Thinking of Work at Scale

The Wrong Level Problem

Most attempts to "improve how we work" fail for a reason nobody talks about: they're operating at the wrong level.

Some companies zoom in too far. They obsess over documenting every micro-step, creating procedure manuals that are exhaustive and exhausting. Nobody reads them. Nobody follows them. The documentation becomes its own kind of work—maintained by nobody, followed by nobody, useful to nobody.

Other companies stay too high. They draw beautiful value stream maps at offsite retreats, post them on the wall, and then... nothing changes. The map shows how work should flow, but it doesn't connect to how work actually happens. It's strategy without execution.

The magic is in knowing when to work at which level.

To think about work at scale—whether you're running a five-person team or a five-hundred-person organization—you need a mental model that lets you zoom in and out deliberately. A hierarchy that connects the big picture to the daily details without losing coherence at either end.

The Four Levels of Work

Work exists at four distinct levels. Each answers a different question, serves a different purpose, and requires a different kind of attention.

Think of this like a map. Value streams are the continent—the overall terrain. Workflows are the highways connecting major cities. Stages are the towns you pass through. Steps are the individual turns and intersections that get you from one town to the next.

You need all four levels to navigate effectively. But you don't need all four levels at the same time.

Value Streams: The Customer's Journey

A value stream is the path value takes from the moment someone needs something to the moment they get it—and you get paid. That's it. No consultant jargon required.

What makes value streams powerful is that they cut across your org chart. Sales doesn't own a value stream. Neither does operations. Neither does finance. The customer's experience spans all of them.

Same stream. Two perspectives. When you see both, you can identify where value gets stuck, delayed, or lost.

When to work at this level:

• → Strategic planning—deciding which flows matter most
• → Diagnosing cross-functional friction—when handoffs between teams are broken
• → Understanding customer experience—seeing what they see, not what you see
• → Executive reporting—communicating operational health in terms that matter

Value streams are too high-level for execution. You can't tell someone to "go work on the onboarding value stream." But they're essential for ensuring that all the work you design actually connects to outcomes that matter.

Workflows: Where Chaos Becomes Finite

If value streams are the journey, workflows are the specific routes you take. Each workflow handles a particular type of work—and here's the key insight: your business only has a finite number of work types.

When you're drowning in chaos, everything feels unique. This customer has a special situation. That project is different. This request doesn't fit the normal pattern.

But underneath the chaos, you're actually doing a small number of things over and over:

Once you name these work types, you can design a workflow for each—a clear path from how it starts to what "done" actually means. Suddenly, a thousand different "things we do" become a manageable set of flows you can actually improve.

When to work at this level:

• → Process design—defining how work should flow for a specific type
• → Tool selection—choosing systems that support your workflows
• → Team alignment—ensuring everyone understands the path from start to done
• → Capacity planning—understanding how much of each work type you can handle

Workflows are the level where most operational improvement happens. They're specific enough to be actionable, but broad enough to see the whole picture.

Stages: The Rhythm of Progress

Inside every workflow, work moves through stages. A stage is a defined position—a place where work pauses, transitions, or waits for something to happen.

Think of stages like columns on a kanban board. "New" is a stage. "In Progress" is a stage. "Waiting on Customer" is a stage. "Complete" is a stage.

Each stage should mean something operationally different. If work is "In Progress," someone is actively working on it. If it's "Pending Confirmation," the work is done but we're waiting for the customer. If it's "Triaged," we understand the problem and it's ready to be scheduled.

Stages give you visibility. At any moment, you can ask: "How many service requests are at each stage?" That tells you where work is piling up, where bottlenecks are forming, where the flow is getting stuck.

When to work at this level:

• → Operational visibility—seeing where work is right now
• → Bottleneck diagnosis—identifying where work stacks up
• → Handoff design—defining who owns what at each stage
• → SLA management—setting expectations for how long work stays in each stage

Most of your daily management happens at the stage level. It's where "how's the work flowing?" becomes a question you can actually answer.

Steps: The Atomic Unit

Steps are where work actually happens. They're the atomic unit—the smallest piece of work that transforms something from one state to another.

A stage might contain multiple steps. "Triage" as a stage might include steps like: review the request, gather additional information, assess priority, assign to the right team. Each of these is a discrete action with its own inputs and outputs.

Here's what makes steps powerful: when you define them properly, they become portable. A well-defined step can be executed by different people, delegated to new team members, or even automated—because everyone (or everything) knows exactly what needs to be true before the step starts and what needs to be true when it's done.

Most workflow failures don't happen at the level of the whole flow. They happen inside steps that were never properly defined. "Prepare kickoff" sounds simple—but what information is actually needed? What artifacts should exist? What does "prepared" really mean?

When to work at this level:

• → Training—teaching someone exactly how to do a specific task
• → Automation—defining precisely what a system should do
• → Quality improvement—diagnosing why a specific step keeps failing
• → Delegation—handing off work with clear expectations

Steps are the most detailed level. You don't define every step in every workflow upfront—you'd never finish. But when a step keeps causing problems, when you're training someone new, when you're building automation—that's when you zoom in to this level.

Working at the Right Level

Ryan Singer, in Basecamp's Shape Up, describes a technique called "fat marker sketches"—using a thick marker to draw ideas so you physically can't add too much detail. The constraint forces you to stay at the right level of abstraction for that stage of work.

The same principle applies to thinking about work. The biggest mistake is jumping to the wrong level:

The discipline is matching the level to the question:

Start broad. Zoom in only when you need to. The hierarchy gives you a map; it doesn't require you to draw every building on every street before you can navigate.

Putting This to Work

Here's how to use this hierarchy in practice:

The Compounding Effect

Here's what happens when you think about work this way:

When you improve a step, every future piece of work that flows through that step benefits. When you clarify a stage, everyone who works at that stage knows what they should be doing. When you design a workflow, every instance of that work type follows a coherent path. When you understand your value streams, every improvement you make connects to outcomes that matter.

This is how you scale operational excellence. Not by documenting everything upfront—by building a hierarchy that lets you work at the right level for the problem you're solving, and zoom in precisely when precision is needed.