A business may contain a thousand imperfections. The operator’s job is not to repair them in order. It is to discover which one currently limits the next valuable result, then make the rest of the organization serve that discovery.
This path is a field instrument. Use it before launching a campaign, hiring a team, or redesigning a process simply because the problem is visible.
Map acquisition, conversion, retention, and capacity with the same period, scope, and definition of valuable output. Pick the lane with the strongest combination of economic impact, evidence, controllability, and learning speed.
flow map → comparable evidence → candidate constraintUse these descriptions as prompts, not as labels to paste over incomplete evidence.
Demand is the constraint when the company cannot create enough qualified opportunities for the downstream system to use. Measure the fit-adjusted input, its source, its response survival, and the cost of acquiring it. A crowded lead report is not the same thing as a healthy demand stream.
The question: if qualified demand increased tomorrow, would the next stage turn it into more valuable output, or would the added volume simply wait in another queue?
Conversion is the constraint when qualified demand arrives but does not become a decision. Look for the exact stage where the buyer loses momentum: diagnosis, proof, proposal, risk, price, follow-up, or implementation confidence. The overall close rate is a summary; the stalled step is the evidence.
The question: what did qualified buyers do immediately before they stopped, and what changed in the wins that moved forward?
Retention is the constraint when the purchased outcome does not become a durable behavior. Read cancellation timing, cohort survival, activation, usage, support, contraction, and customer fit. The stated reason for leaving is useful context, but the preceding behavior often gives a better operating clue.
The question: which leading behavior disappears before value disappears, and which part of the value path could have changed it?
Capacity is the constraint when work accumulates at a scarce role or handoff. It may be a senior decision, a delivery specialist, a quality review, a founder approval, or rework caused by standards that were never written down. Queue age and cycle time make the invisible cost visible.
The question: what step is continuously asked to process more than it can complete, and how much of its time is lost to work that should never have reached it?
Good diagnosis narrows the field. It does not turn an attractive story into a permanent explanation.
Use a ranking heuristic with humility. Economic impact multiplied by evidence strength and controllability, divided by time to learn, gives the team a disciplined way to compare candidates. It is not a universal business equation. It is a way to surface the assumptions behind “this is the priority.”
Then write the counterfactual. If the selected constraint improved, what would rise next: qualified input, decisions, durable value, or processed work? If no downstream movement is plausible, the proposed constraint is probably not the one to attack.
The discipline is simple enough to repeat and strict enough to stop expensive distraction.
Locate the system’s current limit with comparable flow and economic evidence. Name the step, not the mood.
Use the existing constraint better. Remove interruptions, rework, missing inputs, and low-value work before adding capacity.
Let adjacent work serve the pace and quality requirements of the constraint. A system can move faster when some local activity slows down.
Invest in people, spend, tools, or distribution only when the constraint remains after the first three moves and the economic case is explicit.
After the intervention, return to the map. The new constraint is not a failure; it is evidence that the system has changed.
No. It may be a decision, handoff, promise, qualification rule, scarce skill, or missing behavior. Department labels are often too broad to show where flow is actually limited.
When work accumulates faster than it is processed, the queue grows and time in the system rises. Use the relationship between work in process, throughput, and cycle time as a diagnostic lens, while keeping the scope and period consistent.
It means the selected constraint receives the scarce attention, sequencing, and decision speed. It does not mean ignoring safety, legal, customer, or financial obligations. It means refusing to launch unrelated optimization projects while the limiting mechanism remains unaddressed.
Define the expected leading signal and the downstream output before the test. Look for movement in throughput, conversion, durable contribution, queue age, or another system-level result while monitoring the guardrails that protect quality and economics.
That is useful information if the test was bounded and the baseline was clear. Record the failed hypothesis, reopen the flow map, and choose the next candidate. The operating system improves when it learns quickly, not when it defends an incorrect story.
Bring the flow map, the comparable numbers, the counterfactual, and the next test. Leave every other improvement idea in the margin until the system tells you it is time.
Begin The Reading