Author: Jurgen Appelo

In the Agile community, there’s a common misconception that agility exclusively involves iterative, incremental delivery, and anything else is dismissed as “waterfall.” However, my Agile Quadrants picture shows a broader spectrum of agility. While not strictly adhering to “hard agile” principles, many products employ “soft agile” approaches, involving feedback cycles through EITHER iterative experiments OR frequent delivery (but not both). And that’s perfectly okay.

Who decides what is Agile?

Agility is a widely discussed and debated concept, but what truly defines it? Here are three questions for you to think about:

• Consider Apple’s strategy in developing the iPhone. The company is known for conducting numerous iterative experiments internally, yet it holds back from releasing incremental updates until the final product launch. Is Apple’s approach really agile?

• Toyota, renowned for its Toyota Production System, efficiently reduced batch sizes and ensured frequent delivery of their cars. However, Lean Manufacturing is all about the production process, not the design process. Was Toyota’s approach really agile?

• On a more personal level, I am working on a novel, now in its seventh iteration, with the book undergoing constant refinement until its final release (in June 2024). I intend to keep the entire manuscript in progress until one final launch date. Am I really agile?

To answer these questions, you must understand the difference between iterations and increments. This will help you better grasp the concept of agility and distinguish between ‘hard agility’ and ‘soft agility.’

Plus, it will make you a better agilist.

Iterations versus Increments

Some people argue that publishing a novel only on its final release date isn’t agile. They label it as “waterfall” or “monolithic.” However, I see things differently. Not all endeavors benefit from incremental releases, and not all products require iterative experimentation, but that doesn’t mean they can’t be agile.

Experimentalism (Iterate!)

Iterative discovery and development is a method where parts of a product are consistently revised and enhanced through experimentation. This approach teaches a team the most effective ways to create and deliver a great product. The primary purpose of iterative cycles is to gain insights through continuous testing. The insights are usually derived from direct experience, input from other colleagues, and feedback from representatives (proxies) of the customer and target users. By conducting numerous experiments, the team lowers the risk of product failure upon its release into the actual market. Think mock-ups, demos, prototypes, Minimum Viable Products, and beta testing.

Another advantage of the iterative process is the potential for continuous enhancement, leading to increased value. This is particularly true in the entertainment industry, where higher quality often translates to many more users and, consequently, greater revenue.

Design Thinking and Lean Startup are two methodologies that advocate for this iterative and experimental approach in new product design, emphasizing learning and adaptation through continuous feedback cycles.

“Experimentalism is the philosophical belief that the way to truth is through experiments and empiricism.”

Managerialism (Plan!)

What is the antithesis of an iterative, experimental approach in product design? It’s likely a process where people assume they have all the answers already. In such a setting, the prevalent mindset is “Let’s get this done!” rather than “Let’s experiment and learn!” This approach is suitable in situations where all variables are known and predictable. However, it can often lead to what’s known as the Ivory Tower Syndrome, where there’s a false belief that one has complete knowledge and experience.

We could label this approach as Managerialism, defined as relying on professional managers and established strategies. With this way of thinking, the focus is on achieving planned outcomes and delivering results, sidestepping the need for experimentation. Think strategic off-sites, project checklists, milestones, roadmaps, and Gantt charts. (If there’s a more fitting term for this, I’m open to suggestions!)

“Managerialism is the reliance on professional managers and organizational strategies to run an organization.” - Wikipedia

Incrementalism (Deliver!)

Incremental development is a method where product parts are progressively created and released to users. This strategy aims to provide value to users more quickly and to incorporate their feedback into further development.

The main objective of incremental delivery is to meet customer needs faster. In areas like software development, this approach enhances the user experience by allowing users to engage with and benefit from the product even when it’s not fully complete. Think velocity, throughput, lead time, limited Work in Progress, and ‘potentially shippable products.’

Another key advantage of this method is that it exposes the product to its actual operational environment early on. Feedback from real users and infrastructure informs ongoing improvements and adaptations. This is particularly crucial in ever-changing business landscapes, where real-world input is more valuable than proxies typically used during iterative-only discovery and development.

Lean Manufacturing and The Toyota Production System are methodologies that embrace this incremental approach, emphasizing fast responsiveness to real-world feedback in product delivery.

“Incrementalism is a method of working by adding to a project using many small incremental changes instead of a few large jumps.” - Wikipedia

Culminationism (Launch!)

What is the opposite of an incremental approach to product delivery? Essentially, it’s about releasing a product in one grand event: a single, comprehensive delivery where all of its value is presented to the customer at once. Think launch dates, publication dates, opening nights, and release parties.

Consider the absurd idea of asking a moviegoer to visit the cinema twenty-five times to see the next James Bond film in five-minute segments. It’s equally nonsensical to deliver one-third of an opera house and then expect a performance of the first act of an opera. Similarly, a train tunnel that’s 99% complete is still not a tunnel but merely a long and expensive dead end. And the organizers of a big surprise party miss the point when the party is broken down into a series of many rather small surprises.

Big releases are often the case with products in construction and entertainment: such products only hold value for the customers once they are fully completed and presented as one experience.

I propose the term ‘culminationism’ to describe this approach, where the product’s value is maximized by delivering it as a single culminating event. (Of course, if you have a better term, I’m open to suggestions!)

“Culminationism is the method of working toward a big launch or one final conclusion rather than delivering incremental changes.”

The Agile Quadrants

I wish people stopped shouting, “You’re doing it wrong,” when they see an approach without iterations or increments. The smarter thing to do would be to evaluate the product’s context and desired customer experience first.

In some domains, there is only value when the entire project is finished and not a second earlier. You cannot shoot bits and pieces of a satellite into space in weekly sprints and expect those increments to be valuable. Alternatively, in other domains, experiments have little benefit. When you run a coffee bar, it’s probably best to start serving coffee and begin listening to customer feedback—no need for iterations.

Typical Quadrant (1)

When planning a trip to familiar destinations like Paris or London, my approach is straightforward: book train tickets and a hotel, plan travel times and pack my suitcase. Since I’ve done this before, I’m dealing with “known knowns” – there’s no need for iterative experiments, and I can’t break down the trip into smaller increments; it’s a singular journey.

In the ‘Typical’ quadrant, projects and products require minimal iteration and few, if any, increments. Take IKEA as an example. As of September 2023, there were 462 IKEA stores globally, which are almost identical. If you’ve visited one, you know them all. When IKEA plans its 470th store, the company likely won’t focus much on experiments and increments. Instead, they rely on project managers, milestones, Gantt charts, and a well-versed organization for execution.

I label this quadrant ‘Typical’ because many of our routine activities fall into this category. From grocery shopping and cooking dinner to planning vacations and organizing game nights, most of us are adept at managing tasks in familiar environments.

Note: While some of you might refer to this as the ‘Waterfall’ quadrant, I see traditional waterfall projects as a subset within. These projects represent a larger, more professional version of ‘typical’ tasks, characterized by a lack of experimentation and no incremental delivery.

Serial Quadrant (2)

During the COVID-19 pandemic, while painting my house, I experienced the advantages of an incremental approach firsthand. Completing each wall and room led to immediate improvements in our quality of living. This process also allowed me to learn and improve my painting skills through an ongoing feedback cycle, as each completed section provided value to its actual users - me and my partner.

The ‘Serial’ quadrant is suited for projects and products that deliver value in numerous small increments but don’t necessarily require iteration. A good example of this is public transportation. Bus and train services are offered regularly, and their schedules are frequently updated based on user feedback. While some iterative exploration may involve computer modeling, the most valuable insights come from just starting with a bus or train schedule and receiving real-time feedback from travelers.

I’ve named this quadrant ‘Serial’ because it involves breaking down the total value of a product into a sequence of smaller deliveries. Each of these increments can be individually consumed and appreciated, providing ongoing value without iterative experimentation.

Original Quadrant (3)

The process of writing my novel is distinctly different from my house painting project. It’s a project steeped in continuous iteration and experimentation. I’ve worked with structural editors, development editors, and soon, a copy editor. The book has evolved through feedback from friends, beta readers, and, eventually, proofreaders before its final release. Unlike my painting project, this one doesn’t involve increments; the readers will need to wait until the book is entirely completed, likely marked by a launch party.

The ‘Original’ quadrant encompasses projects and products that cannot or should not be divided into small increments. This could be due to logistical challenges, as with constructing skyscrapers, or to preserve the customer experience, such as in feature films. Apple’s strategy of organizing significant launch events for its major products is a prime example. The build-up of anticipation and marketing before the launch only enhances the customer experience. Likewise, the development of computer games usually happens in the third quadrant, not the fourth, as their environments are more similar to motion pictures than business software.

I’ve termed this quadrant ‘Original’ because it typically involves unique, never-before-seen works that benefit from a single, impactful delivery. There is only one opening night of a musical, and it better be great, or else word will get out that it’s not worth seeing.

Agile Quadrant (4)

In the various startups I’ve participated in, we aimed to adopt an agile approach characterized by iterations and increments. We consistently worked to mitigate risks and enhance quality through iterative experimentation while regularly releasing new product versions to actual customers. The incremental approach allowed us to deliver value sooner and gain insights from real user experiences. In a startup environment with limited resources, the last thing you want is a culminationist, managerial approach.

The ‘Agile’ quadrant is tailored for projects and products where both experimental learning and rapid feedback from incremental delivery are essential. Methodologies like Scrum and the principles of the Agile Manifesto merge these two approaches, with iterative experimentation happening through the release of product increments to gauge customer feedback. In “pure Agile,” iterations and increments are practically the same.

I refer to this quadrant as ‘Agile’ in honor of the Agile Manifesto, authored twenty-two years ago by seventeen signatories. However, it’s crucial to recognize that they wrote the manifesto specifically for software development projects, which almost invariably benefit from iterative and incremental methods.

Six Transitions

Now that I have described the two dimensions and four quadrants, we can better understand the journeys that teams worldwide might go through in their attempts to “be more agile.” Looking at the big picture, we can identify six common transitions:

Typical to Serial (A)

The shift from the ‘Typical’ to the ‘Serial’ approach underpins the Toyota Production System and Lean Manufacturing. It represents the move towards incremental delivery as opposed to the traditional method of Big Design Upfront (BDUF).

Toyota revolutionized car production by reducing batch sizes with smaller inventories, enabling the company to accelerate product delivery and respond more swiftly to actual customer feedback. However, Lean Manufacturing had little to say about the design of cars, which could still happen in either the Typical or Original quadrants.

Typical to Original (B)

Moving from the ‘Typical’ to the ‘Original’ approach aims to dismantle the Ivory Tower Syndrome. This shift emphasizes replacing Big Design Upfront with iterative experimentation.

The fundamental change here is for managers and designers to abandon the notion that they fully understand customer desires from the outset. Instead, they should pursue ongoing validation of their assumptions through cycles of experimental feedback. A notable example of this transformation is LEGO’s journey from a command-and-control product design model to embracing open innovation. This shift involved engaging with user groups and fan sites, allowing LEGO to incorporate direct user input into their design process (see the book Brick by Brick).

Serial to Agile (C)

Moving from the ‘Serial’ to the ‘Agile’ approach involves incorporating discovery processes into the existing framework of serial product delivery. This transition is characterized by introducing iterative experimentation alongside the already established planned incremental delivery.

An illustrative example of this shift might be seen in a typical coffee bar that introduces customer workshops and ideation sessions to involve coffee lovers in beverage experiments, potentially leading to more innovative additions to its menu.

Original to Agile (D)

Shifting from the ‘Original’ to the ‘Agile’ model involves breaking down a project or product, which is inherently iterative and experimental, into smaller segments that deliver value to customers more frequently.

For example, in the construction industry, some companies build series of tiny homes rather than erect large apartment buildings. These homes are sold and occupied individually, creating a feedback loop that facilitates continuous improvement in their construction process. Another practical example can be seen in the approach of some book authors who release their novels chapter by chapter. This method allows them to adapt the story’s direction based on reader feedback.

Personal note: I would not recommend writing a novel in such a manner as the delivery of individual chapters does not improve the reader experience, and it’s incredibly hard to create an engaging story when you don’t iterate over the entire scope of the manuscript. There’s a reason movie productions don’t get started until people agree on the script (the story).

Typical to Agile (E)

Among all possible shifts between the four Agile Quadrants, the most significant is the transition from ‘Typical’ to ‘Agile.’ This change involves introducing experimental processes and incremental delivery into a setting that previously relied solely on BDUF.

Tesla serves as a prominent example of this transformation. They have revolutionized the automobile industry by demonstrating the feasibility of reducing car production batches to as small as one unit, exemplifying extreme incremental delivery. Additionally, Tesla is known for continuously experimenting with both delivered and undelivered cars, aiming to enhance the overall user experience through relentless iteration. This approach reflects a comprehensive shift to an Agile methodology, combining frequent, small-scale delivery with ongoing, innovative experimentation.

Original to Serial (F)

The transition from the ‘Original’ to ‘Serial’ quadrant represents a unique case where the overall production process doesn’t change but instead undergoes a shift in its primary focus midway. This involves toggling between the phases of discovery, which occurs in the ‘Original’ quadrant, and delivery, happening in the ‘Serial’ quadrant.

A prime example of this can be found in the production of TV series like “Game of Thrones” or “Stranger Things.” During the discovery stage, these shows are produced for an entire season, comprising perhaps eight or ten episodes, with a focus on iteration and experimentation. However, no incremental content is released to the audience during this phase. Once production is complete, the entire season is made available to streaming platforms like HBO or Netflix. Then, the TV show is delivered to viewers in a serial manner, with one episode released each week. Viewer feedback on individual episodes cannot influence the already completed season, underscoring the distinct separation between the discovery and delivery stages in this process.

Hard Agile versus Soft Agile versus Waterfall

In my interactions with the Agile community, I often encounter the belief that agility is only possible with frequent, incremental delivery (as in quadrant 4), and anything less is quickly labeled as a “waterfall” project. However, the four Agile Quadrants open up a broader range of possibilities.

Books like Creativity Inc. about Pixar Animation Studios, How Big Things Get Done covering large construction projects, and Blood, Sweat, and Pixels on the games industry showcase intensive experimentation in highly challenging projects. These narratives all fall into quadrant 3, characterized by numerous iterations but culminating in a single launch event. While the books don’t call these approaches’ Agile,’ they certainly embody a form of agility, although different from the standard definition. They rely on iterations, mock-ups, prototypes, and digital modeling but do not produce a “potentially shippable product” at the end of each sprint.

Labeling every non-pure Agile approach as “waterfall” is not helpful. Big Design Upfront only exists in the first quadrant, which we can indeed label as “Not Agile.” Instead, it would be more accurate to describe quadrants 2 and 3 as examples of “Soft Agile.” While they don’t adhere strictly to the “hard agile” principles set out in the Agile Manifesto, they still involve feedback cycles. Quadrant 2 benefits from feedback via incremental delivery, and quadrant 3 through feedback from iterative experiments. Both are valid forms of feedback in their respective contexts. The glass is half-full, not half-empty.

You can now answer the questions posed at the beginning: Is Toyota agile? Is Apple agile? And is my writing process agile?

Congrats, you are a better agilist now. 🙂

Jurgen Appelo is a writer-speaker-entrepreneur. He is available for webinars, conferences, and company events. Contact him for details.

The unFIX model helps organizations move across the Agile quadrants. Contact the unFIX team for details on workshops and other assistance.