State Leverage: Doing the Right Things in the Right Order

Ordinary people often overvalue “hard work.” Especially in today’s era of intense competition, many companies default to the belief that increasing efficiency requires everyone to work harder and spend more time. While effort may be a necessary condition, if you only think about working hard all day, you are missing the bigger picture.

Hard work is more like a scalar—it only has intensity. Action, however, has at least three components: direction, timing, and sequence.

The mental tool for this lesson is called “State Leverage.” This is a term I’ve coined; while there isn’t a unified term in academia, many existing research results support it.

Simply put, State Leverage is about whether the world becomes friendlier to your next step after finishing the current one.

Let’s look at two small insights from daily life.

The first comes from housework. You bought several packs of ribs from the supermarket and plan to freeze them for later. You have two choices: one is to cut them first, pack them into portions, and then freeze them. When you want to cook, you just open a bag and throw them in the pot. The other is to freeze them whole and cut them when you need to eat. The second approach allows you to be lazy temporarily, but it brings a lot of trouble later—you have to thaw the ribs before cutting, which requires planning, not to mention that thawed ribs are harder to cut.

Cutting them as soon as you buy them is “front-loading” the labor. This not only reduces the total workload but also pushes the situation into a low-friction state.

The second insight comes from the classic novel Journey to the West. Tang Sanzang and his disciples arrived at the Jisai Kingdom and found an injustice. To fulfill a vow, Tang Sanzang went to a monastery to sweep a pagoda. He and Sun Wukong started sweeping from the bottom floor upwards. When they reached the top, they heard monsters talking and solved the case. This method of sweeping was for the sake of the plot, but it raises a question: shouldn’t you sweep a pagoda from the top down?

If you sweep from the top down, the dust from the upper floors falls down and is swept away with the lower floors. But if you sweep from the bottom up, the moment you finish the bottom floor and start the next one, the dust falls right back down. You could sweep all night and the pagoda would still be dirty.

Actually, both methods take similar physical effort—fatigue mainly comes from moving the broom and adjusting posture. But sweeping a pagoda doesn’t satisfy the commutative law: doing A then B is very different from doing B then A.

Sweeping from the bottom up creates a temporary state of cleanliness that is immediately undone by the next action. Only sweeping from the top down creates irreversible progress.

The lesson from ribs and pagodas is the same: the order of actions is at least as important as the effort invested.

State Leverage refers to actions that change the system’s state, reshaping the cost, probability of rework, and room for choice for subsequent actions.

What you need to optimize is not just the action itself, but the state left behind after the action.

Let’s discuss three types of leverage.

1. Preventive Leverage #

This refers to actions taken at the “upstream” of a project, equivalent to the design phase. The earlier a decision is made, the stronger its locking effect on the whole.

NASA has a famous observation: while the design phase of a project accounts for only about 15% of the total budget, it commits roughly 75% of the total lifecycle cost. How testing, manufacturing, integration, operation, and maintenance are done is largely determined by early design decisions.

If you find a small error during the design phase, it’s easy to fix. But if you wait until the implementation phase to find the problem, the cost of redesigning and revalidating is enormous. Like building a skyscraper: a single line on a blueprint is easy to change, but a mistake found on-site leads to hammers, steel, budgets, delays, arguments, and legal procedures.

According to a 2002 NIST report, more than half of software defects are not found upstream but are dragged into downstream development and usage. The later a defect is found, the higher the repair cost. If testing infrastructure could be improved to find problems earlier, the report estimated that over $22 billion in losses could be avoided annually.

So, design seems “cheap,” but it is actually the most “expensive” part of a project.

This is why Danish economist Bent Flyvbjerg says in How Big Things Get Done, the larger the project, the more careful you should be in the design phase. His mantra is: “Think slow, act fast.”

Prevention is better than cure. Drinking more coffee in the early stages to avoid countless hours of overtime later—that is Preventive Leverage.

2. Sequencing Leverage #

This means that different tasks in a project have information dependencies and must be carried out in a strict order.

For example, when renovating a house, you can’t paint the walls before doing the plumbing and electricity. If the plumbing needs changing, the walls have to be broken open again. You can’t lay the floors before remembering to install the central air conditioning. The correct order is: plan, plumbing/electricity, masonry/woodwork, painting, and finally floors and furniture.

Many people view a project as a simple to-do list. But a complex project is a dependency network. For this, you need a tool called the Design Structure Matrix (DSM).

The DSM maps who depends on whom and who influences whom in a square matrix. It shows which tasks can move forward independently, which are coupled, and where there are loops. It tells you how to rearrange the sequence so that information flows unidirectionally and progress accumulates monotonically, avoiding “waiting for each other.”

A classic case is using DSM to streamline Ford’s automotive hood system development. A hood seems simple, but it involves a series of interlocking decisions.

The styling team wants lower lines, which affects internal clearance. Clearance changes affect the hinges’ position and trajectory. Hinge changes affect the latch, assembly, crash safety, and even how robots install it in the factory.

The trouble isn’t the volume of work, but the entanglement. If styling changes the shape first, it’s just a few strokes, but the structure team later finds the original parts don’t fit. Rework begins. Manufacturing then has to re-evaluate the process. Safety tests’ assumptions become invalid. The project is consumed by back-and-forth pulling.

DSM identifies the strongest coupling and potential rework loops. You can then rearrange the order: which problems must be coordinated early, which can wait for stable conditions, and which steps shouldn’t be rushed.

If you make everyone start at the same time, you are just manufacturing scrap together.

3. Constraint Leverage #

This comes from the Theory of Constraints (TOC) proposed by physicist Eliyahu M. Goldratt.

The core logic is simple: the output of any system is limited by its narrowest point—the “bottleneck.” Don’t rush to optimize the whole process; solve the bottleneck first.

If your marketing brings in 100 leads a day, sales converts 50, but the delivery team can only serve 10, then delivery is your bottleneck. If you don’t improve delivery but instead double marketing to 200, is it useful? You haven’t increased output; you’ve just created internal chaos and customer complaints.

The overall throughput of any system—be it an assembly line, a R&D team, or your personal life—is determined only by the slowest link. Bustling activity in non-bottleneck areas is just a “mirage of prosperity.”

A review of hundreds of TOC cases found that once a company identifies the bottleneck, improvements are massive: lead times reduced by 69%, cycle times by 66%, and financial performance increased by 82%.

Goldratt’s five steps:

1. Identify the bottleneck.
2. Exploit the bottleneck (make sure it doesn’t waste time).
3. Subordinate everything else to the bottleneck.
4. Elevate the bottleneck (add resources to widen it).
5. Go back to step one.

The logic is simple, but there’s an emotional hurdle: if you aren’t at the bottleneck, you don’t need to be so busy. But bosses often can’t stand seeing employees idle, not realizing that effort outside the bottleneck is often wasted or even harmful.

I call these “State Leverage” because they have a deeper mathematical soul: Dynamic Programming, proposed by Richard Bellman in 1954.

Most people work using a “Greedy Algorithm”: “How can I do this task with the least effort right now?” Like throwing ribs in the freezer—it’s the easiest solution now, but it makes things harder for the “future me.”

Greedy algorithms often result in a series of local optima that create a global disaster. Bellman’s Dynamic Programming saves this.

The Bellman Equation states: Total Reward = Immediate Reward + Potential Value of the next state.

In Dynamic Programming, the purpose of an action is not just the present, but a “State Transition.” Like a professional billiards player: they don’t just sink the ball; they control the spin so the cue ball stops in the perfect position for the next shot.

• Preventive Leverage sacrifices immediate reward to push the system into a “no-rework” high-value state.
• Sequencing Leverage uses DSM to plan the correct state transition path.
• Constraint Leverage says the variable determining the “value of the next state” is the bottleneck.

Simply put, when deciding your next move, don’t just look at its immediate value. Look at what state it leads to: is the road ahead wider or narrower? Are costs lower or higher?

Bellman’s Principle of Optimality: Regardless of your past or starting point, your current decision must ensure the best possible future from the new state.

Everyday Applications #

• Preventive Leverage: The French culinary concept of “Mise en place” (everything in its place). Chefs spend a lot of time prepping ingredients before turning on the heat. Once the fire starts, everything flows smoothly. Your work environment should be in a low-friction state. Before writing, gather all data and outlines so you can stay in a pure output state.
• Sequencing Leverage: Stop the urge to do things that “look like progress” but aren’t. Don’t spend 15 minutes on PPT fonts before the core theme is decided.
• Bottleneck Leverage: Don’t work hard everywhere; work only on the “vital point.” Value lies in the link that determines the rhythm of the whole.

You might think your inefficiency is due to lack of willpower, but your bottleneck might be physical: lack of sleep. A project might be slow not because of execution, but because only the boss has decision-making power, and his bandwidth is the bottleneck.

Same tools, same effort. But because the location, order, and focus differ, the result undergoes a qualitative change.

Doing it at the “right spot” vs. not: One is accumulation, the other is consumption; One becomes smoother, the other more blocked; One earns interest, the other pays penalties.

Are you just finishing a task, or are you creating a better future state?

References

[1] Hirshorn, Steven R., et al. 2016. NASA Systems Engineering Handbook. [2] Tassey, Gregory. 2002. The Economic Impacts of Inadequate Infrastructure for Software Testing. NIST. [3] Flyvbjerg, Bent. 2023. How Big Things Get Done. [4] Eppinger, Steven D. 2012. Design Structure Matrix Methods and Applications. [5] Zambito, Antonino Paolo. 2000. Using the Design Structure Matrix to Streamline Automotive Hood System Development. [6] Goldratt, Eliyahu M. 1984. The Goal. [7] Mabin, Victoria J. 2003. “The Performance of the Theory of Constraints Methodology.” [8] Bellman, Richard. 1954. “The Theory of Dynamic Programming.”