Imagine two café fit-outs. Both involve the same deliverables: a kitchen, a counter, seating for forty, a coffee machine and a cool room. A quoting estimator would price them similarly. But the first is inside a busy shopping centre that trades seven days a week: work can only be noisy outside trading hours, deliveries must use a loading dock booked by the centre, and the opening date is fixed because the lease rent starts then. The second is in a standalone heritage building in a quiet street, where neighbours have objected to the development and council approval for the exhaust system has not yet been granted.
Same deliverables, entirely different jobs. The first is a problem of building around a live operation against an immovable date. The second is a problem of consent: whether, and when, the people whose agreement is needed will give it. A deliverables list will never tell you which kind of job you have. The constraints will.
This article explains how to read a job’s constraints, find the one that will break first, rank conflicts before delegating the work, and then tailor the way the job is run, whether planned in detail, staged and experimental, or fast and repeatable, without weakening the accountability every job needs. It is general information for owners and managers who plan and delegate projects.
Constraints are not risks or assumptions
Three things are often mixed together in project documents, and each needs different treatment:
| Item | What it is | How to treat it |
|---|---|---|
| Constraint | A boundary condition that already applies | Design around it |
| Risk | Something that might happen | Assess, reduce, transfer or accept |
| Assumption | Something believed but not yet verified | Test it early |
Treating a constraint as a risk produces months of effort addressed to a problem the business has misdiagnosed. “Neighbour objections” listed as a risk gets a communication plan; recognised as the binding constraint, it changes the order of the whole job, because consent must come before money is committed. Treating an assumption as a constraint quietly freezes a decision nobody made. Testing a genuine constraint, on the other hand, usually wastes effort: it is not in doubt.
Types of constraint
Constraints fall into a few types, and the type suggests the response:
- Physical and operational: something exists and cannot be stopped or moved, such as a trading shopping centre or a production line that must keep running. Design the method around it; usually solvable with planning and money.
- Time: a date set by someone else, through a lease, contract, season or commitment already made. Work backwards from it and decide what can be sacrificed. Often underestimated.
- Money: a firm budget cap. Scope becomes the variable. If time and money are both fixed, then scope or quality must flex; if neither may, the plan is not a plan.
- Consent and legitimacy: someone whose agreement is needed does not agree yet, such as neighbours, a landlord, a council or a key customer. Seek consent before committing money. This is the type most often misclassified as a communication task.
- Regulatory: a condition set by an authority. Treat it as a dependency with a decision date and a named contact, not just a duration.
- Ethical or welfare: a condition on how the work is done, such as safety or animal welfare. It constrains method and often conflicts with cost.
Most jobs have three or four types. The one that binds is rarely the type the business is most comfortable managing.
Find the constraint that breaks first
Every set of constraints contains one that will break first under pressure. Finding it is the most valuable thing you can do at the start of a job. Ask: if pushed, which of these gives way? Then confirm it by asking what would have to change for it to stop binding. If nothing plausible would, design the job around it.
The binding constraint also tells you what kind of leadership the job needs. Domain experience is the entry ticket. In the shopping centre fit-out, the job needs someone who can stage noisy work around trading hours, coordinate a loading dock and protect a fixed opening date. In the heritage fit-out, it needs someone who can work patiently with neighbours and council, sequence approvals before spending, and tell the owner early if consent is unlikely. Same trade, different temperament.
Rank conflicts before you delegate
Constraints often conflict: a fixed date against a budget cap, keeping customers supplied against moving equipment quickly, using existing staff against redeploying them elsewhere. When nobody says which wins, the conflict does not disappear. It is resolved silently, at working level, by whoever is standing in front of it, without the authority, information or accountability to make that call. A site supervisor choosing between a faster and a safer method is making a decision the owner declined to make.
So, before delegating a job, write down for each pair of constraints that cannot both be fully met which one yields. It is uncomfortable, and it is the point. The escalation rules article covers sorting objectives into protections that must not be traded, minimum outcomes and variables.
Then tailor the method
Once you know what binds, choose how to run the job. There is no single approach that suits every project. Three broad patterns are useful:
- Plan-driven: detailed planning, firm baselines and change control. Suits work that is predictable, regulated or expensive to get wrong, where the solution is well understood.
- Change-driven: staged, experimental and iterative. Suits work where requirements or solutions cannot be known in advance, and learning is the main task.
- Time-driven: fast and lean. Suits repeatable work under time pressure, where only the essential controls are kept. Urgency does not remove governance; it makes it more important to know which controls are essential.
A single job can contain all three. A shop fit-out may use a plan-driven approach for the building works, an experimental approach for a new point-of-sale and ordering system, and a fast routine for restocking.
Six questions help decide how much control each part needs:
| Question | Higher answer suggests |
|---|---|
| How bad is failure? | More assurance and clearer decision rights |
| How uncertain are requirements or solutions? | More experiments and shorter feedback loops |
| How hard is it to undo a commitment? | Staged commitment and firmer gates |
| What legal or regulatory obligations apply? | More formal evidence and checking |
| How unfamiliar is this work to us? | More learning, specialist help and review |
| How fast can conditions change? | Shorter planning horizons and more frequent review |
The result should be a short, written reasoning: what is being increased, reduced or changed, such as planning depth, approval levels, review frequency, testing intensity or the contract type, and why. The difficult, complex or ambiguous article covers matching the approach to the kind of problem.
Tailor the method, not the accountability
Tailoring is not permission to drop governance. Deleting fields from a form is not tailoring either. The way of working can change; what accountability must prove should not. Every job, however it is run, should be able to answer:
- What outcome are we pursuing, and why is it worth it?
- Who can decide what?
- Which risks and assumptions matter most?
- Which commitments are hard to reverse?
- What evidence will show it is working?
- Who owns the result after handover?
Equally, state what cannot be tailored away: safety obligations, legal requirements and financial limits remain whatever the method. And review the choice: a job that starts in stable conditions may meet a supplier failure or a regulatory change, and a job that resolves its big uncertainties may be able to simplify.
When conditions change faster than the plan
A plan is a bet that conditions will hold long enough for the plan to finish. Some jobs run in conditions that change slowly: a known building, a stable supplier, settled rules. Others run in conditions that shift during the job: a customer still deciding what they need, a supplier market in flux, a rule under review. The faster conditions change relative to the length of the job, the less value there is in detailed long-range planning, and the more value in short planning horizons, early warning and decisions that can safely be made later.
A practical test is to compare two durations: how long the job will take, and how long the key assumptions behind it are likely to stay true. If the assumptions will probably outlast the job, plan in detail. If they will not, plan the next stage in detail and the rest in outline, and set dates to re-check the assumptions that matter most. Neither approach is better in general. The mistake is using one where the other fits, or failing to notice when conditions have moved from one to the other.
A worked example
This is an illustration. A small food manufacturer must move to a new factory within ten months because its lease ends. The deliverables are clear: move three production lines, a cool room, the office and stock.
The owner lists the constraints and types them:
- Customers must keep receiving orders throughout (operational).
- The old lease ends on 30 June (time).
- The budget is capped at $350,000 (money).
- The new site needs council approval for trade waste and a food business registration (regulatory).
- Two experienced line operators may leave rather than travel further (people and consent).
The binding constraint is customer supply: losing a major supermarket customer would cost more than any overrun. The owner ranks the conflicts in writing: supply continuity beats the moving date, which beats the budget; if the budget and supply conflict, the business will spend more, such as temporary contract packing, rather than miss deliveries. Retaining the two operators is treated as a minimum outcome, so the owner talks to them first about travel and shift times.
The approach is tailored by part. The building works and equipment move are plan-driven, with a detailed schedule and firm contractor dates. The council approvals are treated as a dependency with named contacts and decision dates, started immediately. Commissioning the lines at the new site is change-driven: the business moves one line first, runs it for two weeks with real product and adjusts the method before moving the others. Restocking and customer communications follow a simple routine.
The project lead is chosen for experience in keeping production running during change, not just for knowing the equipment. When the second line’s move slips, the lead knows what to do without asking: hold the date for that line until the first is stable, and use the agreed contract packer for one week to protect supply.
Halfway through, conditions change: the water authority asks for a larger grease trap than planned, adding cost and three weeks to the plumbing. Because the ranking is written down, the decision is quick. The budget yields, the lead re-plans the plumbing stage, and the owner reviews whether the remaining contingency still covers the commissioning trials.
How this applies to a small Australian business
- List constraints separately from deliverables, risks and assumptions.
- Type each constraint: physical, time, money, consent, regulatory or ethical.
- Find the binding constraint, and design the job around it.
- Choose a leader whose strengths match the binding constraint.
- Rank conflicts in writing before delegating.
- Tailor the method by part of the job, with written reasoning.
- Keep the accountability questions the same for every job.
- Review the approach when conditions change.
Signals worth watching
- Constraints recorded as risks, or assumptions recorded as constraints.
- Communication plans aimed at what is really a consent problem.
- Staff making trade-offs the owner never ranked.
- One method applied to every job regardless of uncertainty.
- “Agile” or “flexible” used to mean no decisions are recorded.
- Tailoring that removes evidence rather than changing how it is produced.
Common mistakes
- Reading the job from its deliverables alone.
- Treating every constraint as equally binding.
- Delegating unranked conflicts.
- Choosing a leader by trade rather than by binding constraint.
- Confusing tailoring with removing governance.
- Choosing a method once and never revisiting it.
Frequently asked questions
How long does reading the constraints take? For most small-business jobs, an hour with the people who know the work. It often reveals that effort has been aimed at the wrong problem.
What if there seem to be two binding constraints? Ask which would give way first if both were pressed. If they genuinely cannot be traded, the job may need redesigning, or the business must decide which outcome it values more.
Is a flexible approach riskier? Not if it has clear boundaries: purpose, safety and legal limits, budget limits and decision rights stay fixed while the details evolve with evidence.
Should small jobs use a formal method at all? Use the lightest method that still answers the accountability questions. A one-page plan can be enough.
Who should rank the conflicts? The owner or sponsor, because ranking them is a strategic decision about what the business values.
Questions to ask
- What are this job’s constraints, and what type is each?
- Which constraint will break first?
- What kind of leader does that constraint call for?
- Which conflicts have we ranked, and which have we left to whoever faces them?
- Which parts of the job need detailed planning, which need learning and which need speed?
- What must every approach still prove?
Bringing it together
Deliverables describe what a job produces; constraints describe what kind of job it is. Separate constraints from risks and assumptions, type them, and find the one that will break first. Choose leadership to match it and rank conflicts in writing before delegating, so trade-offs are made deliberately rather than by whoever is under pressure. Then tailor the method to each part of the job, planned, experimental or fast, while keeping the same core accountability. Good tailoring changes how evidence is produced, never whether it is produced.
Source: KEVOS notes, drawing on teaching material on project constraints, tailoring and plan-, change- and time-driven approaches, A. Davies (2019) on adaptive project management for the Association for Project Management, and H. I. Ansoff’s environmental-dependence hypothesis. Examples and figures in this article are illustrations. This article is general information.
