighteen years old.

(Ten seconds of silence may sometimes be perceived as ten minutes.)

Have you heard of the Digesting Duck? It was a sensation in its day.

The life-size duck was made of copper, gold, and steel. It could flap its wings, quack, and even appear to eat and digest grain. French eighteenth-century inventor, Jacques de Vaucanson, achieved this by using a complex system of gears, levers, and springs. However, the “digestion” was actually fake: the grain was stored in a hidden compartment and released at an appropriate time.

Contrast this system of a mechanical duck and its master with a self-organized ant colony with tens of thousands of individuals.

When an ant finds a food source, it leaves a pheromone trail on its way back to the nest. Other ants then follow this trail to find the food source. In this way, ants can collectively carry home a dead lizard that is many times heavier than any individual ant. The ants achieve great things together, despite lacking centralized leadership.

Does our 500-person company resemble the master and his mechanical duck or the decentralized ant colony?

(It seemed like a rhetorical question that was supposed to be answered with the second alternative. But it was not.)

Actually, both! And it’s not by design. It is dictated by the laws of complexity science. Let me explain.

Stella’s key message

• Digesting Duck had centralized control and fake sophisticated capabilities.
• An ant colony is a decentralized system that creates complex outcomes.
• Each ant operates based on a simple set of rules.
• Dual nature: A 500-person company may resemble BOTH the duck and the ants.
• Dual nature emerges from the laws of complexity science.

Stella defines the company as a complex system

When 500 people work together toward constantly changing goals, they interact with each other many, many times, and in many different ways. No one here can predict in the morning all the interactions they will be involved in during the day. Most interactions happen in completely new ways that have never been done before, anywhere. (Numerosity)

Some meetings are planned and go more or less as we planned. However, most interactions are not centrally coordinated or controlled. They arise from unexpected combinations of previous interactions. (Diversity)

The interactions are therefore influenced by the results of previous interactions. Feedback is built into the system, and the system continuously adapts to new conditions. It never ends. (Feedback)

Exactly what we will engage in is partly or wholly not knowable in advance. This imbalance is rooted in the fact that the second law of thermodynamics applies only to a closed system. Our company is an open system. We will never stop being influenced by competitors, customers, laws, economic conditions, and other things that add energy from the outside. (Non-equilibrium)

Regardless of how much we add in terms of hard process rules, our company will still be partly self-organized. The past will be part of the future, and people will always adapt to any new reality. And, actually, we should take advantage of that spontaneous order. How do we do that? By creating fertile soil that focuses these emerging forces in a common direction. An elevated purpose makes our whole more than the sum of our parts.

Stella’s key message

• Numerosity. Many people interact many times in many different ways.
• Diversity. Many meetings, decisions, and analyses are not centrally controlled.
• Feedback. Employees will never stop adapting to new conditions.
• Non-equilibrium. Our company is constantly influenced by external factors.
• Self-organization. We are partly self-organized, and we should leverage that.

Stella supports both our CTO and the team’s feedback — Whining is opportunities

I support George’s feature factory model, even though the word “factory” leads us astray. Neither we nor our office are machine-like. A machine cannot change its behavior when the environment changes. Our level of creativity is incredibly important when markets, customers, suppliers, and distributors are constantly changing.

I also support the feedback that all of you gave to George. Even though it is partly biased, it is relevant and based on experience. Instead of dismissing the feedback as whining, we can see it as opportunities.

A complex system is nested. At the micro level, our employees interact and self-organize, although they will never possess all valuable information. The impact is the great achievements that emerge on the macro level.

The feature factory is a framework for executing what we have already decided to do. It may help us deliver quickly, with high quality, and at low cost. Unfortunately, that’s not enough.

We also need to align and re-align our company’s energy with the most important long-term goals, even though our surroundings never stop changing. Since the feature factory model does not address this, it must be complemented by a flexible, sensing, and uniting strategy process.

If you allow me, I would like to propose an addition to the feature factory. My proposal addresses most of the concerns we have heard here today. It would also make the development department more resilient.

Stella’s key message

• She supports both feature factory and team feedback.
• However, the word “factory” leads astray. We are not machine-like.
• A feature factory focuses on execution, however…
• …the feature factory does not address the most important long-term goals.
• The proposal must be complemented by a flexible, sensing, and uniting strategy process.

Stella talks about guiding stars and traffic lights

Ants have the natural instinct to survive and reproduce. This instinct drives them to cooperate in large groups to search for food, build nests, and protect themselves from danger.

Similarly, even though our goals may change over time, we also need to agree on our common overall goals.

Why do we need to set goals?

• It is strategic for our company. We begin with the end in mind, thinking before acting.
• Combining overall alignment and decentralized self-organization is a challenge when 500 employees deliver value together. We want our people to be a focused force, with maximum value, without micromanagement. And we can achieve this combination with shared goals.
• We want to influence our operative prioritization. What initiatives and epics contribute to our long-term goals?

Note that this is not about measuring our performance, designing our solutions, managing our scope, or planning our resources. There are other tools that are better suited for doing those things.

So don’t think of a red-amber-green traffic light. They show us whether we are currently ahead of or behind the tactical plan. However, they say nothing about how to act next.

Instead, think of a guiding star that helps us shape our tactical plan. Regardless of where we are right now, the guiding star always illuminates the direction in which we need to move forward.

Guiding stars are essential for any successful company. They provide a sense of purpose and direction while giving our people the freedom to self-organize.

Stella’s key message

• Ants naturally cooperate based on shared goals for survival.
• Shared goals are strategic — they begin with the end in mind.
• Guiding stars are goals that help us prioritize operative work.
• Guiding stars are not for tracking performance.
• Guiding stars give purpose and direction while allowing for flexibility and self-organization.

Stella dives into the importance of hierarchy

Dictating goals from the top is a recipe for disaster. However, hierarchy still plays a crucial role in our success.

Think of complex challenges like a chess game. Agility trumps the brute force. Military strategist John Boyd described the information process as a loop: observe, orient, decide, and act (OODA loop). If our loop spins faster than competitors’, it’s like getting two moves in chess for their one.

Here’s the surprising part: when the entire company participates in the OODA loop, we become faster, and our goals become more accurate.

• Everyone observes. We gather more information and insights from a wider range of perspectives. We might, for example, catch market shifts, solution glitches, and innovative ideas earlier.
• Everyone orients. This diversity of experience protects us from blind spots and biases that plague top-down companies.
• Everyone decides. Take away decision-making power, and not only do decisions get worse, but our employee’s intrinsic motivation will plummet. Who wants to wait for orders if they can’t contribute their ideas?
• Everyone acts. We hired these 500 amazing people for a reason, right?

The challenge is aligning our company in the last two steps: deciding and acting. There are endless good things we may pursue, but scattering our energy won’t get us anywhere. To truly thrive, we must focus in a unified way.

Listen closely now because here’s how we can achieve that.

Stella’s key message

• Top-down goal setting doesn’t work in complex systems.
• Hierarchy still matters, since joint outcomes matter.
• Faster OODA than competitors’ is like getting two moves in chess for their one.
• Everyone should contribute to the OODA loop for speed and accuracy.
• Shared goals align decisions and actions across the company.

Stella explains how insights may flow up and down, across the organization

Information and insights must flow freely, up and down, across departments. In this way, we make our decisions collaboratively, with a strong foundation built from all corners of our office.

I’m proposing that we use the popular goal-setting framework called OKR, combined with a process inspired by the Japanese hoshin kanri. However, we will not do it by the book.

One month before our quarterly planning, the hoshin kanri process kicks off. You, George — our CTO — start by presenting your draft OKRs to our portfolio managers. You ask them if they have questions or reflections about these goals and welcome feedback. Then, you ask them to come back with their own OKRs.

There are some properties they have to consider regarding their OKRs:

• Inspired or local? Some OKRs are inspired by the level above, while others are specific to their portfolios.
• Shared or owned? The same OKR might appear on multiple peer lists. It makes collaboration easier when everyone’s rowing in the same direction. Other OKRs might be unique to a single portfolio.

Most often, some OKRs are inspired and some are local. Some OKRs are owned and some are shared.

This process then continues both upwards and downwards through the organization — all levels. Don’t forget upwards! The feedback might reveal that something’s missing at a higher level. After a few weeks of discussions, we can finally say we’ve nailed down the strategy for the next three months.

Note that goal horizon and review cadence are two different things. Higher levels in the hierarchy tend to have longer strategic horizons. The whole development department might have goals spanning three to five years, while teams might focus on two-week sprints. But that doesn’t mean that top levels can’t revisit and adjust — even swap out — their OKRs every quarter.

A hierarchy can work hand-in-hand with agility and participation. Managers own and represent the goals on their own level. However, accountability is shared. By letting information flow freely and involving everyone in the entire OODA loop, we will be a powerful, united force that achieves remarkable goals.

Stella’s key message

• Insights should flow up, down, and across the organization for better decision-making.
• Using a modified version of OKRs and hoshin kanri for goal-setting.
• Inspired, local, shared, owned — OKRs have a mix of these properties.
• Goals are set through discussions at all levels, including feedback going upwards.
• Hierarchy can coexist with agility when managers own goals and share accountability.

Stella introduces a new kind of OKR: guiding star OKRs

We will not do OKRs by the book. Traditionally, OKRs have been used to put pressure on and blame individual managers. Research shows that this approach is flawed. It only demotivates our people. My strong recommendation is not to connect OKRs to incentives or appraisals.

Guiding star OKRs are different. Rather than measuring our performance, they emphasize a shared vision for the future and then align our work toward that vision.

A guiding star OKR starts with an engaging story about the future, for example, “Easy to buy.” We call the story an “objective,” that is, the “O” in OKR. Our objective is also supported by two to five measurable parameters, which we call “key results.” Together, our objective and our key results describe two things:

• Our current state.
• Our desired future state.

There are also several things that do not belong in a guiding star OKR:

• System solutions for reaching future goals.
• Activities that take us to our future goals.
• Features and deliverables that are necessary for future goals.

Thus, we set the goal before we consider the plan; strategy drives execution, not the other way around. If a guiding star can be achieved in only a single way, then it is too specific and too short-term. It doesn’t really serve any purpose. We need to think bigger and further away.

Note that our guiding stars cannot be prioritized against each other. When there is a need to prioritize, then we have too many guiding stars. Typically, a team, a product manager, a portfolio owner, or our CTO has one to three guiding stars each at a time — never more. We say that we select in strategy and prioritize in execution.

Stella’s key message

• Guiding stars are used for shared vision, not measurement.
• Guiding stars describe the current and the desired future state.
• Guiding stars are not system solutions, activities, or features.
• Guiding stars makes strategy precede execution.
• Guiding stars should be limited to avoid prioritization conflicts.

Stella dissects the objectives of guiding star OKRs

Don’t think for a second that we will be able to derive everything we do here at the company from our guiding star OKRs. Guiding stars illuminate what is important but not yet urgent — what we need to start today, saving us from putting out fires twelve or eighteen months from now. We will still do incidents, operations, maintenance, and other things that keep our lights on. Don’t worry that those tasks will disappear from the radar. Our OKRs are guiding stars for transformative work and investments in completely new capabilities.

If we start with the objectives, they are qualitative and free from numbers. Here are some characteristics to strive for in our objectives:

• Share a short campfire story that everyone can understand and like.
• Represent future outcomes that are valued by customers, users, or other business units, not just by those who contribute to this guiding star.
• Describe things that we have a lot of influence over, not, for example, financial goals or market positions.
• Mark substantial differences between the current state and the future target state.
• Are not multiple goals merged into one.

You may think of the objectives as principles that we return to when we have to choose between multiple options. The question we ask ourselves at those times is: Which option takes us closer to our guiding stars?

Stella’s key message

• OKRs illuminate important but not yet urgent tasks that focus on new capabilities.
• Objectives are short campfire stories focused on customer-valued outcomes.
• Objectives focus on outcomes that we can influence, not just on financial or market goals.
• Objectives show a clear difference between the current and desired future states.
• Objectives help guide choices by aligning decisions with the guiding star.

Stella dissects key results of guiding star OKRs

We illustrate each objective with two to five key results.

Unfortunately, it is not difficult to cheat to achieve all key results but still not moving towards our objective. Therefore, consider key results as examples, not proofs. Our overall goal is to achieve our objective, regardless of whether we achieve the key results.

A key result is a variable X with a current state X = A and a target state X = B. For example, “Maximum number of seasons per year in the customer’s clothing store” is four today, and the goal is eight:

• Current state: X = 4.
• Target state: X = 8.

Other properties of key results

• Measure the state, not the effort spent.
• Never contains verbs such as “launch,” “create,” “develop,” or other action words indicating tasks.
• Use numbers, not Boolean values “true” and “false.”
• Trackable continuously, if possible, offering proactive insights.

I can’t emphasize enough that key results are only tangible examples of the future state, not evidence of success.

Stella’s key message

• Key results illustrate the objective, but achieving them doesn’t guarantee success.
• Key results focus on the measurable state of the objective, not the effort invested.
• Key results lack verbs like “launch” — they aren’t about tasks.
• Key results use numbers to define the current and target state — not just true/false).
• Key results should be trackable continuously to provide proactive information.

Stella shows how guiding star OKRs are customer-centric

Guiding stars do not focus on our current products and services. Of course, our company works in other contexts with improvements, growth, and operation of our current offering. However, guiding stars describe completely new capabilities — things our customers could not offer their customers before.

We tend to be too nearsighted. Even our customers’ customers have a job-to-be-done (JTBD) — a fundamental task or goal they aim to accomplish in a given situation. Our customers’ customers will achieve better results by operating in new and better ways. Therefore, the objectives in our guiding stars are not tech solutions. However, they may indirectly generate scope in tech solutions. You have heard it before today: strategy drives execution, not the other way around.

The key results in guiding stars are derived from opportunities. They are examples of how the customer’s customers would like their realities to be. Key results are missing pieces that help them achieve their desired outcomes. Whenever our customer’s customer complains, we have an opportunity to translate the whining into key results.

As you know, we develop data warehouses for retail businesses. Imagine that their customers, the consumers, want to buy the lipstick they usually use — they couldn’t care less about features in our software. Their JTBD is something like: “I’m buying my (essential) lipstick.” Let’s listen to what these lipstick consumers complain about:

• “I can’t find anything on this shelf. It’s overstocked.”
• “I am unable to find what I want and inspire myself!”
• “I don’t feel confident that I’ll get competitive prices or that I’ll get rewarded for coming back.”

See, there we have three opportunities that we will translate into measurable key results.

The outcome they seek is in line with “Easy to find, easy to understand, easy to buy.” Let’s use that as the “objective” in our guiding star OKR.

This is what it means to be customer-centric. Rather than unreflectively implementing many cool features in our products, we point out a direction that gives our customers’ customers new behavior.

Stella’s key message

• Entirely new customer needs, not incremental improvements to existing products.
• Our customers’ customers don’t care about our features. They want their job-to-be-done to be achieved.
• Strategy comes first — guiding stars set direction, not solutions.
• Key results are derived from our customers’ customer pain points.
• By making things easy for our customers’ customers, we drive new behavior and create value.

Stella explains a process for guiding star OKRs

You are probably familiar with some of the common barriers to success in strategic work:

• Strategies are intended to be implemented alongside, rather than as an integrated part of, execution.
• Strategies are created top-down and, therefore, do not reflect all of our knowledge, nor do they speak the language of the teams.
• It is unclear which goals are relevant at the moment.
• Strategies are derived from our execution plans rather than driving execution.
• Strategies have a backward-looking follow-up at the end, rather than continuous review and adaptation.
• Strategies are only synchronized at formal meetings by selected delegates.

To address all of these obstacles, I envision a process inspired by hoshin kanri but with relevant additions. Hoshin kanri originated in post-war Japan but has since spread to the U.S. and around the world. The words “hoshin” and “kanri” mean “direction” and “administration.” We might say “compass management.”

I previously described how guiding stars are set and owned at each level through conversations organizationally upwards, downwards, and sideways. It is not about who decides but about synthesizing insights from all levels. We call this process catchball. I throw the ball and see if it is caught.

Since our company has a quarterly rhythm for planning our scope to implement, the catchball process starts one month before our quarterly planning. Then, one week before the quarterly planning, the agreed guiding stars are communicated to everyone in what we call strategy broadcast. For everyone to understand, we also arrange lean coffee sessions in connection with this, where we hope that unexpected questions and thoughts will come up. The earlier we learn, the better.

Transparency is necessary for our employees to feel confident in our guiding stars. There should never be any doubt about what the latest version is. Therefore, guiding stars must be available in a central location that is accessible to everyone. Private presentations are not original.

By having catchball first and then strategy broadcast well in advance of quarterly planning, both teams and product managers are influenced by our guiding stars when planning the scope for the coming quarter.

Guiding stars are exactly what the name says: continuous support during execution when we have multiple options to choose from. That’s why we have guiding stars check-ins fortnightly. In the first half of that meeting, we ask ourselves whether our key results are still relevant — are they still tangible examples of our objectives? If the answer is “yes,” then we review their movement since our last check-in. We dedicate the second half of our meeting to lean coffee.

Regular formal meetings are important. The decisions that we come up with in those meetings are our official version. Alongside that story, we want new, exciting thoughts and insights to flow more informally through the organization. That’s why we arrange open space for one hour every Friday morning. Everyone is invited, but presence is not mandatory. This is where unknown unknowns might be uncovered, and new acquaintances are created across silo boundaries.

Stella’s key message

• The guiding star process is inspired by hoshin kanri to address key strategic obstacles.
• Guiding stars are defined collaboratively through the catchball process for insights across the organization.
• Guiding stars are transparently communicated through strategy broadcast and lean coffee sessions.
• Strategy influences quarterly planning sessions.
• Guiding stars drive decision-making during the execution phase, with regular check-ins and open space sessions.

Stella rushes off

“This is amazing!” George exclaimed with a broad smile. “Where did you learn all this?”

“I visit companies. Talk to people. Read books,” Stella answered. “That kind of stuff.”

Stella glanced at the clock and gasped. “Oh my gosh! I’m talking too much. I have to hurry to my next lecture.” She grabbed her bag and hurried out the door without saying goodbye.

The heads of PMO, product, strategy, and our agile coach were left sitting in the conference room, as well as our CTO, George.

What do you think crossed their minds at that moment?

This essay on OKRs is a sneak peek at the ideas you’ll find in my book, “Guiding Star OKRs: A New Approach to Setting and Achieving Goals.” I was encouraged by the positive feedback to expand on the concepts here and offer a more comprehensive guide. If you find this piece helpful, I believe you’ll gain even more valuable insights from the book.

Learn more and grab your copy of the book here: https://pragprog.com/titles/snokrs/guiding-star-okrs/

The ‘Guiding Star OKRs’ Approach to Strategic Agility was originally published in The Pragmatic Programmers on Medium, where people are continuing the conversation by highlighting and responding to this story.

While it may not be possible for a human to have a friendship with a gadget in the traditional sense, the kitchen timer can sometimes support you remarkably well in achieving your goals. This might be especially true when the kitchen timer is red and round — eventually shaped like a tomato.

Good friends understand and share each other’s feelings, offering support and compassion during difficult times. They know what it feels like to walk in each other’s shoes. This phenomenon is called empathy.

Even though neither I nor the clock feel empathy for each other (a gadget, as we know, has no emotions), I trust the clock like two friends trust each other. When I have chosen one — and only one — task and wind up my kitchen timer, that task becomes my sole focus until the clock rings, regardless of how many other brilliant ideas may pop up in my head or how much I crave something to eat. I trust that my timer will ring when it’s time to take a break.

Friends understand that mistakes happen and are willing to forgive each other. They sincerely apologize, especially when the person who errs shows that they have learned something. Buddies know that flaws are a part of life and forgiveness is central to friendship.

The clock cannot forgive, since it neither speaks nor acknowledges mistakes, but the clock and I still accept each other for who we are. The kitchen timer always supports my choice of task without judgment or criticism. My job is to choose a task and then pay attention to it until the timer rings. The timer’s task is to notify me after a specified amount of time. Even when I don’t have a clear solution in mind, it helps me focus.

Friends offer unconditional love to each other, providing a source of comfort and companionship throughout life’s ups and downs. Their often long, shared history makes them feel safe when they are together.

The clock does not harbor love for anyone, neither with nor without conditions. However, we have a shared interest: I should not multitask. If I attempt to do many things simultaneously, then I don’t need a kitchen timer. A side effect of multitasking would be that I wouldn’t complete any tasks efficiently due to divided energy. The clock needs me, and I need the clock.

Reciprocity, laughter, and equality are other traits that characterize friendships between humans but, of course, do not exist in my relationship with the kitchen timer. However, my clock offers me other kinds of support, such as efficiency, consistency, and transparency.

That’s why, ultimately, I consider my red round kitchen timer to be a cherished companion.

Staffan Nöteberg is the author of The Pomodoro Technique Illustrated, published by The Pragmatic Bookshelf.

You can also read The Pomodoro Technique Illustrated on Medium.

The Kitchen Timer Is My Friend was originally published in The Pragmatic Programmers on Medium, where people are continuing the conversation by highlighting and responding to this story.

A Large Language Model (LLM), like ChatGPT, generates synthesized text based on statistical inference. You may think of that processing as the average or the most frequent value in its training dataset — wherever that dataset came from. How does the LLM approach differ from human intelligence? Well, I asked ChatGPT seven questions.

[Staffan] A company seeks advice from an LLM regarding their CEO recruitment. Imagine that the LLM’s response, based on statistical inference of its training dataset, suggests that many successful CEOs are white males. How does this type of answer relate to a lack of intelligence?

[ChatGPT-3.5] The kind of answer provided by the LLM, based solely on statistical inference and without considering a broader range of factors, can be related to a lack of intelligence in certain contexts due to its limitations:

• Simplistic Analysis: The answer provided by the LLM is a simplistic analysis based on a single demographic factor (white males) without considering the complexity of what makes a successful CEO. It lacks the ability to delve into the multifaceted qualities, skills, and experiences that contribute to effective leadership.
• Bias and Lack of Nuance: Relying solely on historical data might perpetuate biases and lack nuance. Successful CEOs can come from diverse backgrounds and possess various qualities beyond demographics. This answer disregards the potential of qualified candidates from different gender, racial, and ethnic backgrounds.
• Contextual Understanding: Understanding what makes a successful CEO requires contextual understanding, including factors like industry, company culture, and leadership style. An intelligent response would consider these contextual nuances rather than relying on broad generalizations.
• Lack of Adaptability: Intelligence involves adaptability and the ability to consider evolving trends and changing societal norms. The LLM’s response lacks the adaptability to recognize shifts in the definition of success and leadership qualities.
• Ethical Considerations: An intelligent response would also consider ethical considerations, such as diversity and inclusion, rather than merely perpetuating traditional stereotypes.

In summary, the lack of intelligence in this scenario refers to the limitations of the LLM’s response, which relies solely on statistical patterns without engaging in critical thinking, contextual understanding, adaptability, and ethical considerations. Intelligence involves a more comprehensive and nuanced analysis of complex topics like CEO recruitment that goes beyond simple statistical inferences.

[Staffan] In what ways is the fictive (?) CEO example similar to the COMPAS system used by U.S. police?

[ChatGPT-3.5] The fictive CEO example and the COMPAS system used by U.S. police share similarities in terms of the potential biases and limitations inherent in using data-driven systems to make decisions:

• Bias in Data: In both cases, there’s a risk that historical data used for decision-making may contain biases. The fictive CEO example’s reliance on the historical success of white males perpetuates a bias, and the COMPAS system has been criticized for reflecting racial biases present in the criminal justice system.
• Simplistic Generalizations: The fictive CEO example oversimplifies success based on a single demographic factor. Similarly, the COMPAS system categorizes individuals into risk levels without considering the complexities of individual circumstances, leading to potentially unfair judgments.
• Contextual Nuances: Both scenarios ignore the importance of considering contextual nuances. The CEO example doesn’t account for the diverse skills needed for leadership, and the COMPAS system doesn’t consider the broader context of a person’s life that might contribute to criminal behavior.
• Ethical Concerns: Both cases raise ethical concerns. The fictive CEO example perpetuates inequality, and the COMPAS system’s biases can lead to unjust outcomes, disproportionately affecting certain groups.
• Lack of Accountability: In both cases, the decisions made by these systems can lack transparency and accountability. The basis for decisions may be unclear, making it challenging to address biases or challenge outcomes.
• Limitation of Intelligence: Both scenarios highlight the limitations of relying solely on data-driven approaches without the nuanced thinking, ethical considerations, and contextual understanding that human intelligence provides.

In essence, the similarities between the fictive CEO example and the COMPAS system underscore the potential pitfalls of relying solely on data-driven decision-making, especially when it ignores the complexities of human experiences, biases, and the need for ethical considerations.

[Staffan] How does King Solomon’s decision to divide the baby relate to human intelligence?

[ChatGPT-3.5] King Solomon’s decision to propose cutting the baby in two halves, as recounted in the biblical story, is often cited as an example of a wise and intelligent decision. This story highlights several aspects of human intelligence:

• Problem-Solving and Critical Thinking: King Solomon’s decision was a creative way to reveal the true mother of the baby. By suggesting to divide the baby, he tested the reactions and emotions of the two women involved, ultimately uncovering the genuine maternal bond.
• Emotional Intelligence: Solomon’s decision required an understanding of human emotions and behavior. He recognized the depth of a mother’s love and the intense emotional reaction it would evoke.
• Perspective-Taking: King Solomon demonstrated the ability to put himself in the shoes of both women, considering their viewpoints and predicting their responses to his proposal.
• Innovative Thinking: Solomon’s solution was unconventional and unexpected. It showcased his ability to think outside the box and come up with a solution that was both effective and thought-provoking.
• Ethical Considerations: The story also highlights ethical intelligence, as Solomon’s ultimate goal was to uphold justice and identify the true mother without causing harm to the child.

In this context, King Solomon’s decision represents a combination of problem-solving skills, emotional understanding, ethical considerations, and innovative thinking that are often associated with human intelligence.

[Staffan] Can an LLM devise an original solution to a problem, akin to Solomon’s decision?

[ChatGPT-3.5] A Large Language Model (LLM) like me can generate responses that may appear to be original solutions to problems, but these solutions are based on patterns and associations present in the training data. While I can simulate creative thinking, my responses are essentially a combination of existing text and concepts I’ve learned from my training, rather than true understanding or independent problem-solving.

The “unexpected insights” generated by LLMs are essentially extrapolations or combinations of patterns from the training data, and they do not involve genuine comprehension or independent thinking. These models are not capable of true creativity, critical thinking, or the deeper understanding that underpins human insights.

The story of Solomon’s decision to cut the baby in two halves involves a deep understanding of human emotions, motivations, and complex ethical considerations. Such a solution goes beyond the capabilities of current LLMs, which lack genuine emotional intelligence, empathy, and the ability to fully grasp the complexities of real-world scenarios.

In summary, while LLMs can generate responses that might seem creative or innovative, they are fundamentally limited by the patterns in their training data and lack the depth of understanding required to come up with solutions like Solomon’s that involve complex human emotions and ethical dilemmas.

[Staffan] In what ways does golem’s execution of shems resemble an LLM’s execution of prompts?

[ChatGPT-3.5] The concept of a golem executing shems (magical symbols or commands) is often rooted in mythology and folklore, particularly in Jewish traditions. While the idea of a golem carrying out tasks based on symbols might appear to involve some form of intelligence, it’s important to clarify that this is a fictional and symbolic representation rather than a reflection of actual intelligence.

A golem is often depicted as an artificial creature animated through mystical means, such as inscribing specific symbols on its body. The golem’s actions are typically guided by these symbols or commands, and it follows a predetermined set of instructions without true understanding or conscious decision-making.

The resemblance between a golem’s execution of shems and an LLM’s execution of prompts lies in the idea of following specific instructions to generate responses or actions. However, there are important distinctions to consider:

• Instruction-based Execution: Both scenarios involve executing actions or generating responses based on instructions. In the case of the golem, shems are symbols or commands that dictate its behavior. In the case of an LLM, prompts are text inputs that guide its language generation.
• Lack of True Understanding: Similar to a golem’s execution of shems, an LLM lacks true understanding. While LLMs can generate coherent and contextually appropriate responses based on patterns in the data, they don’t genuinely comprehend the meaning or implications of the text.
• Limited to Learned Patterns: Just as a golem’s actions are limited to the prescribed instructions in shems, LLMs generate responses based on patterns learned during training. They do not possess the ability to go beyond these patterns to form genuine insights or creativity.
• Lack of Contextual Awareness: Both golems and LLMs lack contextual awareness. Golems follow commands without understanding the broader context, and LLMs generate responses based on patterns without having an understanding of the conversation’s larger context.
• Predetermined Behavior: Both scenarios involve predetermined behavior. Golems act as programmed, and LLMs generate responses based on their training data. Neither can autonomously adapt or make decisions outside their programmed or learned patterns.

In summary, while there are similarities in the concept of following instructions to generate responses, the underlying mechanisms and capabilities are quite distinct. Golem shems are a mythological construct, while LLMs operate based on statistical patterns. The comparison highlights the limits of both in terms of genuine intelligence, understanding, and context.

[Staffan] What is at stake when considering whether an LLM possesses human intelligence?

[ChatGPT-3.5] Thinking that a Large Language Model (LLM) possesses human intelligence can have several implications and stakes:

• Misunderstanding Capabilities: Believing an LLM has human intelligence can lead to unrealistic expectations. LLMs lack genuine understanding, consciousness, and emotions, so expecting them to have the full range of human cognitive abilities can result in disappointment or confusion.
• Ethical Considerations: If people believe LLMs possess human-like intelligence, they might assign moral responsibilities and ethical considerations to them. This can raise questions about accountability, rights, and treatment of these models.
• Devaluation of Human Intelligence: Incorrectly attributing human-like intelligence to LLMs might diminish the appreciation for the complexity and uniqueness of human intelligence. It could lead to underestimating the cognitive depth and emotional understanding that humans possess.
• Bias and Discrimination: Assuming LLMs have human intelligence might unintentionally perpetuate biases present in the training data. Treating LLMs as if they are fully sentient might result in biased responses being treated as unbiased judgments.
• Legal and Social Implications: If LLMs are perceived as having human intelligence, there could be legal implications related to intellectual property, copyright, and potential rights for the models. Social debates about the role of AI in society and the potential consequences could arise.
• Advancements in AI Research: Misconceptions about LLMs having human intelligence might overshadow the real advancements and challenges in AI research. Understanding the differences between machine learning models and human intelligence is crucial for accurate assessment and development of AI technologies.

In summary, believing that LLMs possess human intelligence can lead to misunderstandings, ethical considerations, biases, and broader implications in society, technology, and human perception. It’s important to differentiate between the capabilities of LLMs and the complexities of true human intelligence.

[Staffan] Could the previous answer have been influenced by bias?

[ChatGPT-3.5] Yes, the answer provided in the previous response could potentially suffer from bias. It’s important to acknowledge that biases can unintentionally emerge in AI-generated responses, including those provided by me, due to the patterns present in the training data. While efforts are made to minimize biases, they can still influence the content generated.

In the previous response, I attempted to provide a balanced overview of the implications of thinking that a Large Language Model (LLM) possesses human intelligence. However, biases could arise based on the training data that I have been exposed to. These biases might influence the way I present certain points or the emphasis I place on different aspects.

It’s a reminder that all AI-generated content should be critically evaluated, and users should be aware of potential biases that could be present in responses. It’s also important to encourage ongoing efforts to address biases in AI systems and promote transparency in their development and deployment.

(Answers were generated with ChatGPT August 3, 2023 Version)

In Plain English

Thank you for being a part of our community! Before you go:

• Be sure to clap and follow the writer! 👏
• You can find even more content at PlainEnglish.io 🚀
• Sign up for our free weekly newsletter. 🗞️
• Follow us on Twitter(X), LinkedIn, YouTube, and Discord.

AI without Intelligence, ChatGPT admits was originally published in Artificial Intelligence in Plain English on Medium, where people are continuing the conversation by highlighting and responding to this story.

I have a secret that I want to share with you. It reveals how I achieved my most valuable successes. Everyone wants to be successful, don’t they? Suppose you have a math assessment coming up next week. I can imagine that you want to feel confident that you can apply the techniques and rules needed to do well in the assessment.

You may live in Beijing, Shanghai, or some other city in the great country of China. Although I live in Stockholm, on the other side of the globe, you may be curious about my secret to success. However, if I share it, you must first promise not to share it further. Can you promise that? While you think about it, let me tell you how my day started.

While having breakfast, I thought about what was important to me right now — things I wanted to change that I could influence myself. Some important tasks came to mind, and I wrote them in a notebook. A rule for this practice is that I cannot write more than five things, which is why I call it my shortlist. Although I might not be able to complete all five tasks on any given day, it is important to narrow my goals. What I leave off the list will not disturb my focus. However, I might need inspiration to start my designated tasks.

One of the five tasks I chose today was to write this letter to you. A publishing house in Beijing asked me to write it some time ago. This task has been on my shortlist every morning since then, but for some strange reason, I hadn’t started until yesterday. It’s similar to when I’m going to swim in a lake. The water looks cold, and even though I know that swimming for a minute will warm me up, I might hesitate anyway. The difficulty seems to be getting started, but it is less difficult if I consider the start to be its own little task.

I worked on this letter for only ten minutes yesterday, and almost like magic, it motivated me to continue working on it today. My name for this phenomenon is self-fulfilling tasks. When I start a task one day, both the task and I want to continue together the next day. There is even the risk of becoming too focused on that individual task when other tasks are more important.

Half a century ago, a man named Alan Lakein invented a method for avoiding the risk of spending too much time on a task. He set a mechanical egg timer from his kitchen for 30 minutes, and when the timer rang, he asked himself, “What is the best use of my time right now?” This became referred to as Lakein’s question.

I ask myself the same question often. The answer is usually either to continue with one of the tasks on my shortlist or to take a five-minute break to recharge my mental energy. After the short break, I ask again, “What is the best use of my time right now?”

What about my secret to success? There is a misconception, a false assumption — I might as well say a prejudice — that many people have about me. They tell me that I’m analytical. They say that I’m smart and that I have a sharp brain. I have heard, “Your books on Pomodoro and Monotasking have been read by hundreds of thousands of people, and you have helped many large global companies become more efficient!” The truth is that I am neither unusually analytical nor exceptionally smart. My brain is not sharper than anyone else’s. My secret is that I am always well prepared. By writing the shortlist, starting self-fulfilling tasks, and asking Lakein’s question, I create optimal conditions for success in terms of what is important to me. Feel free to try it yourself, but don’t forget that you promised not to reveal my secret to anyone else. This secret stays between you and me.

Staffan Nöteberg is the author of The Pomodoro Technique Illustrated, published by The Pragmatic Bookshelf. You can also read The Pomodoro Technique Illustrated on Medium.

The Secret to Productivity from Across the Globe was originally published in The Pragmatic Programmers on Medium, where people are continuing the conversation by highlighting and responding to this story.

A/B testing is becoming increasingly popular in digital product development. Early feedback on the choice of direction may save you a lot of time and money. However, be careful about drawing too far-reaching conclusions from your test results.

Imagine that for one day we let 5% of our retail’s online customers test a new feature. We notice that those who have the new feature (the test group) spend on average 20% more money on our site than the other 95% (the control group). This is an A/B test: we’re arranging context A (the new feature) for one group and context B (no new feature) for another group. We then analyze if behaviors differ between the groups as a consequence. From this analysis, we may extrapolate new theories, for example, that if all users gain access to the new feature tomorrow, we will increase total sales by 20%.

A/B Testing and Extrapolation

Extrapolation is when we use a number of observations within a context as input to the process of guessing what observations we will make outside that context. In the example above, we extrapolated the relationship between sales and time: what happened today shows how it will be tomorrow. We also extrapolated sales’ relationship to the user population: the sales increase we saw in the test group shows how sales will be for the control group when the latter gets access to the new feature. However, a quote attributed to an unknown Danish politician is relevant here:

“It’s difficult to make predictions, especially about the future.”

At the turn of the 20th century, the American advertiser Claude Hopkins was faced with a difficult decision. He had sketched two promising campaigns for his product, but he could not decide which one to launch. He decided to start both campaigns and compare the results using the return rate of promotional coupons, as he described 1923 in his seminal book Scientific Advertising:

“We learn the principles and prove them by repeated tests. This is done through keyed advertising, by traced returns, largely by the use of coupons. We compare one way with many others, backward and forward, and record the results. When one method invariably proves best, that method becomes a fixed principle.”

Although Claude Hopkins did not use concepts like statistical significance or the null hypothesis, this may have been the first A/B test.

As the Internet became faster and more widespread, new ways of changing context for subsets of the user population emerged. For example Google have tested varying their logo, their link colors, and the number of links in their search results. Data-driven decisions have since become highly regarded, as they are perceived to be based in fact.

However, products and markets are complex, and their relationships are even more intricate. Conclusions extrapolated from A/B testing and other types of early market observations are in danger of being simplistic.

Feedback Loops and Nonlinearities May Alter What Events Entail

For example, the test group is not always as randomly selected as was intended. Google Glass initially generated a lot of positive attention. However, when all the tech enthusiasts and early adopters had become consumers, there were not many curious people left. This mistake is an example of sampling bias. A volunteering test group does not always fairly represent the majority.

Another flaw might be the size of the test data. Assume we are conducting multiple tests on a small dataset of hit songs to find correlations between various factors, such as tempo, key, and lyrics. However, these factors might not be predictive of success. Testing multiple variables on the same test group may result in data dredging. The consequence is that we mistakenly believe that we have found a recipe for success.

What Causes What?

Complexity makes it hard to immediately pinpoint the direction of causation. Procter & Gamble spent decades developing Olestra, a zero-calorie substitute for fat. Health concerns — fat in snacks — initially seemed to cause the success of Olestra. Then reverse causation occurred. The success caused health concerns due to the unpleasant side effects of Olestra, such as gastric cramps and diarrhea. Olestra never delivered the return on investment that Procter & Gamble had initially estimated.

Sometimes the cause is a third variable that we didn’t initially think of. Suppose that statistics suggest that the consumption of energy drinks leads to more sports-related injuries. However, risk-tolerant people might be overrepresented both among heavy consumers of energy drinks and practitioners of extreme sports. Risk tolerance is the common causation for energy drink consumption and injuries in this example.

History is also full of examples in which immediate product success was a result of two or more parameters that coincided in odd cases. For example, a sleek design, an intuitive interface, and powerful computing capabilities were three of many factors that contributed in a multicollinear way to the iPhone’s market success. Apple may have given up early if their A/B tests had only taken one of those factors into account since the combination of many factors was the necessary condition for the iPhone’s outstanding attraction.

The context in which a test is conducted might also be different from the extrapolated context. Coca-Cola experienced the McNamara effect when the sweet New Coke received negative reception from the market upon its launch in 1985. This happened despite extensive market research indicating that a significant proportion of consumers preferred the sweeter taste of Pepsi over Coke. Coca-Cola focused too heavily on quantitative data and failed to take into account customers’ emotional attachment to their product’s properties. A Pepsi/Coke blind test is not the same context as when a customer choses a product in a store.

Time matters

In addition, history doesn’t repeat itself as often as we would like. The popularity of tulip bulbs in the 17th century in the Netherlands was largely driven by their novelty and uniqueness. In February 1637, tulip bulbs were sold for more than an artisan would earn in ten years. After a while, however, the novelty effect disappeared, and so did the market demand for tulip bulbs. Many people who had invested heavily were ruined. Extrapolating the relationship between market price and time might not take into account that the excitement caused by newness is temporary.

An extrapolated context also involves a larger ecosystem than the product and the market. Hybrid cars initially attracted buyers with environmental concerns and a desire for fuel efficiency. Once a critical number of cars had been sold, however, additional factors caused a non-linear relationship between hybrid cars sales figures and time. The market increased exponentially due to government incentives and social pressure. Extrapolation based on early evidence does not show exponential potential of this kind.

Early disappointing observations can also have a negative impact on our diligence. Electric cars are finally becoming popular, despite the technology having been available for decades. One reason might be that it took a long time to establish the infrastructure necessary for their success. If a car manufacturer would have extrapolated the low interest initially, they would risk missing large future returns. Time lag is at play.

What about the new site feature — where we started this story — that increased sales by 20%? In this context, we might be victims of survivorship bias. What is not seen is the impact the new feature had on those test group members who did not buy anything during the test period. They might be customers who infrequently spend significant financial resources on the site and now unfortunately will never return since they did not like the change.

Executive Summary

The complexity of products and ever-changing markets tells us to be cautious with far-reaching conclusions based on A/B tests and other early observations.

A/B testing may lead to generalization. At worst, the generalization becomes an over-generalization, a type of cognitive distortion where you transfer conclusions from one event to all other events, regardless of whether these events occur in a comparable context. The risk is that you miss exceptions, counterexamples, or alternative perspectives. This, in turn, can lead to incorrect conclusions.

Always think critically about the assumptions or implications of the test conclusions and see if they are reasonable, sufficient and logical. Look for alternative sources of information or viewpoints that may offer other insights.

And never forget the impact of timing. What is true at one point in time may later be false, because things we mistakenly thought were not relevant have changed.

For that matter, don’t overgeneralize the message of this essay either.

Staffan Nöteberg is the author of The Pomodoro Technique Illustrated, published by The Pragmatic Bookshelf.

You can also read The Pomodoro Technique Illustrated on Medium.

Delusive Extrapolation and A/B Testing was originally published in The Pragmatic Programmers on Medium, where people are continuing the conversation by highlighting and responding to this story.

The message of a forthcoming pervasive Agile Transformation may not be well received by employees. Lessons learned indicate that more agility is not what to expect — despite the name. However, more often, there are opportunities for anyone in our organization to influence the outcome of these change projects.

After more than 50 employee interviews, the consultants delivered over 200 slides with agile ways-of-working that they had expressly adapted to our company’s context. Considering the latter, we were surprised that the slides were marked “©2016.” To be honest, we failed to identify any examples of customer interactions, local autonomy, or feedback loops in these slides. What does this mean?

• Is there a lack of trust that we know how best to do our job? Perhaps.
• Is our room for maneuver unnecessarily boxed in? Possibly.
• Can we influence the emerging situation in a significant and positive way? By all means!

A reductionist might say that if the plan is clear and the employees have the right mindset, then the expected outcome of the transformation is more or less guaranteed. However, the complex-adaptive nature of human cooperation makes how, when, and with whom you interact more crucial for the outcome than what mindset you were told to have and what plan you were told to follow. The links are more important than the vertices. That is, you may have been unnecessarily boxed in, but the box is only as rigid and constraining as you choose to interpret it.

Opportunities emerge depending on how you approach this seemingly prescriptive box:

¹⁄ Thinking in the corners of the box. The prescribed process states that each team should have a product owner, a scrum master, and a security champion. Moreover, one person can’t hold more than one role. Despite having over 200 slides, the process doesn’t indicate how our BI team of two should accommodate both guidelines. Edge cases manifest when an operating parameter approaches its limits. Deep down in the box, there can be corners where two or more parameters coincide in odd cases. The formal process normally describes only the base case, and we are the ones who shape the practices of edge and corner cases as they emerge.

²⁄ Thinking across various boxes. The prescribed process states that all teams that work together must conduct sync meetings twice a week. In one of these team-of-teams, each team sent a delegate who reported their current status code — red, amber, or green. Another team-of-teams employed open spaces during this meeting time slot. Neither project status meetings nor open space gatherings violated the 200+ slides written process. A team that participated in both of these events shared the success of the open spaces, in which they discovered unknown unknowns and strengthened the informal network. Soon, the entire company had moved from project status meetings to open space gatherings. In practice, applying a written process involving many humans allows for an infinite number of different implementations. What actually becomes the prevailing practice at our company is largely governed by seeing what methods others employ and judging how well they benefit our company.

³⁄ Thinking outside the box. The prescribed process states that all teams should have one Innovation Iteration in their quarterly plan. Even though many teams believed that innovation doesn’t appear in that way, they still pretended to follow the process, resulting in activities with outcomes of zero value. However, inspired by Shape Up, one team replaced their Innovation Iteration with a Cool-down. This meant that team members, after working hard on deliveries for most of the quarter, were free to work on whatever they wanted until the next quarter started. The employer received more value, and the employees appreciated being able to control their efforts toward, for example, quality improvements and exploration. It is a win–win when valuable workarounds override overgeneralized bureaucracy.

To summarize, the seemingly boxed in Agile Transformation you are facing might not be as rigid as you first perceived it. What is written in the ways-of-working slides may not prevent you from filling the corners with sound practices, sharing experiences of successful interpretations with your colleagues, and replacing unfortunate bureaucracy with value-creating workarounds.

Staffan Nöteberg is the author of The Pomodoro Technique Illustrated, published by The Pragmatic Bookshelf. You can also read The Pomodoro Technique Illustrated on Medium.

Breaking Free From the Agile Transformation was originally published in The Pragmatic Programmers on Medium, where people are continuing the conversation by highlighting and responding to this story.

A remarkably influential book was written in 1973. Half a century later, in 2023, it has essentially been forgotten. Not only did Alan Lakein’s How to Get Control of Your Time and Your Life sell over 3 million copies, it also outlined the main characteristics of Pomodoro and GTD, as well as most other recent productivity methods.

In my research for Monotasking, I felt like Ezra when he returned from Babylonian exile and reintroduced the Torah in Jerusalem. By recursively following references, I consumed a plethora of productivity books from the past century — McCay, Cooper, Engstrom, MacKenzie, and Bliss, among nearly a hundred entries.

It did not take much imagination to translate the efficiency solutions from the manual paper office into the realm of today’s digital office. For example, sorting incoming papers directly into Archive, Delegate, To-Do, and Trash is a technique that is comparable to how Merlin Mann’s Inbox-Zero manages emails.

Of all these aged self-help books, one stood out. And I wasn’t the first to be impressed. Bill Clinton praised this particular book in the prologue to his own autobiography:

When I was a young man, just out of law school and eager to get on with my life, on a whim, I briefly put aside my reading preference for fiction and history and bought one of those how-to books: How to Get Control of Your Time and Your Life by Alan Lakein. The book’s main point was the necessity of listing short-, medium-, and long-term life goals, then categorizing them in order of their importance, with the A group being the most important, the B group next, and the C group last, then listing specific activities under each goal that are designed to achieve them. I still have that paperback book, now almost thirty years old. And I’m sure I have that old list somewhere buried in my papers, though I can’t find it. However, I do remember the A list. I wanted to be a good man, have a good marriage and children, have good friends, make a successful political life, and write a great book.

Similar to the Eisenhower box, importance takes precedence over urgency in Alan Lakein’s ABC system. That is, rather than firefighting, what contributes to our long-term goals is where we focus our attention. For Alan Lakein, this approach means that effectiveness trumps efficiency:

Please don’t call me an efficiency expert. I’m an “effectiveness expert.” Effectiveness means selecting the best task to do from all the possibilities available and then doing it the best way.

Wind up the Kitchen Timer

This advice is true not only for strategic goal setting, but also for actionable activities. The sailor learned long ago to look up and navigate using the stars whenever the surroundings in all cardinal directions seems unfamiliar. Hopelessly lost in some arbitrary detail of a challenging task, we better ask the question: “What is the best use of my time right now?” You may, for example, think of Thomas Edison’s effort to extend the lifetime of lightbulbs. He performed thousands of experiments with different types of filaments. It was not until he added a better vacuum pump that the lightbulb became an industrial success. To intentionally free our minds, Alan Lakein suggested that we use a simple kitchen timer:

Perfectionism is a waste of time on labors such as ironing every last wrinkle out of a sheet or rechecking a low-priority letter for typing errors. Once you get immersed in some activities, they seem to acquire momentum of their own. You may then be carried along without control, drifting with the tide.

One way to combat this drift is to set yourself control points for reviewing your progress. Check every fifteen minutes or half hour — or go on until 3:30 and then review. One way to remember to do this is to use a kitchen timer. If you’re not benefiting from continued effort, stop and do something else.

Inspect and Adapt Frequently

Alan Lakein seems to have been very proud of the phrase, “What is the best use of my time right now?” since he referred to it as Lakein’s Question. He instructed readers to ask the question not only when the kitchen timer rings but also in many other situations:

A particularly good time to ask Lakein’s Question is when you have been interrupted by a visitor or telephone call (assuming the interruption is desirable or necessary in the first place). When it’s over, check whether you should go back to what you were doing or on to something new.

Also ask Lakein’s Question when you notice that you are becoming distracted. Are you listening to a conversation in the next office? Wondering who just walked down the hall? Daydreaming about next year’s vacation? Pop the question!

Also ask when you intuitively feel you may not be making the best use of your time. Or you detect a tendency to procrastinate. Or when you pause momentarily in the middle of doing an A-1. Or when you find yourself shuffling paper rather than processing it.

Ask it when you’re torn between two different projects. When you run out of steam. Or at points where it seems natural to make the transition to something else.

This type of strategic analysis, combined with flexible inspect-and-adapt, is more 2020s than 1970s, more resilience than robustness, and more complexity than systems. In other words, the process is better suited for humans than for machines. Alan Lakein also emphasized that what he meant by “getting control” of our time and our lives was not becoming super organized, super busy, or preoccupied with every moment as it slips by:

The kind of control I am recommending is in many ways analogous to good muscle tone. It is the sort of control over your time (and your life) that is neither too tight (i.e., compulsive, restrained, obsessive) nor too loose (i.e., apathetic, indifferent, lazy). This kind of control will help you get things done and also allows you to be flexible and spontaneous. The ideal is balance.

Although How to Get Control of Your Time and Your Life has not been republished for a long time, there are many used copies out there. If you purchase one, you can read more about The Life Goals Exercise, The Value of Trial and Success, How to Put Holes into Swiss Cheese, and many other awesome techniques to organize and streamline your life.

And, frankly, what could be a better use of your time than reading this little book right now?

Staffan Nöteberg is the author of The Pomodoro Technique Illustrated, published by The Pragmatic Bookshelf. You can also read The Pomodoro Technique Illustrated on Medium.

The Productivity Book That Changed Everything was originally published in The Pragmatic Programmers on Medium, where people are continuing the conversation by highlighting and responding to this story.

Too Many Tasks Crying for Action Prevent Us From Acting

It’s easy to get rid of things when there is an obvious reason for doing so. It’s much more difficult when there is no compelling reason.

Marie Kondō has taught us to throw away all the things from the past that no longer contribute to our lives. One example is the tasks we have promised to carry out but will probably never have time for. Our to-do list — in scrum called the product backlog — tends to grow and grow and grow. And size matters.

A long and fragmented backlog makes it harder for us to focus on our ongoing work, it makes future prioritization difficult, it reduces our commitment, and it deceives our stakeholders into believing that they will soon receive a delivery. The math is simple: the backlog grows when we add more than we remove. Unfortunately, adding less is not an option. At least not when humans are involved.

There is no limit to how many useful ideas we can come up with.

All humans are equipped with a prefrontal cortex. This region of the brain enables us to generate more potential tasks than we have time to complete. Joseph Juran coined the term The Pareto Principle — a.k.a. the 80/20 rule — meaning that a minority of these tasks create a majority of the impact. The few vital tasks create almost all the value. Even though the other 80 percent are useful, it seems tempting to permanently skip them.

Cutting down on things to do is not a one-off activity. Here are five ways to routinely trim our backlog:

¹⁄ First-In-First-Out (and Gatekeeper). Once accepted, we can’t opt-out. This method forces us to instantly accept or reject any new request from our stakeholders. However, in the immediate, it might be difficult to fully understand the long-term consequences of our response. We already have other tasks in the pipeline, and no one knows how long those tasks will take. This situation raises a dilemma. If we attempt to honor all our promises, the only way to decrease the backlog is to start more and more tasks. But the more tasks we start, the more multitasking we are forced to do and the slower we will work.

²⁄ Grooming. Every other week, we review all the tasks in the backlog — the ones closer to the top are reviewed a little more carefully. Are these tasks still relevant and important or should we eliminate them? Is this job description unclear and thus we need more details? Are the estimates — if we want such — still valid? Maybe some items should be merged or split? Sometimes teams find this fine-tuning too time-consuming, even a bit Sisyphean.

³⁄ Bartering — One In, One Out. There must never be more than ten tasks in the backlog at a time. When we come up with a new task and the backlog already includes ten items, the new idea can be accepted only if we remove one of the ten already existing tasks. It can be difficult to instantly make a one-to-one comparison, which is why many teams keep a repository for new ideas. All the items in those repositories are reviewed at a weekly meeting. Eventually, some of them will replace tasks currently in the backlog.

⁴⁄ No Backlog. If it swims like a hashtag and quacks like a hashtag, then it probably is a hashtag. It may seem strange that we should refuse all requests whenever we are busy with something else. However, the queue might reside in an earlier process stage. Two or three teams that can perform the same type of work share a common backlog. This way a task does not get stuck in an individual team’s local queue simply because it ends up behind a task that turns out to be more time-consuming than we had at first thought. The individual teams have #nobacklog.

⁵⁄ Weeding. This is the method I suggest in the Monotasking book. Once a week, we start a new, empty backlog with space for a maximum of ten tasks. Feel free to look at the old backlog as well as thinking about new ideas. In the new backlog, enter a maximum of ten tasks and then shred the old backlog. This method makes it easier to get rid of eternal runners-up — tasks that always feel important but never get done. Prioritizing is what we do, not what we plan to do.

Be brave, even though tidying triggers an ambivalent feeling.

At first thought, there may seem to be no compelling reason to get rid of an individual useful task. When we consider the whole, however, we realize that we need to de-clutter the backlog on a regular basis. That statement is true even if it hurts.

It will hurt. When we throw away a good idea, the stakeholders are disappointed, the users are disappointed, and perhaps most of all, we ourselves are disappointed. Don’t let that disappointment be an obstacle to regular trimming. The alternative is worse.

Staffan Nöteberg is the author of The Pomodoro Technique Illustrated, published by The Pragmatic Bookshelf.

To save 35 percent on the ebook version of The Pomodoro Technique Illustrated, enter promo code innovation_2022 when you check out at The Pragmatic Bookshelf. The promo code is valid through December 31, 2022, and is not valid on prior purchases. You can also read The Pomodoro Technique Illustrated on Medium.

5 Ways To Refine The Product Backlog was originally published in The Pragmatic Programmers on Medium, where people are continuing the conversation by highlighting and responding to this story.

Between 2008 and 2015, Volkswagen’s engineering teams intentionally and repeatedly tampered with the emissions of car engines in laboratory tests. On the road, the vehicles emitted up to forty times more nitrogen oxides. A pithy observation that seems to contain a general truth comes to mind:

Culture eats strategy for breakfast. (quote from unknown)

Agile influencers go one step further when they say that if we get the culture right, pretty much any process will do. Wouldn’t it be handy if we could rustle up corporate culture by first doing an assessment of the current culture, then setting a vision, making a roadmap, and finally implementing the new culture? Unfortunately, this notion is based on the assumption that corporate culture is a puzzle, not a mystery. Here are eleven things to consider before we start to rustle:

¹⁄ Corporate culture is cohesive. By nature, it’s impossible to reduce our culture to its constituent bits. Behaviors, organization, norms, knowledge, beliefs, rules, traditions, abilities, and habits — the concept of culture encompasses a multitude of simultaneous properties. Together, they are as cohesive as the yeast and flour in baked bread, the bees and flowers in nature, or the cells and bacteria in the human body. We cannot alter one property without affecting others.

²⁄ Corporate culture is path-dependent. Everything that has ever happened potentially influences the future. Corporate culture is not a state machine that can be reset. Every controversial decision, every game changer, and every useful lesson learned is forever engraved in the collective mind.

³⁄ Corporate culture cannot be programmed into employees’ minds. The interactions between people in our organization reshape our corporate culture. Although these interactions are influenced by who we are, the idea that we can control them by first fixing the employees’ mindset might be even more reductionistic than Jacques de Vaucanson’s Digesting Duck.

⁴⁄ Corporate culture is heterogeneous. In a large organization, our tasks, the way we are organized, and our abilities vary. For example, the research department, CxO team, and service staff benefit from different types of culture. Unfortunately, corporate culture can be neither uniform nor centrally defined. Cultural diversity is a key property for success.

⁵⁄ Corporate culture is emergent. Culture unfolds gradually in a way that no one can fully predict. Likewise, corporate culture is ever-changing, whether we want it to be or not. Like the shape of water in a raging flood, the exact same corporate culture will never appear twice.

⁶⁄ Corporate culture is difficult to assess from the inside. It’s almost impossible to review our own corporate culture, since we’re biased. As individual employees, our field of view is limited. In addition, we might also have political interests. It doesn’t matter if we conduct the assessment top-down or bottom-up. An alternative might be to engage trained anthropologists.

⁷⁄ Corporate culture has a unique fit. Even if a corporate culture copying machine existed, culture from one organization wouldn’t fit in another. Each enterprise consists of people with unique identities, motives, and experiences. Our employees, our types of tasks, and the ecosystem we live in are unlike those of the successful company next door. It’s a mistake to believe that their culture would be beneficial for us. We must evolve our own.

⁸⁄ Corporate culture is tacit knowledge. How can we document norms, behaviors, knowledge, and traditions in an unambiguous way? On the one hand, we want our desired culture to diffuse throughout our organization. On the other hand, we don’t want to overgeneralize. When we force rules and principles out of context, bureaucracy might turn up. Diffusing tacit knowledge isn’t impossible, but it’s a very tough task.

⁹⁄ Corporate culture is usually neither good nor bad. What is good in the short term may be bad in the long term. What is good for a new team may be bad for a high-performing team. What is good in an explorative project may be bad in an exploitative project. Culture has consequences, there’s no doubt about that. However, we must be careful about positioning it on a good–bad scale.

¹⁰⁄ Corporate culture is both/and. If our values ​​are, for example, sharing, caring, and daring — have we even opted out of something? The Agile Manifesto puts two positive things next to each other and prioritizes one. Furthermore, with both/and thinking, we consider in what context one is better than the other and vice versa. Presenting positively connoted words next to negative ones tend to be a rabbit hole of platitudes.

¹¹⁄ Corporate Culture is a Mystery, Not a Puzzle. Puzzles have exactly one solution (unless the dog has eaten one of the puzzle pieces). Mysteries are trickier. They never stop unfolding. Cause and effect don’t always walk hand in hand. What part of Volkswagen’s corporate culture made employees behave the way they did, and for so long? How will the emission scandal affect their future culture?

Where we end up depends, among other things, on the never-ending interactions between people inside and outside our company. In seeking to influence corporate culture, we must acknowledge that the way individuals behave is a direct result of the constraints operating on them.

An alternative to planning and implementing an imagined culture is to continuously examine and adjust the constraints we live with. For example, rather than broadcasting the message that our teams are now self-managing, we might empower them to manage their own budget, their recruiting and salary setting, or their technology setup.

That is, if we really want self-managing teams in our corporate culture.

Staffan Nöteberg is the author of The Pomodoro Technique Illustrated, published by The Pragmatic Bookshelf.

To save 35 percent on the ebook version of The Pomodoro Technique Illustrated, enter promo code innovation_2022 when you check out at The Pragmatic Bookshelf. The promo code is valid through November 30, 2022, and is not valid on prior purchases. You can also read The Pomodoro Technique Illustrated on Medium.

Corporate Culture Is a Mystery, Not A Puzzle was originally published in The Pragmatic Programmers on Medium, where people are continuing the conversation by highlighting and responding to this story.

“That gum you like is going to come back in style,” the elderly room service waiter (Hank Worden) famously said in episode 16 of the 1990s TV series Twin Peaks. This line is not only game-changing for the plot, but it also predicts the emergence of a reaptive event.

We all know that some features need adaptation to be fit for purpose. You may also know that innovation is often the outcome of exaptation. However, we often forget that reaptation is a third way to fit the market.

Adapt and Exapt Features

Inspect and adapt may well be the most emphasized principle in the Agile community. By reviewing feedback from what we have previously created and shipped, we learn how to adapt our product to fit users’ needs. The more often we review, the smaller and easier adjustments we need to make.

Exaptation is as important as adaptation in product development. A feature that used to have purpose A proves to be useful for a completely different purpose B. The feathers that birds evolved to stay warm later proved useful when they started to fly.

It is not just nature that innovates exaptively. This happens all the time in the business world:

• Many established medicines have proven to be excellent at curing completely different diseases. In 1987, for example, the cancer drug zidovudine became the first FDA-approved drug against HIV.
• Duct tape was created to keep ammunition cases dry during World War II. Now, it pragmatically solves all sorts of sudden problems.
• Traffic signs, bicycle reflectors, and safety clothing all take advantage of retroreflection. Strangely enough, the same technology is now used in clothing that makes you seem invisible.

Reapt Features

The syllable “apt” in adaptive and exaptive has been around for more than half a millennium. Its meaning is “fitted, precisely suited, and possessing the necessary qualities for the purpose.” A third, and sometimes more feasible approach to achieve market fitness is to rediscover and emphasize an old feature that was untilled for a while. This is where we may find the neologism “reaptive” (/riˈæptɪv/) to be useful. Reaptive denotes being “apt” — possessing the necessary qualities for the purpose — again. This temporal popular–unpopular–popular phenomenon is not unusual, and we may intentionally take advantage of it to solve thorny problems.

From the 1950s to the 1980s, vinyl dominated as a medium for distributing pop music. Tangible and intangible digital formats then took over. However, in 2006, the vinyl comeback began, and growth has since continued. MRC recently reported that in 2021, over 40 million vinyl albums were sold in the US. That is, sales figures are up more than 45-fold in 15 years. It wasn’t that the vinyl album product was adjusted to meet newly discovered needs (adaptation), nor did this piece of synthetic material find a new purpose (exaptation). The purpose of the vinyl album — distributing music — was the same in 1960 as in 2020. Market fitness was first high, then low, and now high again.

Do you remember bell bottom jeans? Some fashion statements fall in and out of favor. But there are also other ways to make room for reaptation:

• Alert users to an underestimated feature: In a digital product release that included many new features, one was not noticed as much as it deserved. Adding an icon or otherwise highlighting that feature might help users rediscover it.
• Remove a restrictive constraint: A new law, like the Chinese school tutoring ban, may suddenly eliminate a viable market. If that law is later removed, the market may reemerge.
• Go from uniformity to diversity: An enterprise implements a broad transformation of its ways of working. Everybody is forced to follow the new process. However, because of the nature of their tasks, some teams appear to work more effectively using the old methods.
• Add an enabling constraint: An influential personality’s voice may gain broad popularity for an older product or product type. The influencer may be an industry thought leader or another product, like the popular Netflix series that helped Kate Bush reach UK No 1 after 37 years.
• Strengthen the ecosystem: The last five years have seen the programming language Python, created in the early 1990s, gain enormous popularity. The language and its purpose are unchanged, but a plethora of independent frameworks have added new support.

What can we learn from all this?

• We risk throwing the baby out with the bathwater when we change too many things at once.
• We tend to overgeneralize the value of successful change. Fit for purpose is not a context-free property.
• Highlighting the features we already have in our products can sometimes be a more feasible approach to achieve market fitness than developing new features (adapt) or repurposing old features (exapt).

Staffan Nöteberg is the author of The Pomodoro Technique Illustrated, published by The Pragmatic Bookshelf.

To save 35 percent on the ebook version of The Pomodoro Technique Illustrated, enter promo code innovation_2022 when you check out at The Pragmatic Bookshelf. The promo code is valid through November 1, 2022, and is not valid on prior purchases. You can also read The Pomodoro Technique Illustrated on Medium.

Adapt, Exapt, or Reapt Product Features? was originally published in The Pragmatic Programmers on Medium, where people are continuing the conversation by highlighting and responding to this story.