Each bid answers a different question: one prices a bed per month with utilities left out, another prices the whole unit inclusive, and a lump sum arrives that never says whether catering sits inside it. The team then adjusts the numbers by hand until they match, and the tender loses its discipline.
The fault is not with the bidders. When the request does not fix the unit of measure, the scope boundary, the duration or what counts as acceptable performance, each bidder fills the gaps privately, with assumptions reasonable to them and inconsistent with everyone else's. The spread came out of the document.
That is cheap to fix while the document is still a draft and expensive to fix afterwards. Once the bids are in, what is left is a clarification round, a rebuild of somebody else's numbers, or a re-tender.
Before the document comes the decision: run housing in-house, or outsource it?
Why bids come back incomparable
A bid is an answer. If the question is loose, the answer is a guess wearing the clothes of a price. In workforce accommodation the gaps recur in the same places, each with its own distortion.
- The unit of measure: per bed per month, per occupied bed, per room, per building, per head per day. Different units cannot be compared without rebuilding both bids.
- Occupancy risk: do you pay for contracted capacity or for beds actually filled? Those are two different products, and a bidder left to guess which one you meant has priced something you did not ask for.
- Inclusions: utilities, consumables, cleaning, security, supervision, maintenance, asset replacement. Anything unstated is priced by one bidder as excluded and by the next as included.
- Duration and volume commitment: a firm term and a rolling arrangement that can be lost at short notice are different commitments, and it is optimistic to expect one price to cover both.
None of that is exotic; any bidder must settle these questions internally before producing a number. If the request does not answer them, you receive prices built on assumptions you will not see until delivery.
What the request has to pin down
Before describing the service, the document must describe the demand it is buying for. A weak request treats this as a covering note; the bidder treats it as the only page that matters.
The demand picture over time
Headcount is the first number most requests send and the one that tells an operator least. Accommodation is a capacity commitment held over time, so what they need is the shape of the curve: how the population builds, where it plateaus, when it falls away. Say which part of that curve is contracted and which is forecast.
- The workforce profile: trades, nationalities where they affect room allocation, any separation by gender, and supervisory grades needing a different standard.
- Locations: where each population is housed, which site it reports to, and whether transport between them is inside the scope.
- The phasing: arrival dates by group, planned demobilisation, and any seasonal swing you already know about.
- The confidence attached to each. Contingency goes where the uncertainty is, and a bidder who cannot tell the firm numbers from the estimates has to treat all of them as soft.
The scope boundary
Managed accommodation, catering, transport and facility management overlap at their edges, and the edges are where disputes live. Who stocks and cleans the kitchen if catering is a separate contract? Who maintains the building fabric, and who the assets inside the rooms? At every interface, state which party performs, which pays, and which answers when it fails.
- Building fabric and fixed elements versus furniture, appliances and bedding in the rooms.
- Utilities: supply, metering, consumption, and who carries the summer peak.
- Catering: kitchen, storage, meal service, and the standard it is judged against.
- Transport: vehicles, drivers, scheduling, and driver hours when shift patterns change.
- Security and access control. Then the resident register on its own: who updates it, and how long after someone moves.
- Waste, pest control and external areas, which drop out of documents and return at inspection.
Specify the outcome, not the solution
Weak documents tend to describe a building: the property type, the room layout, the fittings, sometimes the make of an appliance. The effect is that you designed the solution, own the risk it works, and ruled out anyone with a better way.
Write what a resident must experience and what you must be able to prove, then let bidders propose how; the differences between them become informative rather than noise. A service level is complete only when it states what is expected, within what window, and what follows if it is missed.
- Cooling, water and power: availability expected, response time by severity, and what follows a breach.
- Cleaning and waste: frequency by area type, and the standard the result is judged by, not a cleaner headcount.
- Maintenance: the split between planned and reactive, and the closure evidence you accept.
- Complaints: a channel residents can use in their own language, a response clock, and an escalation path with names on it rather than job titles.
- Occupancy administration: check-in, check-out, room moves, and an accurate register available on request, not at month end.
Records deserve their own clause, because they are the only part of the service you can verify after the fact. State what you receive and how often: occupancy against contracted capacity, maintenance raised and closed with ageing, complaints and outcomes, and permit status by building. Say who owns them, too: an operator who keeps your housing data and hands over nothing at exit has built a switching cost you never agreed to.
The pricing schedule: making the numbers line up
If bidders choose the shape of their commercial submission, you will spend the evaluation rebuilding it into something comparable, and you will make errors doing so. Issue the pricing schedule as a locked structure, require it back unchanged, and keep commercial commentary in a separate document outside the evaluated price.
- Define the unit precisely and use it throughout, including whether it is charged on contracted capacity or on actual occupancy.
- Separate one-off charges from recurring ones. Fit-out, mobilisation and demobilisation must not dissolve into the monthly rate; once they do, they disappear from any comparison run over a different term.
- List every inclusion as its own line, even those you assume are bundled, with a mark against each: included, excluded, or priced separately.
- Ask for the rate that applies when volumes move either way, and the notice that triggers it. Flexibility is priced in advance or bought later under duress.
Require the price to hold for a stated period and say what adjusts it afterwards. A schedule built this way separates the bidder who is cheap because they are efficient from the one who is cheap because they left something out. Whether the price itself comes down is a separate question.
For the items that actually move the cost
The clauses that get left out
These are the easiest clauses to leave out of a first draft, and the most expensive to negotiate after award, when you have nothing left to trade with.
- Mobilisation: what must be true on arrival day, who proves it, and what happens if it is not.
- Scaling: the mechanism, notice period and price, in both directions rather than only upwards.
- Inspection rights: unannounced entry, bringing your end client or an auditor, and the records behind what you are shown.
- Rectification: how a failure is raised, how long the operator has, and what follows if the clock runs out.
- Records handover: what you receive periodically and at exit, in what format, and who owns it.
- Exit and transition: notice, the condition buildings are returned in, continuity of accommodation, and cooperation with an incoming operator.
- Termination triggers tied to compliance and safety, not only to payment.
Exit is the clause worth insisting on hardest, because it keeps the contract honest for its whole life. An operator who knows you can leave cleanly behaves differently from one who knows you cannot.
Set the evaluation criteria before the bids arrive
Criteria written after the submissions are opened are not criteria; they are a justification. Once the prices have been seen everyone is anchored, and the weightings quietly rearrange themselves around the answer someone already preferred.
Agree the model while the document is still in draft: what is scored, at what weight, what evidence earns a score rather than asserts it, and what is a pass-or-fail gate before the commercial envelope is opened. Compliance evidence, resident welfare, on-site supervision and the credibility of the mobilisation plan should each be able to fail a bid alone.
Decide in advance how to treat a price far below the rest. A low bid is information, not a prize. Ask the bidder to walk through the assumptions behind it and test them against the scope boundary you wrote. Where the gap turns out to be a service they excluded, you have found the real difference between the offers, and it was never price.