When a Heating-Oil Quote Stops Being a Promise
The Competition and Markets Authority’s heating-oil intervention, published on 28 August, concerns a relatively small number of households. Its significance is much larger.
The regulator found that hundreds of customers who had ordered oil through an intermediary site saw those orders cancelled when prices surged and supply tightened earlier this year. Suppliers have agreed to compensate customers who then had to buy replacement fuel at higher prices, or to honour original orders where no replacement was bought.
For a typical 500-litre order, the CMA estimates that the price rose from about £320 in February to roughly £520 in March. That was a sharp wholesale shock, not a simple case of retailers inventing margin. Yet customers had made a purchase decision at one price and then encountered a different commercial reality after the transaction.
That distinction is where the wider lesson lies. In volatile markets, a price shown to a customer cannot be treated merely as a moving data point. It has to be connected to stock, credit, delivery capacity and the point at which the business accepts contractual risk.
A pricing engine is also a promise-management system.
A refund is not a substitute for certainty
Heating oil is an unusually revealing market. Demand is seasonal, buyers are dispersed, deliveries are physical and urgent, and about 1.5 million UK households rely on kerosene for heat and hot water. Customers cannot always defer the purchase, switch product or travel to a competitor. When supply is constrained, the cost of a cancelled order is not simply the time spent finding another seller; it can be a cold home and a much higher replacement price.
The same structure exists elsewhere. Building materials, replacement parts, medical supplies, event tickets, freight capacity and even certain digital services can be sold through systems where availability changes faster than the customer-facing proposition. An intermediary may collect the order, while a distributor manages stock and a third party performs fulfilment. At normal levels of volatility, these arrangements appear seamless. Under pressure, their hand-offs become visible.
A refund closes a payment loop. It does not automatically address the customer’s exposure to the market between ordering and replacement. In the CMA case, compensation recognises that the customer’s loss was created by that gap.
Businesses should take notice because customers increasingly understand this intuitively. They know that prices move. What they do not accept readily is being invited to transact, receiving apparent confirmation, and then discovering that the seller reserved the right to revisit the price once the market moved against it.
The quote needs a contract model behind it
Many organisations have invested heavily in real-time pricing. They have invested less in defining what a quote means operationally.
There are legitimate commercial differences between an indicative price, a timed quote, an order request, an accepted order and a guaranteed allocation. The problem comes when the customer journey obscures those differences. A button labelled ‘buy’, a payment authorisation and an email headed ‘order confirmation’ all create expectations, whatever the underlying terms say in the small print.
The remedy is not to freeze prices indefinitely. It is to design a clearer risk model.
First, businesses need a single definition of acceptance across their website, contact centre, intermediary partners and fulfilment teams. If a price depends on supplier confirmation or available transport capacity, that limitation should be stated before payment rather than revealed afterwards.
Second, commercial teams should identify which costs are genuinely variable after an order is placed and which risks the business is prepared to absorb. A modest contingency margin or a capped re-pricing rule may be cheaper than widespread cancellation, remediation and reputational damage during a spike.
Third, intermediaries need explicit responsibility for the customer outcome, not just for lead generation or checkout conversion. The customer does not experience a fragmented supply chain; they experience the brand that took the order. Contracts between platforms, suppliers and delivery partners should therefore set out escalation routes, data-sharing expectations and who funds redress when fulfilment fails.
Research must observe behaviour under pressure
There is also a market-research lesson. Customer satisfaction studies conducted in stable periods will rarely reveal the real strength of a pricing proposition. People may say they value choice, convenience or a competitive headline rate. During disruption, they tend to value certainty, explanation and the ability to act.
That means research teams should test the moments that operations teams prefer not to discuss: a delivery date moving, a basket repricing, a stock allocation being withdrawn, or a supplier failing to meet a committed service level. These are not edge cases if a category is exposed to commodity prices, weather, global logistics or constrained infrastructure. They are foreseeable stress conditions.
Useful research should combine transaction data with complaint reasons, call-centre transcripts, abandoned orders and customer-service recovery outcomes. It should distinguish between customers who object to a price rise and those who object to the way it was communicated. Those are commercially different problems. The first may be unavoidable; the second is often designed into the process.
The CMA’s heating-oil action should not be read as a niche story about rural fuel. It is a reminder that digital commerce has made pricing immediate, but has not removed the need for firms to decide when an offer becomes a commitment.
In a volatile market, that decision belongs in product design, procurement, legal terms, customer operations and board-level risk management. If those functions do not agree on what a confirmed order means, customers will supply their own interpretation — and regulators may eventually do the same.
Comments