Approach
Decisions that set the cost are made before anyone breaks ground
Industrial projects are rarely lost on site. They are lost in the first fraction of the schedule, where the scope is still soft, the lead times are still assumed, and the contract is already being drafted. Our work sits in that window — and in the governance that has to hold afterwards.
Entry decision · Should this project be taken at all, on these terms?
A fixed-price contract is a capital allocation decision with asymmetric downside. It deserves the arithmetic any capital decision gets: quantified exposure before the bid, assumptions stress-tested against real supply-chain and productivity scenarios, and a process with enough authority to say no when the risk profile does not match the margin — including when the client matters and the pipeline is thin.
Selectivity is not timidity. It is the most expensive discipline to acquire and the cheapest one to own.
What you getA bid / no-bid position with the numbers behind it, and the conditions under which the answer would change.
Definition before price · What exactly is being built?
A scope defined to seventy per cent at contract award is not a project with a thirty per cent gap. It is a project with open-ended cost exposure, and the price agreed on that basis is a floor rather than a ceiling. Money withheld from early engineering is the most expensive money on the project: it returns as rework, change orders and standing time, at a multiple, before mechanical completion.
What you getA scope defined to the level where it can be priced, a work breakdown into packages with boundaries that hold, and an interface matrix that names who owns each seam.
Risk · Do the critical assumptions survive contact with the people who deliver them?
A register filled in over two hours to satisfy a governance requirement produces a document. A working session that puts engineering, procurement, construction and operations in the same room produces decisions. We run the second kind, and we write down what was decided, not what was listed.
What you getThe assumptions that failed the test, the decisions taken in response, and the ones deliberately left open with an owner and a date.
Long-lead equipment · Is the schedule built on fixed lead times or assumed ones?
A schedule built on assumed lead times is fiction. Transformers, large rotating equipment, pressure vessels and heat exchangers are ordered against a delivery market, not a plan, and qualifying an alternative source takes months of technical audit and trial orders, not weeks. Where delivery exceeds the schedule, that is a go / no-go input, not a detail for later.
What you getA concentration map of the critical equipment categories by source geography, lead time and alternative-source status; and conditional commitments placed early, where the cancellation cost is a fraction of the delay it prevents.
Assurance · Would an outside reader with no stake call this project ready?
The cheapest moment to stop a bad project is before it starts. The second cheapest is at the end of pre-FEED — not after forty per cent of the capital is committed. That is the whole argument for an independent review reporting to the owner rather than to the project.
What you getAn assurance report written to be acted on: what is ready, what is not, and what has to be true before the investment decision is taken.
Governance in execution · Does the reporting describe the project or describe the reporting?
Optimism bias is not deception. It is what happens when each level makes one small, defensible adjustment: the subcontractor to the contractor, the contractor to the owner's team, the team to management, management to the board. No one lies and the picture still drifts.
The correction is structural — data taken from the project systems rather than through layers of interpretation, escalation triggered by risk rather than by a fixed percentage, and independent review at the milestones rather than only at sanction.
What you getOne page that shows the state of the project, a baseline that is not quietly rebased, and escalation protocols that define what reaches the board and what stays with management.
Three questions before signing an EPC contract
A short diagnostic. Any answer of “no”, or “we will deal with it as we go”, means the project is already in trouble.
- Is the scope defined above ninety per cent — or are we pricing at sixty to seventy?
- Has there been a real risk workshop with engineering, procurement, construction and operations in one room — or only a register?
- Are the lead times for critical equipment fixed by commitment — or assumed by the planner?
Safety is designed in, not added afterwards
Fire safety, process safety and environmental constraint are project conditions, not a compliance layer applied at the end. Designed in, they shape the layout, the packages and the contract. Added afterwards, they arrive as rework at the least forgiving moment in the schedule.
On a large facility several bases apply at once — the national code of the jurisdiction, the licensor's package, the client's corporate standards and the insurer's requirements. Establishing which one governs where they diverge is part of the engineering work, not a formality. We never claim compliance with a specific standard without the document that proves it.
A facility can pass every audit and still carry a systemic exposure. Lagging indicators alone will not show it; the state of near-miss reporting and whether commercial pressure is quietly eroding a standard will.
A safety requirement written into the contractor's scope is enforceable. The same requirement issued as a comment on a drawing is a negotiation.