Limitation of Liability

Without this sentence, nobody knows how large one incident can get. A collaboration tool sold for $230 a month goes down for a day, the customer can't take orders during that day and loses $375,000, and there is nothing in the contract connecting the money coming in to the money going out. A deal worth $2,800 over a full year can produce a $375,000 claim.

So the selling side started hesitating in front of exactly the deals it wanted. When Sofia brought Pagedeck's first enterprise contract in, the question Grace asked after reading it through was not about price. "If this goes badly once, is that the end of the company?" A single contract worth $27,000 a year is not revenue if the whole business is standing behind it as collateral.

The buying side isn't comforted by an open limit either. If the vendor's entire assets come to $750,000, a $3,750,000 loss still collects $750,000 at most. Unlimited liability is only unlimited on paper. Both sides ended up preferring a number written down in advance, with price and terms set on top of that number.

The ceiling isn't one figure. Three devices stack to produce it.

1. The cap amount. This sets the maximum payout. The most common form is everything the customer actually paid in the twelve months before the incident. For a customer on $750 a month, the ceiling is $9,000. It's keyed to money already received rather than total contract value, because that keeps the size of the deal and the size of the risk in the same order of magnitude.
2. The types of loss. This decides which losses count at all. What the customer spent recovering from an outage is direct loss; the revenue they couldn't book during those hours and the damage to their reputation are indirect. Indirect loss is normally excluded outright. There is no reliable way to size it, so leaving it in keeps the dispute large no matter how low the cap amount goes.
3. The carve-outs. This lists what sits outside the ceiling. Willful misconduct and gross negligence, breaches of confidentiality, personal data leaks, and the third-party claims handled by the indemnification clause are the regulars. For anything on this list, the cap does not apply.

You need all three to know the real exposure. A $9,000 cap does nothing during a data leak if data leaks are carved out.

The clause usually runs both ways. If there's a ceiling on what you pay out, the same ceiling applies to what you can collect when a customer stops paying. Pushing the number down is not automatically a win.

Two values move in negotiation: the multiplier and the period. Buyers ask for twenty-four months instead of twelve, and for two or three times fees instead of one. The heavier the promises in the warranty clause, the higher that multiplier goes — committing to 99.9 percent uptime while capping damages at a token amount prices the promise and the commitment differently.

It appears in nearly every master service agreement (MSA) signed with an enterprise customer. An MSA sets the terms that govern the whole relationship so they don't get renegotiated per order, and this clause is among the first three the other side's legal reviewer marks up. A good share of the weeks that enterprise sales cycles spend in redlines is spent here.

In self-serve products where one person signs up and pays by card, the same content lives in the terms of service. There's no counterparty to negotiate with, so the wording stays as written.

The first place the numbers get tested is the service level agreement (SLA). An SLA is a contractual promise about service levels such as uptime, and a broken promise is often settled with credits against next month's invoice rather than cash. Those credits act as a first ceiling; losses that credits don't cover run into the second ceiling this clause defines.

"Our liability is capped at twelve months of fees and they're asking for two times."

On a $45,000 contract, the worst case moves to $90,000, and someone has to decide whether that's acceptable.

"They want personal data incidents carved out of the cap."

A request to drop the ceiling for leaks. The maximum loss from one incident becomes open-ended, which makes it a founder-level decision.

"Worst case on this contract, what do we actually owe?"

Asking whoever read the clause to answer with a single number.

At Pagedeck the clause turns up the same way each time. Sofia brings back the redlined draft, marks what the other side changed, and passes it to Grace. Grace reads three things: what the ceiling is, what has been moved outside it, and whether the same terms bind the customer as well.

If those three sit inside what the company can absorb, it goes through. If they don't, the price gets recalculated, because a contract that carries more risk has to be worth more.

The most common mistake is reading the capped figure as the maximum possible loss. The number is right there, so that reading comes naturally. But carve-outs always travel with the clause, and the incidents that actually shake a company are mostly the kind that sit on that list — data leaks above all. The cap also covers damages only. Staff time spent responding to an investigation and legal fees land outside it. Treating a leak lightly because the cap says $9,000 is itself the sort of decision that pulls an incident into the carve-out territory.

The second mistake is assuming a lower cap is always better. From the seller's chair it looks like less exposure, which is fair enough. But the clause binds both parties in most contracts, and it's easy to forget you're also somebody else's customer. The bigger problem shows up at the table. A ceiling set far too low reads as a signal that you don't intend to keep your promises. Offering a $3,800 liability cap on a $150,000 annual contract usually ends the review right there.