One decision, in full.

A 3.25 MTPA brownfield liquefaction export facility being taken to a final investment decision inside a live operating regasification terminal. I was the single named person accountable for getting it there.

COVID-19 arrived, and across the industry final investment decisions were postponed or canceled. The blocker was rarely the virus itself. It was that nobody could put a number on it: the EPC contractor and its suppliers would certainly claim a Force Majeure event, and the exposure was open-ended. A board cannot sanction against an unbounded number, so the default everywhere was to wait.

The obstacle was how the risk was being transferred, not the size of it. A lump-sum settlement asks the owner to buy an unknown at a fixed price, and asks the contractor to accept a figure that may not cover them. Neither side can sign that, so the project waits. What nobody had done was price COVID as what it actually was: a cost that would rise, peak and then decay, at a rate that could be observed month by month.

We negotiated the claim before it happened. There was no precedent for executing an EPC contract through a pandemic, so neither party had a basis on which to price the consequences — a number could not be agreed, but a mechanism could.

We ran joint qualitative risk workshops with the contractor, with shared responsibility for the mitigations. That put both teams on the same problem rather than on opposite sides of it — set the contractual argument aside, and two organizations working a risk together get further than either will alone.

Once that was agreed, we negotiated the structure. First we separated the direct cost and schedule impacts — people getting sick — from the indirect ones: the loss of economic and productive activity caused by the preventative measures themselves. Then every cost was sorted into three categories, and each was paid for differently.

• One-off costs, committed once and not recoverable if the pandemic ended — paid at pre-agreed milestones.
• Time-related costs, which scale with how long the preventative measures run and how many people are on site: transport, PPE, cleaning, social distancing, back-up crews for those quarantined — paid monthly against a forecast curve.
• Reimbursable costs, payable only if and when actually incurred — COVID-19 testing.

That confined the exposure to the period the measures were genuinely needed. It also carried a ramp-off: as the demonstrated impact on the contractor fell, the monthly payments reduced and could be stopped, at which point we carried the residual risk ourselves.

It asked the contractor to accept an exposure nobody in the world could quantify, on a job nobody had done before. It had to be built with them rather than presented to them, and it had to be built by people who understood both the commercial instrument and the execution work it was pricing.

The only LNG project in the world to reach FID in 2020. Payments under the mechanism stopped at 25 per cent of the project schedule, at a total cost of about 1 per cent of project cost.

Being the only project in execution proved worth considerably more than that 1 per cent. Neither the contractor nor our own team had competing priorities anywhere else. Supplier shops had spare capacity because market demand had collapsed, which showed in both quality and schedule, and site labor was available where others had stood crews down. COVID absenteeism on site stayed below 1 per cent, so productivity never suffered for want of people, and detailed engineering and procurement ran ahead of schedule.

Second-order advantages followed. We took engines canceled by the collapse in aviation to get our aeroderivative mechanical drives early. And the equipment carrying Russian and Ukrainian content was already fabricated, so the project was untouched by the supply shock two years later.

About 1 per cent of project cost, to convert an unbounded exposure into a bounded one that could be switched off. What it set was a precedent: a claim agreed in the contract before the event that would give rise to it, priced as a curve rather than a number.

I would not do it differently. We reached FID, we paid less than we had assumed, and no further COVID claim was ever made — which is the evidence that stopping when we did was right rather than lucky.

Stopping was not a judgment call made in a meeting. It came out of the same internal qualitative and quantitative risk process I use for most key project decisions, and that is the point of using one: it gives the decision a structure and a methodology rather than an opinion. The conclusion was not open to argument, because the process it came from is the company's own.