A programme award is the customer's forecast, not our order book

When a vehicle maker chooses a supplier for a new model, it sends a nomination letter. The letter names the part, the assembly plants that will receive it, the price for each year of the programme and an estimate of annual volume. Suppliers celebrate these letters, announce them to investors and add them to a figure usually called booked business. Then they buy tooling, lease or build floor space, hire and train people, and wait for the vehicle to launch.

Almost everything in that sequence treats the volume estimate as a commitment. Almost nothing in the letter makes it one. Purchasing terms in our industry generally let the customer call off what it needs week by week, and most nominations state plainly that volumes are for planning purposes. The supplier's capital is fixed on the day it is spent. The customer's demand stays variable for as long as the programme runs. I learnt this on a minivan launch in 2000, and the industry relearnt it on electric vehicles between 2023 and 2025.

The quoted volume is someone else's hope

At Tallisford we were nominated in 2021 and 2022 to supply battery cooling plates and chillers for seven electric programmes from three customers. Added together, the quoted volumes implied about 2.1 million vehicle sets a year by 2026. Building for that would have needed three new plants and roughly US$470 million of capital, more than twice our usual annual capital budget. Our commercial team had worked hard for those awards, and our customers wanted to see concrete.

By 2025 the seven programmes were running at 38 per cent of their combined quoted peak. One had been cancelled, two had slipped by more than a year, and two customers had moved part of their plans from fully electric vehicles to plug-in hybrids. None of this was dishonest. Each customer's estimate reflected its own plans, which in turn reflected expected consumer demand, battery costs, charging networks and federal incentives, all of which moved, including the end of the federal purchase credit for new electric vehicles in September 2025. A supplier cannot forecast those things better than the vehicle maker. It can only decide how much of the error it will carry.

The error is not symmetrical. If volume arrives above the quote, the supplier earns overtime margins for a while and the customer asks for a price reduction to share them. If volume arrives below, the supplier carries depreciation on equipment that is half used, and the piece price, which was calculated by spreading fixed cost over the quoted volume, no longer covers that cost. Across a portfolio of programmes, a supplier that builds to quote is handing its customers a free option on volume.

Price the shortfall before you pour the floor

The answer is not to refuse awards. It is to write the risk into them. In 2022 I recommended, and our board approved, three changes to how Tallisford commits capital to new programmes.

The first was staging. We built one plant, in Tennessee, sized for the programmes most likely to launch on time. We bought land and permits for a second in Kentucky but did not build on it, and we dropped the third. Each later step required evidence: a customer's production schedule rather than a planning estimate, or a contract term that moved the risk.

The second was a shortfall clause. For new awards we asked customers to agree that if annual volume fell below 70 per cent of the quoted figure, they would pay either a higher piece price or a lump sum recovering the unamortised cost of equipment dedicated to their programme. Some customers accepted. Some accepted a weaker version that covered tooling but not buildings. One refused, and we declined that programme, worth about US$110 million a year at its quoted peak. A competitor took it, and I do not know yet whether that competitor will be glad.

The third was a change in what we call booked business. Awards now reach our board at three volumes: the customer's quote, our own estimate and a low case at half the quote, each with the capital it requires and the return it earns.

The results were mixed in the way real decisions usually are. Between 2023 and 2025 we avoided about US$230 million of capital that would have stood idle, and recovered US$41 million under shortfall and tooling terms. But the Tennessee plant was itself built for our optimistic case and ran at 52 per cent of capacity in 2025. Staging protected us from two plants. It did not protect us from our own enthusiasm about the first.

What a board should see beside an award

Directors of supplier companies usually meet awards as good news, presented in aggregate: lifetime revenue booked this year, new programmes won, share of a growing market. I would ask them to look at three other numbers before approving the capital that follows.

The first is the volume at which each programme repays its dedicated capital, set against the quoted volume. If a programme needs 90 per cent of quote to break even, it is a bet that the customer's estimate is nearly right, and the paper should say so in those words.

The second is the share of dedicated capital protected by contract. Tooling that the customer owns or pays for is protected. A building, a brazing furnace or an end-of-line test rig usually is not. A board approving US$185 million for a plant should know how much of that sum a customer has agreed to share if the vehicle sells badly.

The third is concentration by technology as well as by customer. Our seven electric programmes in 2022 came from three customers, but they shared one assumption about how quickly North American buyers would adopt battery vehicles. Spreading the awards across customers did nothing to spread that assumption.

None of this is comfortable for a supplier's sales organisation. Customers resist shortfall terms, competitors may not ask for them, and a declined award is visible in a way an avoided idle plant never is. But a supplier that cannot say no to an estimate is not managing its capital. It is lending its balance sheet to its customers' product plans, interest-free and unsecured. The award letter deserves its celebration. It is not the order book, and a board should never be asked to approve concrete as though it were.