Operations & Systems Thinking

Operations & Systems Thinking

A business without systems is a person doing everything. A business with systems is an organization that produces results consistently, regardless of which individuals are present. The transition from the first to the second is the most important scaling inflection point for any operator.

What Systems Actually Are

A system is a set of inputs, processes, and outputs that consistently produces a defined result. In business, systems are built on three components:

1. Documented processes: The sequence of steps and decisions required to produce the output
2. Clear owners: The person or role responsible for each step
3. Feedback mechanisms: The way the system signals when it is off-track

Most businesses have the first without the second, and almost none have the third. A checklist is not a system if no one is checking the checklist. A handoff process is not a system if the definition of "done" is unclear between the sender and receiver.

SOPs: Converting Individual Knowledge to Organizational Knowledge

The test for whether you need an SOP: if the person who does this task left today, how long before things broke? If the answer is "immediately," the process lives in a person's head, not in the organization.

SOP components that actually work:

• Trigger: What event initiates this process?
• Owner: Who is responsible from start to finish?
• Steps: Numbered, action-verb-led instructions specific enough that someone unfamiliar with the process could execute them
• Decision branches: What happens if X? What happens if Y?
• Output: What does done look like? What gets handed to whom?
• Review schedule: When does this SOP get reviewed for accuracy?

The last point is often skipped. A 12-month-old SOP in a fast-moving company may be more dangerous than no SOP — it creates false confidence in a process that no longer reflects reality.

The Delegation Ladder in Practice

Most founder delegation failures happen because the delegation level was never defined. The founder thought they were delegating full ownership. The employee thought they needed approval at every step. The result: nothing moves without the founder and the founder blames the employee.

The five levels:

1. Look into this and report back — pure information gathering, no action
2. Give me options with your recommendation — analysis and judgment, decision stays with you
3. Decide and let me know what you decided — action taken, founder informed
4. Act and report at the next regular check-in — full execution within defined parameters
5. Own this completely — results ownership, method autonomy

For any new delegation: start at level 1 or 2, accelerate to 4 or 5 as the employee demonstrates judgment. Skipping levels with inexperienced employees creates expensive mistakes.

The Org Chart as Information Architecture

The conventional view of an org chart is a hierarchy of authority — who reports to whom. The more useful view: the org chart is a map of how information travels, decisions are made, and accountability is assigned.

When an org chart creates a bottleneck where a single leader must approve everything, information queues up at that node. When an org chart has unclear ownership between two functions, decisions fall into a gap. When reporting lines are misaligned with actual work flow, coordination cost is paid daily in meetings that exist to compensate for the structural misfit.

Designing structure for information flow:

• People who need to coordinate frequently should be in the same team or report to the same leader
• Decision rights should sit with the person who has the most relevant information and accountability for the outcome
• Spans of control (how many direct reports a manager has) should reflect the complexity of the work: 5-8 for complex, judgment-intensive work; 8-15 for more procedural roles

Identifying and Relieving Bottlenecks

The Theory of Constraints (Goldratt's framework) provides a five-step operational improvement process:

1. Identify the constraint — where is work piling up or waiting? The stack before the constraint reveals it.
2. Exploit the constraint — maximize throughput at the bottleneck without additional resources. Cut waste in the constraint process first.
3. Subordinate everything else — align non-bottleneck steps to support the bottleneck's pace. Do not optimize elsewhere.
4. Elevate the constraint — invest additional resources to increase bottleneck capacity.
5. Repeat — once the constraint is resolved, a new one will emerge elsewhere. The process is continuous.

In service businesses, the bottleneck is often the founder's review. The operational intervention is not to work faster — it is to build a system where fewer decisions require the founder's involvement.