Carlill v Carbolic Smoke Ball Co [1893]: The Company That Won by Losing

Carlill v Carbolic Smoke Ball Co [1893] 1 QB 256, Court of Appeal (England and Wales) Area of Law: Contract Law — Formation, Unilateral Offer, Acceptance by Performance, Consideration

There is something almost comic about the Carbolic Smoke Ball Company. They lost the case. Badly. Unanimously. Three judges, not one dissent, zero sympathy. And yet, within weeks of the Court of Appeal’s decision, their managing director Frederick Roe placed a new advertisement in the Illustrated London News, reframed the entire defeat as proof that the product worked, and the company continued trading. The men were rogues — Lindley LJ himself implied as much — but they were commercially brilliant rogues. What I find consistently underappreciated about this case is not how Mrs. Carlill won, but how narrowly the court defined the winning conditions, and how much it left unsaid about what would have happened if the facts had shifted even slightly. That silence has followed contract law ever since.

What the Court Was Actually Deciding

Four distinct legal questions were before the Court of Appeal, not one. Textbooks often treat this case as being about unilateral contracts, full stop. But the real texture of the judgment is in how the court dispatched four separate lines of attack that the company advanced.

The company argued, first, that the advertisement was mere commercial puffery, a boast, not a promise. Second, that even if it were a promise, it was too vague to be enforceable, because it had no time limit for contracting influenza. Third, that acceptance required prior notification to the company, which Mrs. Carlill had not given. Fourth, that there was no consideration, because she had given nothing in exchange for the promise.

The court rejected all four. But the rejection of the third argument is where the real doctrinal work happens, and where most student analyses spend the least time.

The Facts the Court’s Reasoning Turned On

The Carbolic Smoke Ball Company published an advertisement in the Pall Mall Gazette, a widely read London daily newspaper, on November 13, 1891. The product itself was a rubber ball filled with carbolic acid powder, designed to be squeezed so that the vapour passed through a tube into the user’s nose. The company claimed it could prevent influenza. The advertisement promised £100 to any person who contracted influenza after using the carbolic smoke ball three times daily for two weeks according to the printed directions. Crucially, the advertisement also stated that £1,000 had been deposited with the Alliance Bank in Regent Street to demonstrate the company’s sincerity.

Mrs. Louisa Carlill, a London housewife, read the advertisement. She bought the smoke ball. She used it exactly as directed from November 20, 1891 to January 17, 1892. She caught influenza. She wrote to claim her £100. The company refused.

Two facts mattered above all others in the reasoning: that Mrs. Carlill read the advertisement before performing, and that the company had deposited the £1,000. Strip either of those facts away, and the outcome becomes uncertain. The court never acknowledged how much weight it was silently placing on them.

How the Court Got to Its Conclusion

Lindley LJ delivered the first judgment. His method was systematic. On puffery, he pointed directly to the bank deposit. He reasoned that a company does not set aside £1,000 in a named institution and announce it to the public merely to express confidence in a product. The deposit had no commercial purpose except to signal genuine legal obligation. Therefore the advertisement was a real promise.

On vagueness, Lindley LJ applied a principle of reasonable construction. The absence of a fixed time limit did not kill the contract. The sensible reading was that the protection period ran for a reasonable time after the prescribed two-week usage. Bowen LJ added that Mrs. Carlill had contracted influenza while still within the use period, so the question of what “reasonable time” meant after the period ended did not even arise on these facts. He made the point and moved on.

On acceptance without notification, this is where Bowen LJ did his most important work. He formulated the rule this way: a person making an offer may, expressly or by implication, waive the need for notification of acceptance. In a unilateral offer to the world, notification is impliedly waived because the structure of the transaction makes it unnecessary. The offer is not “I will pay you if you promise to use the smoke ball.” The offer is “I will pay you if you actually use the smoke ball and still catch influenza.” Performance of the condition is simultaneously the acceptance and the notice.

On consideration, Lindley LJ identified two independent bases. The inconvenience to Mrs. Carlill of inhaling carbolic vapour three times a day for two weeks was a detriment sufficient to constitute consideration. And the commercial benefit flowing to the company from increased public use of the product was a benefit to the promisor that also grounded the promise. Either one would have been sufficient.

Where the Reasoning Breaks Down

The court was correct to find a binding contract. I have no quarrel with the outcome. But the reasoning contains two structural gaps that have been generating problems in contract law ever since, and which standard analyses of this case consistently ignore.

The revocation mid-performance problem. Bowen LJ said, almost in passing, that the unilateral offer was “an offer to become liable to any one who, before it is retracted, performs the condition.” Before it is retracted. Think about what that means. On Bowen LJ’s own formulation, the Carbolic Smoke Ball Company could have published a notice the day after Mrs. Carlill started using the ball, withdrawing the offer. If revocation is effective upon communication to the public (as the general rule for unilateral offers holds), then a person halfway through performance could find themselves with no contract and no remedy. The court said nothing about this. Nothing at all. The case did not require them to answer it, because Carbolic never tried to revoke. But the question they left open, whether a unilateral offeror can revoke after the offeree has begun performance, is one of the most contested problems in contract law, and Carlill is the case that created it without resolving it.

English courts would later grapple with this in Errington v Errington and Woods [1952] 1 KB 290, where Denning LJ held that a unilateral offer cannot be revoked once the offeree has entered on performance of the act. But that case relied on Carlill without Carlill having actually decided the point. It was a court building on silence.

The awareness problem. Mrs. Carlill read the advertisement before she performed. The court’s reasoning on acceptance depends entirely on the idea that performance is acceptance. But acceptance requires knowledge of the offer. What if someone had used a smoke ball gifted to them by a friend, without ever seeing the advertisement, and then contracted influenza? They would have performed every physical condition. Would they have a contract? On strict classical theory, no, because you cannot accept an offer you do not know exists. The court’s ratio says “acceptance by performance,” but that formulation only works if the performer was aware of the offer when they performed. Since Mrs. Carlill was plainly aware, the court never engaged with the awareness question directly. The gap remains.

The court was simply wrong to leave these questions unaddressed, and I will explain why that matters: not because it changes the result in Mrs. Carlill’s case, but because every unilateral contract case that came after it inherited the uncertainty.

The Ratio, Stated Precisely

A public advertisement, if it contains a definite promise in clear terms, demonstrates genuine intention to be legally bound, and is capable of acceptance by any member of the public who performs the specified conditions, constitutes a valid unilateral offer. Acceptance occurs upon performance of those conditions without prior notification to the offeror, provided the offeror has impliedly waived notification by the nature of the offer. Consideration exists in the detriment suffered by the offeree in performing the conditions, and independently in any benefit accruing to the offeror from that performance.

That is the ratio. Not “advertisements can be binding,” which is how it is often paraphrased. The ratio is conditional on specific factors: definiteness of terms, demonstrated intention, and implied waiver of notification. An advertisement without those features remains an invitation to treat. The line matters enormously.

Obiter Worth Noting

Bowen LJ made an observation that has shaped the law in ways the case report rarely flags. He distinguished between two categories of public communication: offers to the world on one side, and “offers to negotiate” or “offers to receive offers” on the other. The distinction he drew is that an advertisement which invites a person to enter into a specific, defined transaction is an offer, while an advertisement which merely invites a person to come and deal is not. This is not ratio. It was not necessary to resolve whether the smoke ball advertisement was an offer. But courts have relied on this distinction ever since to classify price lists, catalogue entries, and online product listings as invitations to treat rather than offers.

An invitation to treat, to be clear, is not an offer. It is a preliminary communication that invites another party to make an offer, which the first party may then accept or reject. The distinction matters because only an offer is capable of acceptance. An invitation to treat cannot be accepted directly. It can only attract an offer in response.

The entire modern architecture of commercial offer-and-acceptance traces back to these few sentences in a case about a Victorian cold remedy. Which produces an immediate problem. Every person who adds an item to a cart on Jumia or Konga, enters their payment details, and clicks “pay now” understands themselves to be accepting something. The law says they are making an offer. The seller’s system generating a confirmation is the acceptance. The contract does not form at the moment of clicking. It forms later, if at all.

The doctrinal answer and the commercial reality are running in opposite directions, and Bowen LJ’s obiter is what set them on different tracks.

Where the Invitation-to-Treat Framework Collapses

The classical analysis holds only where the seller retains genuine discretion after the buyer acts. That is the entire purpose of the invitation-to-treat classification. The gap between the buyer’s action and the seller’s response is where the seller’s judgment is supposed to live. The buyer “offers,” the seller decides whether to accept. Amazon, Jumia, Konga — their terms of service almost all say something equivalent to “your order constitutes an offer to purchase” and “we accept when we dispatch or send a confirmation.” That framing is legally deliberate.

But consider digital products that deliver automatically upon payment. No human is reviewing the order. No discretion is being exercised. The system charges the card and simultaneously releases the download or activates the license. The seller eliminated their own discretion by design, and when you eliminate the discretion, you eliminate the commercial justification for the invitation-to-treat classification.

That situation maps onto the vending machine analysis that Lord Denning applied in Thornton v Shoe Lane Parking [1971] 2 QB 163. His reasoning was that an automated machine makes the offer itself, and the customer accepts by inserting coins. The operator pre-committed to the transaction by setting the machine running. There is no invitation waiting for a human offer in response. A digital storefront that charges you and instantly pushes a product has done the same thing. Calling your payment an “offer” in that moment is a legal fiction with no commercial substance behind it.

The error cases expose this most sharply. When retailers have listed products at dramatically wrong prices online and buyers completed checkout, courts have used the invitation-to-treat framing to protect the retailer. The automated confirmation email was held not to constitute acceptance in those circumstances. The buyer who thought they had a contract had nothing. The framework that was built to preserve seller discretion ended up being used to protect sellers from the consequences of their own system errors.

Then there is the refund problem. When payment goes through but the digital product fails to deliver, the seller holds your money during whatever days their internal processing requires. If we take the invitation-to-treat framing seriously, that money was never consideration for any concluded contract. You made an “offer,” they rejected it by failing to deliver, so no contract formed. Under what legal basis are they holding your funds during that window? In practice, nobody frames it that way. The seller calls it a “refund,” which concedes that a contract existed, was performed on your side, and failed on theirs. The language of commerce contradicts the legal classification the moment something goes wrong.

The days-long delay compounds the injury. Your money is gone, the product was not received, and you are waiting on the seller’s timeline with no contractual leverage to accelerate it because the contract formation question is unresolved. Meanwhile the seller collects float on pooled consumer funds across thousands of similar transactions.

The invitation-to-treat doctrine was built for a world of physical goods, human intermediaries, and sequential transactions with breathing room between each step. Digital commerce compressed all of that into a single automated moment. The law has been retrofitting old categories onto new facts ever since, and the fit gets worse as the automation gets faster.

What This Means in Nigerian Law Today

Nigerian courts have consistently applied Carlill as persuasive authority on unilateral contracts and the definition of offer. It is cited routinely in Nigerian legal scholarship on contract formation. The principles on acceptance by performance, implied waiver of notification, and the detriment-based analysis of consideration map cleanly onto the Nigerian common law framework. Nigerian contract law, inherited through the received English common law, does not differ from English law on these fundamentals.

The digital commerce problem is not merely theoretical in Nigeria. Millions of transactions happen daily on Jumia, Konga, Paystack-integrated storefronts, and social commerce platforms where sellers list prices, receive automated payments, and either deliver or fail to deliver without any human touching the transaction sequence. Nigerian courts have not yet squarely resolved whether such a payment constitutes an offer or an acceptance, and the Federal Competition and Consumer Protection Act 2018 (FCCPA), the primary statute governing consumer transactions and unfair commercial practices in Nigeria, does not explicitly address the contract-formation sequence in digital transactions. That gap is open and growing.

But here is the observation that standard Nigerian legal commentary has entirely missed.

If the Carbolic Smoke Ball Company were operating in Nigeria in 2024, they would face a legal problem that has nothing to do with unilateral contracts, and it would be the bigger problem. Section 124 of the FCCPA makes it unlawful for any person, in trade or commerce, to engage in misleading or deceptive conduct. Section 125 specifically prohibits making false representations concerning the efficacy of goods or the standard or quality of goods.

The carbolic smoke ball did not prevent influenza. The company knew or ought to have known this. Under the FCCPA, the claim that the product would immunise users against influenza was a representation as to its efficacy, and that representation was false. Mrs. Carlill’s remedy would not have depended on proving a unilateral contract. She would have had a statutory cause of action from the moment she purchased the ball based on the misrepresentation about what it could do, before she ever contracted influenza.

The implication is striking. Carlill built its fame as a contract law case. In a Nigerian 2024 context, it would be decided primarily as a consumer protection case. The contract doctrine it established would become secondary. And Nigerian legal commentary, by focusing exclusively on the contractual principles, has missed a chance to connect this case to the modern statutory framework that now governs exactly this kind of commercial conduct.

The case of Nigerian Bottling Co. v. Ngonadi [1985] 1 NWLR (Pt. 4) 739 is instructive by analogy. The Supreme Court there applied the neighbour principle from Donoghue v Stevenson to hold a manufacturer liable to a consumer for a defective product. The same logic of protecting the consumer from commercially predatory conduct underlies both that decision and what the FCCPA now enshrines in statutory form. Carlill belongs in conversation with both.

I am not entirely sure the Nigerian Supreme Court would today follow the Carlill analysis on the notice question without modification, particularly in a case where the offeror attempted to revoke mid-performance and the offeree had not yet completed the prescribed act. The revocation problem is real, and Nigerian courts would have to confront it on facts that have not yet arisen.

What studying this case made me think about was not the legal doctrine at all. It was Frederick Roe. The man lost in court, turned the judgment into marketing copy, and kept selling. The company did not collapse under thousands of claims as Finlay warned — almost nobody actually claimed the £100. Which raises a question I find genuinely unresolved: if the risk of mass liability was so small in practice, why did the court expend so much doctrinal energy building a framework for it? Did the judges understand the commercial landscape well enough to know they were creating a rule for an edge case? Or did they think they were deciding something routine?

The court got the outcome right. Mrs. Carlill deserved her money. The reasoning on puffery, on performance as acceptance, on the dual basis for consideration, all of that holds up. But the court wrote as if it were resolving every question in unilateral contract law when it was actually, and only, resolving the narrow question of whether this particular advertisement, on these particular facts, gave rise to a binding obligation. The generalisations it left behind have caused problems ever since.

And the question I leave with you is this: in a world where a Nigerian company launches a viral social media campaign promising a reward to any user who completes a challenge and tags the brand, which thousands begin performing before the company quietly deletes the post, who exactly bears the risk? Carlill tells you the offer was valid. It does not tell you whether the deletion was an effective revocation. That question is still waiting for its court.