Few clauses absorb as much negotiation time as the liability cap, and few phrases are deployed in that negotiation as confidently as "market standard". The phrase does real work: it suggests a settled norm the other side is unreasonably departing from. It is worth knowing what the phrase can honestly describe — and where it is simply a negotiating posture — because the difference decides whether you are arguing about convention or about risk.

The honest version is this: there are recurring structures, and there are recurring carve-outs. There is no single number. Anyone who tells you that one specific multiple is "the" market standard across commercial contracts is describing their template, not the market.

What the cap does — and what sits around it

A liability cap sets the maximum a party can be made to pay for breaches of the contract. But the cap never operates alone, and reading it in isolation is the classic mistake. Three neighbouring mechanisms determine what a cap is worth: the exclusion of indirect and consequential damages, which often removes lost profit and business interruption from recovery entirely; indemnities, which frequently sit outside the cap; and insurance obligations, which determine whether a promise to pay is worth anything in practice.

A seemingly generous cap combined with a sweeping exclusion of indirect loss can leave you recovering less than a modest cap with narrow exclusions. The question is never "how high is the cap" but "what would we actually recover in our realistic worst case".

The structures you will actually encounter

Caps in commercial contracts cluster around a few recognisable shapes:

  • A cap tied to the fees: the amounts paid or payable under the contract, sometimes measured over a defined recent period rather than the whole term. This is the dominant pattern in services and subscription agreements, because it scales the exposure to the size of the deal.
  • A fixed amount: common where the fees are small relative to the damage the work could cause — a modest engagement touching critical systems or sensitive data — or where the parties simply want certainty.
  • Per-claim versus aggregate: whether the cap applies to each event separately or to all claims over the contract's life. The same headline figure means very different things under the two readings.
  • Separate, higher caps for specific risk areas — data incidents being the frequent modern example — sitting between the general cap and a full carve-out.

Which structure fits, and at what level, depends on the economics of the deal: what the fees are, what a failure would actually cost, and who can insure or absorb the risk more cheaply. That is a facts question, not a convention question.

The carve-outs are the more standardised part

If anything in this area deserves the label "market standard", it is the carve-outs — the categories where liability remains unlimited regardless of the cap. Gross negligence and unlawful intent lead the list, and under Swiss law this first carve-out is not a concession at all: an advance exclusion of liability for such conduct is void under the framework of Art. 100 CO, so the clause merely restates what the law imposes.

The genuinely negotiated carve-outs come after that: breach of confidentiality, infringement of third-party intellectual-property rights (usually via the IP indemnity), bodily injury, and increasingly breaches of data-protection obligations. Here the parties really are allocating risk, and here the discussion belongs: a supplier holding your data resisting a data carve-out is telling you something about where it expects problems.

When deviating from the pattern is justified

"Market standard" is a starting point, not an answer, and deviating is legitimate whenever the standard pattern misprices the actual risk. The clearest case is asymmetry between fee and exposure: where a low-fee service could cause damage far beyond anything fee-linked, a fee-based cap is structurally inadequate, and a fixed amount or a dedicated regime for the critical risk is the reasoned position. Insurance is the second honest argument — a cap aligned with the coverage a party actually carries is more defensible than any abstract formula. Dependency is the third: sole-source suppliers, migration costs and integration depth all change what a breach really costs.

What does not justify deviation is habit — "our template says" — in either direction. If you demand an unusual position, expect to explain the risk that motivates it; if you concede one, price it.

Reading a cap before you sign

The practical test fits in four questions. Set the aggregate number against your realistic worst-case loss. Check what the exclusions remove from recovery before the cap even applies. Establish which indemnities sit outside the cap, in both directions. And confirm the insurance behind the promise. Whether the result is acceptable depends on your facts — the label "market standard" settles nothing by itself. If a cap in a contract on your desk feels either aggressive or oddly generous, that instinct is worth a conversation, and we are happy to have it.