A handshake over dinner might be a contract. A signed, notarised 40-page agreement might not be. Here are the six things that actually decide it.
Two founders shake hands over dinner and agree to split equity 60/40. Nothing is written down. Eighteen months later, one of them is holding a term sheet and the other is holding a grudge, and a court is being asked whether that dinner produced a contract. Meanwhile, three floors up in the same building, a procurement team has a beautifully drafted 40-page supply agreement, signed by both parties, notarised, with a gold seal on the cover page, that a judge will hold unenforceable because the person who signed for the supplier had no authority to bind it.
The handshake might be a contract. The 40-page document might not be. That gap is what this article is about.
Most people assume a contract is defined by its form: a written document, signed, ideally lengthy. It isn't. A contract is defined by whether six specific things are present. Get all six and you have an enforceable agreement, whether it is a 200-page merger agreement or a text message. Miss one and you have a document, which is a very different thing to own when the relationship goes wrong.
Every enforceable agreement rests on the same foundation. The terminology varies slightly between jurisdictions and textbooks, but the substance is consistent across common law systems:
The order matters less than the completeness. Courts do not award partial credit. An agreement missing consideration is not 83 percent enforceable.
An offer is a clear proposal of terms, made with the intention that acceptance will create a binding agreement. The test is objective: would a reasonable person in the recipient's position understand that agreement would close the deal?
The distinction that trips people up is between an offer and an invitation to treat. A price list, a catalogue, an advertisement, a shop window display, and most requests for proposal are generally invitations to treat, not offers. They invite others to make offers. This is why a retailer is not obliged to sell you a television mispriced at £2 instead of £2,000 — the display was an invitation, your presentation at the till was the offer, and the retailer is free to decline it.
Acceptance must mirror the offer. This is the mirror image rule: accepting on varied terms is not acceptance at all, it is a counter-offer, which rejects the original and puts a new one on the table.
This is precisely what causes the battle of the forms. A buyer sends a purchase order on its standard terms. The seller acknowledges on its own standard terms, which conflict. Neither party reads the other's boilerplate. Goods are delivered and paid for. When something goes wrong, the question of whose terms govern becomes genuinely difficult, and different jurisdictions resolve it differently — some apply a last shot rule where the final set of terms sent before performance wins, others knock out the conflicting clauses and substitute default law. If you have ever wondered why whether a purchase order is a contract is a more complicated question than it sounds, this is why.
A party generally cannot impose acceptance through inaction. "If I do not hear from you by Friday, I will assume you agree" does not create a contract on Friday. The exceptions are narrow: a prior course of dealing between the parties, or conduct that unambiguously signals acceptance, such as beginning performance.
Consideration is the element most often missing in agreements that feel binding but aren't. It means each party provides something of legal value — money, goods, services, a promise to act, or a promise to refrain from acting they were otherwise entitled to take.
Courts generally do not assess whether consideration is adequate. Selling a building for £1 can be perfectly valid, because the law enforces bargains rather than policing whether they were good ones. But consideration must be sufficient, meaning it must be real and must have some value in the eyes of the law.
Note that jurisdictions with civil law traditions, and agreements executed as deeds in England and Wales, can dispense with consideration entirely. A deed is binding because of its form, which is one reason guarantees and certain property transfers are executed that way.
Capacity asks whether the parties were legally capable of committing. Minors, individuals lacking mental capacity at the moment of agreement, and those severely intoxicated to the point of not understanding the transaction may be able to void agreements they entered.
In commercial practice the more common failure is corporate authority. An entity acts through people, and not every person can bind it. A sales representative may have apparent authority to agree standard terms but no actual authority to accept an uncapped indemnity. A director may need board approval above a certain value. When a signature is challenged, the question is whether the signatory had actual authority, or apparent authority the counterparty reasonably relied on.
This is why diligence requests ask for board resolutions and delegation-of-authority matrices. A signature from someone without authority produces a document whose enforceability against their own entity is uncertain, which is often a worse position than no document at all.
An agreement to do something illegal is void, and courts will generally not assist either party. Beyond outright criminality, agreements can fail on public policy grounds: unreasonable restraints of trade, clauses attempting to exclude liability for death or personal injury caused by negligence, or terms prohibited by consumer protection or competition law.
Partial illegality does not always destroy the whole agreement. Courts may sever an offending clause and enforce the remainder, particularly where a severance clause exists and the illegal portion is not central to the bargain. A non-compete drafted too broadly may be narrowed or struck while the rest of the employment agreement survives.
The parties must have intended to create legal relations and must have agreed on the same thing. Courts apply an objective test — what the parties outwardly manifested, not their private mental states.
Intention is presumed in commercial dealings and presumed absent in social and domestic arrangements. An agreement between friends to share a lottery ticket may or may not be binding depending on how formally it was treated; an agreement between two companies almost certainly is. This presumption is why letters of intent, heads of terms, and memoranda of understanding require careful drafting: they often mix binding provisions (confidentiality, exclusivity, governing law) with non-binding commercial intentions, and the document must say which is which.
Agreements affected by these are typically voidable rather than void, meaning the wronged party can choose to rescind or affirm.
Most contracts do not need to be in writing. Oral agreements are generally enforceable, and email exchanges regularly form binding contracts without anyone intending ceremony.
But statute of frauds provisions, in their various national forms, require writing for certain categories. These commonly include transfers of interests in land, guarantees of another's debt, agreements that cannot be performed within a year, and in some jurisdictions sales of goods above a threshold value. Get one of these wrong and you have an agreement that satisfies all six elements and is still unenforceable for want of form.
The practical case for writing has little to do with validity anyway. Writing is about evidence. When a dispute arises three years later, the question is rarely whether a contract existed. It is what its terms were, which obligations survived, and who agreed to what. That is a documentation problem, and it is why a contract's life after signature matters as much as its formation — as covered in what an executed contract actually means.
Offer, acceptance, consideration, capacity, legality, and mutual assent. Some formulations list four or five by combining offer and acceptance into "agreement" or folding mutual assent into acceptance, but the substantive requirements are the same. All must be present simultaneously. An agreement that satisfies five of six is not partially enforceable — it generally is not a contract at all, though separate doctrines such as promissory estoppel may occasionally provide a remedy where a party reasonably relied on a promise to their detriment.
Usually yes. Most contracts require no particular form, and an oral agreement containing all six elements is enforceable. The exceptions are categories covered by statute of frauds provisions, which commonly include land transactions, guarantees, and agreements incapable of performance within a year. The real difficulty with oral contracts is evidential rather than legal: proving the precise terms years later, when both parties remember them differently and neither has a record, is extremely hard.
A void contract never existed in law — typically because its purpose was illegal or an essential element was entirely absent. A voidable contract exists and is enforceable unless the affected party chooses to rescind it. Misrepresentation, duress, undue influence, certain mistakes, and lack of capacity generally make an agreement voidable rather than void. The distinction matters because a voidable contract can be affirmed, and the right to rescind can be lost through delay or by continuing to take its benefits.
In common law systems, yes: each party must give something of legal value, which is what distinguishes a bargain from a gift promise. Adequacy is not assessed — nominal consideration is generally acceptable — but the consideration must be real, must not be past, and must not merely repeat a pre-existing obligation. Agreements executed as deeds are the main exception in England and Wales, as deeds are binding on their form alone, and many civil law jurisdictions have no consideration requirement at all.
It depends on the entity's constitution, its delegation-of-authority policy, and applicable company law. Directors typically have broad authority, officers have authority within their remit, and employees have whatever authority has been actually or apparently conferred. A counterparty can often rely on apparent authority where the signatory was held out as able to bind the company, but relying on that is risky for high-value agreements. Requesting evidence of authority — a board resolution, a signing matrix, a power of attorney — is standard practice above material thresholds.
There is generally no enforceable contract, and the parties are left to other remedies. Restitution may recover the value of benefits conferred, promissory estoppel may protect a party who reasonably relied on a promise, and a quantum meruit claim may recover reasonable payment for work performed. None of these give you the bargain you thought you had. They give you something considerably less, which is why confirming the elements before performance begins is far cheaper than litigating them afterwards.
HERO is a document editor built for agreements that have to hold up — where defined terms, cross-references, and clause structure stay intact rather than flattening into formatted text the moment you export. Teams drafting contracts in HERO can see which provisions depend on which, catch a definition that broke when a section was renumbered, and keep the reasoning behind a negotiated term attached to the term itself. Book a demo and bring your most heavily amended agreement.