Workforce Housing Demand and Capacity Planning

Workforce accommodation is easy to treat as a purchase: headcount is known, a price per bed is requested, a contract is signed, and the file closes until renewal. That works while supply is loose, while there is always another building to fall back on.

It stops working the moment several employers need capacity in the same region inside overlapping windows. When construction and industrial activity concentrate into the same corridors at the same time, accommodation stops being a commodity bought on demand and becomes a constrained resource secured ahead of need.

Nobody holds a reliable read on when supply in a given region will tighten, and a plan that waits for one is built on luck. Planning, though, does respond to effort. What will you need, where, when, and how firmly are you willing to commit before you know?

When accommodation stops behaving like a commodity

A loose market gives a buyer what it never charges for: choice, time, and the ability to walk away. A tight one withdraws all of it, in an order that catches planners out. Availability tends to go first; price moves last. So a buyer watching price reads a comfortable market right up to the point where nothing worth taking is left. Watch how much capacity within reach of the site is still uncommitted. Cost is the lagging indicator.

Tightening is usually local and temporary. Mobilisation clusters, and a corridor that has filled tends to clear when the teams occupying the stock demobilise rather than when anyone builds more. That leaves planning with an uncomfortable consequence: capacity has to be secured while the need for it is still uncertain. Postponing the decision does not remove the uncertainty, it just removes the choice.

Headcount gives a flat number; a pipeline gives a curve

Most accommodation plans start from headcount: a known workforce, an allowance for growth, a bed requirement derived from the two. The output is a flat number, and a flat number is wrong in every period. It runs too high while scopes sit between awards and too low when several mobilise together. Averaging a curve produces a figure that describes no month you will actually live through.

Planning from the pipeline produces the curve. The inputs are scopes of work: each one has a location, a ramp, a plateau, a release date, and a probability that it happens at all.

1. List every scope, including the ones you have not won yet. Work under tender and work you merely expect belong on the same sheet. Give confidence its own column and let everything onto the sheet.
2. Locate each scope precisely enough to matter: the site or the corridor rather than the region. Two scopes in one region can draw on completely different accommodation markets.
3. Give each scope a shape over time instead of a single number: when people arrive, how steeply the ramp builds, how long the plateau holds, when release starts.
4. Weight each scope by likelihood, and keep the weighting visible so that anyone reading the curve can see which part of a peak rests on work you have not won.
5. Add the shapes together by location and by period. The sum is the curve.

The curve is a shape, and the shape is the useful part: where the peaks fall, which of them hang on the same uncertain award, and what is left standing once the speculative scopes are stripped out. Nobody should read it as a forecast.

The same picture from inside a single scope: worker housing on giga-project scopes→

Base load and surge

The curve separates into base load and surge. Base load is the portion present in nearly every scenario, including the pessimistic one: what you still need if you win nothing new and something already awarded slips. Surge is everything above that line. It is real demand, conditional on the pipeline converting.

The two are worth buying with different instruments, and treating them alike is where money leaks. Base load is stable and long-lived, so it earns better unit economics and an operating relationship worth building; a longer term on it buys you something. Surge is uncertain in size, in timing and often in location, so pay for flexibility with your eyes open. The premium on a short, releasable commitment is the price of an option, and options cost money.

A long lease sized to the peak is usually the worst structure available, and a run of spot purchases is barely better. The lease pays for empty beds through every trough. Spot buying lands at the worst possible moment almost by definition, because your moment of need is everyone else's in the same corridor. A portfolio beats both: anchored capacity for base load, contracted expansion rights for the predictable part of surge, and a small tolerance for spot exposure. It works because each instrument is matched to how certain the demand behind it is.

Who carries that structure: running housing in-house versus outsourcing it→

The risk runs in both directions

Over-commitment is the visible risk: capacity is contracted, an award slips or is lost, and you pay for beds nobody sleeps in. The subtler cost arrives afterwards, when the committed asset starts shaping decisions it should not. Teams begin placing people where the capacity sits instead of where the work is, and nothing in the accounts records that as a cost.

The opposite error costs nothing until the week you actually need the beds. Arriving late in a corridor that has tightened means taking what is left instead of what was specified, accepting a building further from site, carrying the transport bill for the life of the scope, and negotiating all of it with a counterparty who knows your mobilisation date.

Forecasting harder solves neither. Scope-level accuracy is rarely available to anyone, and a plan that depends on it is fragile by construction. The risk is managed in the contract instead, by deciding in advance what happens when the plan turns out wrong.

• Phasing tied to project milestones rather than calendar dates, so capacity switches on when the work does.
• Expansion rights priced in advance, which turn a future negotiation in a tight market into a decision that is already yours to make.
• Step-down and release provisions, so a scope that ends early or never starts can be exited without paying it in full.
• Notice periods that mirror your own client contract rather than the accommodation market's defaults.
• Clarity on who carries the cost of idle capacity, agreed while both parties still expect it to be occupied.

Demand clusters, and national averages hide it

National supply figures are close to useless for planning a single site. Your market is narrower than that: the buildings within an acceptable travel time of the site, operated to a standard your client and the competent authorities will accept. It can be tight while the national picture looks comfortable, because activity concentrates into corridors.

Concentration is also why competitors matter in this decision when they do not matter in most procurement. Where purpose-built stock is deep, another employer mobilising nearby barely registers. Where it is thin and much of what exists was converted from something else, one large mobilisation by somebody else can absorb the supply you were quietly assuming would be there.

Assessing local depth means asking narrower questions than a national view suggests. Who in the corridor can actually operate accommodation, as opposed to offering you a building. How much of the stock was built for shared accommodation and how much was converted into it. Whether the fit-out supply chain exists locally or has to be brought in. And what else is mobilising into the same corridor in the same window.

An industrial corridor with its own dynamics: worker housing in Jubail and Yanbu→

Review triggers, and who owns them

The curve rests on assumptions that change constantly: awards land or do not, schedules move, scopes grow, clients defer. A curve built once and filed is worse than no curve at all, because it carries the authority of analysis while describing a pipeline that no longer exists. Review works better when pipeline events trigger it than when the calendar does.

• An award won or lost, which changes the shape of the curve and the confidence weighting of everything adjacent to it.
• A material schedule revision, since a slipped mobilisation moves a peak into a different window and often a different market condition.
• A scope change affecting headcount or location. This is the trigger most often missed, because it reaches you as a technical variation and nobody reads it as an accommodation matter.
• Any approaching option, notice or renewal date, at which flexibility bought earlier is either exercised or wasted.
• A shift in local supply: a competitor mobilising, stock leaving the market, an operator withdrawing from the corridor.

The review needs an owner, and ownership is where most of these plans quietly die. Accommodation demand originates in planning, gets bought by procurement, and is lived with by HR and site management. It falls in the gaps between the three and ends up belonging to nobody. Naming one person who holds the curve and can trigger a commitment decision is a smaller change than it sounds, and it is usually the change that makes the rest of this work.

None of this makes a pipeline predictable. What it does is let you commit to capacity while the pipeline is still uncertain, and unwind the commitment when the pipeline changes. That is the part of the problem you control.