Direct vs consequential damages, in plain English

Every meaningful commercial contract contains some version of this sentence: "In no event shall either party be liable for any indirect, incidental, consequential, special, or punitive damages, including without limitation lost profits, lost revenue, or lost data." It's the most copy-pasted paragraph in contract law, and it sits two inches above the limitation of liability cap.

Most people reading that sentence have a rough intuition that "consequential" means "downstream," but the actual line between direct and consequential damages is blurrier than the clause suggests. The same dollar amount, say, $2M in lost revenue from a botched integration, can be direct damages in one framing and consequential in another. Which side of the line it lands on determines whether anyone writes a check.

This is a walkthrough of what direct and consequential damages actually are, why the categories matter so much in a contract, and where the boundary tends to move.

The two buckets

When a contract is breached, the non-breaching party is usually entitled to damages, a dollar recovery meant to put them back in the position they would have been in if the contract had been performed. Courts and contracts split those damages into two main buckets:

Direct damages

Direct damages are the losses that flow immediately and naturally from the breach itself. They're the damages that anyone standing on the street could predict without knowing anything special about the non-breaching party's business.

If a vendor fails to deliver software they sold for $100,000, the direct damages are roughly the $100,000 the customer paid (or the cost of covering with a replacement). The harm is the non-delivery itself, measured against what was promised.

Consequential damages

Consequential damages (sometimes called indirect or special damages) are losses that flow from the breach but go beyond the immediate transaction, they depend on facts specific to the non-breaching party's situation.

Same vendor, same non-delivery. But the customer was launching a new product on a timeline that depended on the software. The launch slips, the PR firm's retainer is wasted, and an expected $2M in first-quarter revenue evaporates. That $2M in lost revenue is consequential damages. It flowed from the breach, but only because of the customer's specific plans, which the vendor didn't necessarily know about.

Why the clause excludes one and not the other

The standard exclusion knocks out consequential damages but leaves direct damages intact. Read the clause again: "indirect, incidental, consequential, special, or punitive." Direct damages are conspicuously absent from the list.

This is deliberate, and it reflects a rough bargain:

• Vendors will stand behind what they sold, if they fail to deliver, they'll refund or make the customer whole on the immediate transaction.
• Vendors will not stand behind the customer's business plan, if the customer built a $100M product launch on a $10K SaaS contract, the vendor isn't on the hook for the $100M.

The logic is economic. If a $10K vendor could be liable for $100M in consequential damages every time something went wrong, nobody would sell the software at $10K. The price implicitly assumes a ceiling, and the consequential damages exclusion is how that ceiling gets written down.

Where the line actually sits

The tidy two-bucket framing breaks down fast. The same loss can be direct in one framing and consequential in another, and some of the biggest fights in contract disputes are about which bucket a particular loss goes in.

Lost profits, the classic ambiguity

The phrase "lost profits" is usually listed as consequential damages in the exclusion clause. But courts have held lost profits to be direct damages in some contexts, specifically, when the profits lost are the profits the contract itself was designed to generate.

Example: A software company sells a dealer a license to resell its product. The software company breaches by refusing to deliver. The dealer's lost reselling profits are direct damages, because selling the product at a margin was the whole economic point of the contract, not a downstream side effect.

Same software company sells the same product to a manufacturer who uses it internally. The manufacturer's lost profits from its own product line, which it couldn't ship because of the software failure, are consequential damages, because the manufacturer's business is a separate economic layer the vendor didn't directly participate in.

Cover costs

If a vendor fails to deliver and the customer has to buy a replacement at higher cost, the difference between the contract price and the replacement price is usually direct damages. That's close to textbook, cover costs are what the non-breaching party pays to put themselves back where they would have been.

Data loss

Almost always listed as consequential. But if the entire purpose of the contract was data storage, a backup provider, say, data loss is arguably direct. Which is why sophisticated backup contracts often carve data loss out of the consequential exclusion, or set a specific cap for it.

Reputational harm

Almost always consequential, and almost always excluded. Reputational harm is hard to quantify and the link between a contract breach and a reputation hit is usually indirect.

The Hadley test

The common law test for consequential damages comes from a 1854 English case called Hadley v. Baxendale, and it's still the framework American courts use. Simplified: consequential damages are recoverable only if, at the time of contracting, the breaching party either:

1. Knew that the specific loss was a natural consequence of breach, or
2. Was made aware of the special circumstances that would make the loss likely.

The practical upshot: if a customer tells the vendor during sales "we're launching on March 1 and a slip will cost us $2M", and the vendor signs anyway, that knowledge can pull the $2M into the category of foreseeable damages, which complicates the vendor's argument that the exclusion clause insulates them.

This is why sophisticated vendors are careful about what they accept in writing during negotiation. Statements that acknowledge the customer's downstream stakes can erode the consequential damages exclusion after the fact.

How the exclusion interacts with the cap

The consequential damages exclusion and the liability cap are two separate limits stacked on top of each other:

• First, the exclusion zeroes out entire categories of damages, lost profits, lost revenue, lost data, reputational harm.
• Then the cap puts a dollar ceiling on whatever damages survive the exclusion, typically direct damages.

The combined effect is that a $50M consequential loss can become a $100K direct-damage claim capped at $100K in fees, a 500x reduction. This is not an accident; it's the design.

Carve-outs that bring consequential damages back

Some contracts carve specific categories back into the scope of recoverable damages, meaning the consequential exclusion doesn't apply to them. Common carve-outs:

• Breach of confidentiality. A leak of trade secrets often produces mostly consequential losses, so exempting this category from the exclusion is common.
• Indemnification obligations. IP infringement, third-party claims, these usually survive the consequential exclusion because they're inherently downstream.
• Gross negligence and willful misconduct. Courts in many US jurisdictions won't enforce consequential exclusions in cases of bad faith anyway, so many contracts make this explicit.
• Data breach. Increasingly called out as a standalone carve-out, especially in contracts involving regulated data.

The carve-out structure is how customers preserve recovery in the scenarios that matter most, without rewriting the exclusion from scratch.

Reading the clause carefully

A few drafting patterns worth noting:

• "Including without limitation lost profits..." The word "including" matters. If it's "including" (non-exhaustive), lost profits is one example of consequential damages. If the clause tried to say "consequential damages means lost profits," the scope would be much narrower.
• Placement of "lost profits." Some clauses exclude "lost profits" as a standalone category alongside consequential damages, which arguably excludes direct lost profits too. Other clauses list lost profits only as an example of consequential damages. The difference matters.
• Mutual vs. one-way exclusion. Most consequential exclusions are mutual, both sides give them up. A one-way exclusion (customer can't claim consequentials but vendor can) is a red flag.
• Survival language. Some contracts expire the exclusion after termination. Worth checking.

Common negotiation moves

Teams actively negotiating this section tend to reach for:

• Carve-outs for confidentiality and data breach. The highest-leverage ask and often granted.
• Preserve direct lost profits. Narrow the exclusion so that lost profits from the contract itself (direct lost profits) aren't swept up with truly consequential losses.
• Cap the exclusion rather than eliminate it. Instead of excluding consequential damages entirely, cap them at a separate (higher) dollar number, preserving some recovery without exposing the vendor to unlimited downstream risk.
• Align the exclusion with the indemnification clause. Make sure indemnified third-party claims aren't accidentally knocked out by the consequential damages exclusion.

The bottom line

The distinction between direct and consequential damages looks philosophical until a contract breaks and the dollars show up. Direct damages are the immediate, predictable loss from the breach itself. Consequential damages are everything downstream, and almost every commercial contract excludes them outright.

What matters operationally is not memorizing the two definitions but understanding that the line between them is fuzzy, that a loss's bucket depends on the framing, and that the carve-outs to the consequential exclusion are often where the real risk allocation happens. The boilerplate you're reading was negotiated in the past for some other deal. Whether it fits the current one is a separate question.