What the theory of constraints is
Think of any marketing system as a chain of steps. A FunnelModel of the customer journey stages. runs from arrival to purchase. A content line runs from idea to brief, draft, review and publication. A lead process runs from enquiry to qualification, call and proposal. The output of each chain is set by its weakest link: either the step with the least capacity, where work queues, or the step that loses the largest share of what reaches it.
Three consequences follow, and they are the reason the framework is useful. First, time or money gained at any step other than the constraint changes nothing at the end of the chain; it only moves work faster into the queue in front of the limit. Second, work piles up before the constraint, so queues are the clearest sign of where it is. Third, each team maximising its own output can make the whole system worse: a demand team that doubles leads for a sales team that cannot call them creates waste, not revenue.
The method is known as the five focusing steps, and it is general enough to apply to funnels, production workflows, sales capacity and budgets alike. The funnel and checkout hub applies one version of the idea to a single funnel: find the step whose improvement would add the most conversions and work there first. This page applies the same logic to any marketing system, including the parts that sit behind the funnel.
The focusing steps
| Step | Question | Marketing example |
|---|---|---|
| Identify | Which step limits the output of the whole system? | Drafts waiting days for legal review; a checkout step that loses buyers who meant to pay; leads that sales cannot call within a day |
| Exploit | How can we get more from that step without new spending? | A pre-approved claims library so legal reviews only what is new; removing optional checkout fields; routing only qualified leads to sales |
| Subordinate | How should every other step be paced to the constraint? | Briefing no more pieces than review can clear; timing promotions to days with spare fulfilment capacity; pacing lead generation to calling capacity |
| Elevate | If the step still limits output, how do we add capacity? | A second reviewer; a rebuilt checkout; another sales development role; extra fulfilment shifts |
| Repeat | Has the constraint moved, and where to? | With review cleared, briefs become the limit; with checkout fixed, the quality of traffic becomes the limit |
The order matters. Exploiting before elevating avoids buying capacity the system does not need, because much of a constraint's time is often lost to avoidable work, waiting and rework. Subordination is usually the hardest step, because it asks other teams to produce less, or produce at a different pace, which can look like underperformance on their own measures. And the repeat step guards against inertia: a policy built around an old constraint, such as a rule that every piece goes through a full legal review, can itself become the next constraint once the original limit has gone. When the system is a marketing team's own way of working, the guide to prioritising marketing operations work weights each improvement by whether it acts on that limiting step.
Finding the constraint in a funnel, a workflow or a team
In a funnel, measure each step: how many people enter, the step CVRShare of visits or clicks that convert., and how many leave. The funnel analysis tutorial builds this in an analytics exploration and does the arithmetic of which step would add the most conversions if improved. The constraint is that step, not simply the one with the lowest rate. A high Cart Abandonment RateShare of shopping carts created that did not become an order. may be the constraint, or it may be ordinary browsing behaviour that no fix would change much. Then look behind the funnel. If stock, delivery slots, appointment capacity or sales follow-up are already full, improving the funnel produces orders the business cannot serve, and the real constraint is operational. The guide to planning marketing around capacity and stock turns that finding into a plan, with a weekly capacity view and four responses: shift, slow, queue or stop demand.
In a workflow, trace recent items from request to completion and split the elapsed time into working and waiting. The constraint is where work waits longest and where the queue grows when demand rises. The guide to the campaign workflow from intake to launch sets out the usual stages and gates; a queue that forms in front of one gate, a stage everybody waits on, or a person whose calendar decides every launch date are the signs to look for. The tutorial on mapping a marketing process draws that trace in lanes by role, with every handover, wait and loop back marked, so the queue is visible before anyone picks a fix.
In a team or a business, the constraint may be capacity that marketing does not control: sales time, customer service, onboarding, fulfilment or cash, where the limit on growth is how quickly spending pays back. It may also be demand itself. If the market is not asking for more, no amount of internal capacity raises output, and the constraint is reach and demand creation.
Whichever system is being examined, pick one output measure for the whole of it, such as orders delivered, pieces published or qualified meetings held, and make it the KPIA metric tied directly to an objective. that judges every change. Local measures, such as drafts written or leads generated, reward activity at steps that are not the limit.
Worked example (illustrative)
A meal-kit service finds its constraint in the kitchen
Bramblequist Meal Kits, an invented meal-kit service delivering on Mondays, Wednesdays and Fridays, keeps raising paid social spend, yet weekly orders have stalled at about 2,400. The website funnel converts steadily. The limit is Monday packing: the kitchen can pack 1,100 boxes for Monday, and that slot sells out by Thursday each week, while Wednesday packs about 600 boxes and Friday about 700 against the same limit. At checkout, 38% of new customers who are offered only Wednesday or Friday abandon. The team works through the focusing steps. To exploit the constraint, the kitchen pre-portions the 5 most popular recipes on Saturday, raising Monday capacity to 1,250 boxes. To subordinate marketing to it, the first-box offer becomes valid only for Wednesday and Friday deliveries, and the Sunday-evening ads that drove Monday demand are paused. Paid reach stays flat. Over the next 6 weeks, weekly orders rise to about 2,750, with Monday full and the other two days busier. The team's next question is whether Friday packing will become the constraint as it fills.
Limits and criticisms
It assumes one binding constraint. A production line makes similar items in a fixed sequence; marketing systems rarely do. In marketing, demand and quality interact: a weak offer depresses every step of the funnel at once, and a landing page, an audience and a creative can share the blame for the same drop. The constraint can be genuinely ambiguous, so its identification should be treated as a hypothesis to check, by looking for where work queues and for whether improving that step actually moves the output measure.
Fixing the biggest drop is not always the best return. The step that loses the most may be expensive to improve, or the loss may be natural, and a cheaper improvement elsewhere may add more output for the effort. The framework says where the limit is; it does not price the fix, so the cost and likely effect of each option still have to be compared.
Constraints move. Relieving one shifts the limit to another step, and seasonality can move it without any change to the system, so the analysis has to be repeated. Teams that keep investing in a constraint that has already been relieved spend money for no gain.
The flow view also leaves things out. Brand building, reputation and long-term demand do not move through a chain of steps, and optimising throughput can quietly crowd out work whose effect arrives later. And subordination is often harder politically than analytically, because it asks teams to accept lower local output for the sake of the whole.
Related frameworks
The AARRR framework divides the customer lifecycle into acquisition, activation, retention, referral and revenue, and the theory of constraints asks which of those stages is currently limiting growth; money spent on acquisition while activation is the constraint mostly fills a leaking bucket. The KPI tree breaks an outcome into its drivers, and the constraint is usually the driver that is both limited and capable of moving the outcome most, which makes the tree a natural map for the identify step. When the question is which customers, products or partners deserve most attention rather than which step limits output, the comparison of the Pareto principle and the theory of constraints sets out which lens fits.
Common mistakes
- Optimising a step that is not the constraint. Faster drafting does nothing for a content line stuck in review.
- Buying more traffic into a full system. If fulfilment, stock or sales capacity is the limit, more demand creates waste and unhappy customers.
- Elevating before exploiting. Adding capacity before removing waste at the constraint pays for slack.
- Judging teams on local output. Leads generated or drafts written reward activity at steps that do not set the total.
- Assuming the constraint is inside marketing. Operations, sales and service often set the real limit.
- Treating the lowest rate as the constraint. The limit is where improvement moves the whole system, not where the rate looks worst.
- Not re-checking after a fix. Once a constraint is relieved, the limit is somewhere else.
