This is an excerpt from the book Kanban Remastered: Agile Lessons from Strategy Games.
In an organization, work can be distributed by two different approaches: “push” or “pull.” You either get new work put on your desk and have to manage what to prioritize, or you ask for new tasks once you are finished with your old tasks.
For example, “Gosplan,” the agency responsible for central economic planning in the Soviet Union, used a push-based logistical system. At Gosplan, mathematical models predicted consumer behavior. Based on these models, the government created plans for the entire supply-chain from the factory to the shops—with catastrophically bad results. For certain goods, there was always either too much or too little available. By contrast, “Kanban” is different:
KANBAN · Kanban is Japanese and literally means “signboard.” In the context of project management, the term is interpreted as “queue limitation.” Kanban is a method designed to reduce unfinished work and wasteful inventory levels; it was originally developed at Toyota in the late 1940s. Back then, marketers at Toyota studied consumer behavior and supermarket stocking strategies and applied the ideas to logistics in industrial production. At Toyota, they had previously produced as much as possible, regardless of the demand from the market. In contrast, in supermarkets, customers take only what they need, expecting that the supermarket will be stocked up the next time they visit. The customer “pulls” an item from the shelves and the supermarket makes sure to refill the shelves. This new Kanban method applied to production provided just as much as what was needed, just in time.
In addition to improving the production flow within a company, the big advantage of Kanban is that it can be applied to the production phases of any existing organizational structure without having to change business processes. With the production flow being made transparent, you can detect where work piles up at one place and then introduce a limit of the amount of work that is in progress. As opposed to a “relay race” model where work is pushed to the next department and not followed through, in Kanban, work items are pulled and work is stopped when the work limit of the next department hits its limit. This ensures that no work piles up or gets lost.
For example, cars need tires, a chassis, and a motor. Without a Kanban system, the three departments responsible for these parts will produce as much as possible: the company will always end up having either too many chassis, motors, or tires. If you instead have the departments check how many items are already in stock and stop producing when the stock is full, you save a lot of money. Sure, you might need these items later, but keeping a large inventory costs money that you could have invested elsewhere.
When presenting the situation in a business this way, it opens an objective discussion at the management level, ideally followed by incremental, evolutionary improvements. As opposed to Scrum which can be executed within a company “by name only,” i.e., by following the ceremonies but never addressing core issues like multidisciplinary teams (see my book Scrum Your Jira! ), Kanban is a organization-wide change management approach. Sure, Scrum involves creating lists of impediments but it leaves it up to you how to deal with them.
SCRUM · Scrum is a set of management tools that focuses a project back on the team level and uncovers internal and external impediments of the production process. By reducing communication paths through small, multidisciplinary teams, as well as frequent releases to the customer for review, the probability for project success can be improved even if the scope is not clear from the start. In addition, work is divided into units of fixed lengths (sprints) which helps to plan future sprints with your team working at a sustainable speed.
Once Kanban is in place, the goal is to focus on the bottlenecks, and manage the flow. Discussions with management can best be lead by making things explicit. Identifying the bottlenecks themselves is just the start of your work. The real game changer is to make explicit the current collaboration policies. This moves the discussion away from the abstract and maybe emotional or anecdotal arguments and toward objectivity.
After an organization-wide introduction of Kanban, usually the first bottleneck is found at the top. If the organization is strongly hierarchical, work (decisions!) piles up at the desk of the organization’s leader. In StarCraft, this is not much of a problem as playing in a team inherently includes decentralized control, but it helps to imagine how cumbersome a game would be if all decisions had to go over a third party’s desk.
Even with a decentralized organization, a common argument against Kanban is that there will be idle time because one part of the organization might not be able to keep up with the rest. As a result, work piles up until it hits the limit. This is true, but the idea that you are doing a good job when everyone is working at 100 percent is not always true. This is comparable to, in StarCraft, trying to keep all building facilities active, whether you need the units at the moment or not—or maybe even refraining from building additional facilities fearing they would not be used all the time.
Ultimately, Kanban is about trying to improve the flow through an organization. It does not matter how many items you produce if your sales department cannot bring them to the customer. Likewise, in StarCraft, you must think about how you will use the units you produce, meaning how you will deploy them on the battlefield. Sure, producing as many units as possible is a viable strategy, just like you could throw your unsold products on the market at a lower price or keep them in storage “just in case” there is a sudden demand. But it is an inefficient strategy. We can do better than mere local optimization.
This rings especially true when looking at the market. While a business does not literally fight an enemy like you do in StarCraft, your competitors try to get your market share. Scouting your enemy and adjusting your strategy are central components in StarCraft, just as they are on the market. You need to have foresight and gain intelligence about them and maybe even think about making some early risky investments before you are absolutely sure what others are doing. Why? Because waiting itself is a risk and you might miss the window of opportunity to be the first on the market with your product.
When all is in place, you can focus on communication. Using Kanban automatically leads to situations where a team is stuck insofar as it cannot pull new items to work on because the subsequent phases or departments have hit their work in progress limit. This encourages collaboration, where one team can help out the next and where teams sit together and think about how to prevent future bottlenecks. This element can be found in StarCraft, too, with very close team communication over audio during the game, as well as a review of the game and discussions about how to optimize team play afterward.
No matter the size or structure of your organization, take small steps. Kanban promotes this approach by starting with what you have in place and pointing out the bottlenecks. Where StarCraft falls short, by comparison, is the visualization of the “workflow.” While tools eventually emerged that visualized some aspects of the replays, like your “actions per minute” symbolizing your workload, I know of no tool that does it the Kanban way and actually analyzes how much time you spend on each team or location and thus points out possible paths for optimization.
The closest software that does a similar job is Evolution Forge, which I will discuss in the last chapter of this book. It helps you to optimize your basic build order in small steps. Behind both approaches (Kanban and Evolution Forge) is the grand idea inherent in nature to leave things as they are and move forward without ever making a step back. Every change you make should improve the situation and larger changes come into place as the sum of a whole number of smaller ones. In that regard, if you want to improve your StarCraft play the Kanban way, manual observation, maybe together with a critical friend, might be the best choice. On the other hand, if you want to learn the idea of Kanban with StarCraft as an easy-to-understand reference, you have come to the right place!
“When you choose to use Kanban as a method to drive change in your organization, you are subscribing to the view that it is better to optimize what already exists, because that is easier and faster and will meet with less resistance than running a managed, engineered, named-change initiative. Introducing a radical change is harder than incrementally improving an existing one.” —David J. Anderson, Kanban: Successful Evolutionary Change for Your Technology Business [Anderson, 2010, p. 1]