"Risky" is the word everyone uses about a bad contract, and it almost never gets defined. People say a clause is risky the way they might say a hike is steep — directionally, but without a precise measurement. That's fine for casual conversation, and a problem when you actually have to decide whether to sign.
This article is a framework for thinking about what risky means in a contract context. It doesn't replace a lawyer's review. It's the mental model that lets you read a contract the second time and notice what you missed on the first. It is informational only and is not legal advice.
The three dimensions of contract risk
Contract risk has three components, and you need all three to make a useful judgment about a clause:
- Asymmetry. Does this clause apply to both parties equally, or only to one?
- Exposure. What is the maximum cost if this clause is triggered against you — and is it capped?
- Ambiguity. Is the language specific enough that you and the other party will read it the same way three years from now?
A clause that scores badly on all three is a red flag. A clause that scores badly on one but well on the others is usually negotiable. A clause that scores well on all three is the kind you sign without thinking — and most contracts have plenty of those, too.
Asymmetry: who carries the obligation?
Asymmetry is the easiest dimension to spot. Read the clause and ask whether it applies to "each party" or only to one. If it's one-sided, the next question is whether that's appropriate for the relationship.
Some asymmetry is normal and fine. A freelance contract where "Contractor shall deliver Services on the schedule below" is appropriately one-sided — the client is paying for delivery, not delivering anything themselves. The asymmetry follows the value exchange.
Problematic asymmetry is when the clause has nothing to do with the value exchange. A limitation-of-liability clause that caps the client's exposure but not the freelancer's. A termination clause that lets the client walk away with 7 days' notice but binds the freelancer until the project ends. An indemnification obligation that only flows one way. These create lopsided risk that doesn't match the underlying deal.
The fix for unjustified asymmetry is almost always to make the clause mutual. "Each party shall indemnify the other..." is a tiny edit that completely changes the balance.
Exposure: how much can this cost?
Exposure is the dollar value at risk. Many clauses look harmless until you ask what the worst-case outcome is.
Take a standard indemnification clause: "Contractor shall indemnify Client against any and all claims arising out of Contractor's work." On its face, it's procedural. In practice, if a third party sues the client over something the contractor did, the contractor could be on the hook for the client's legal fees, the third party's damages, and any settlement — none of which is capped by anything in the contract.
Exposure questions to ask about every significant clause:
- Is the obligation capped? A cap of "fees paid in the prior 12 months" is reasonable. Uncapped exposure is rarely appropriate for ordinary commercial work.
- Are there carve-outs? Caps that don't apply to gross negligence, willful misconduct, or IP infringement are standard — and reasonable.
- Is the worst case bounded? "Indefinite confidentiality" technically lasts forever. "Liquidated damages of $50,000 per breach" can stack quickly. Look for ceilings.
Unlimited liability is one of the most common contract red flags, and it's specifically the exposure dimension at fault. A clause can be perfectly mutual and clearly worded and still be ruinous if the upside is bounded and the downside isn't.
Ambiguity: will we read this the same way later?
The third dimension is whether the clause will mean the same thing when there's a dispute as it does today. Ambiguity isn't always intentional — drafters write quickly, copy from templates, and don't always think about edge cases — but ambiguous language is disproportionately bad for the weaker party. When a clause can be read two ways, the party with more lawyers usually wins the reading.
Words and phrases that often hide ambiguity:
- "reasonable" — reasonable to whom? Bring a standard or a benchmark.
- "from time to time" — how often? On what notice?
- "as may be requested" — by whom, in what form, with what deadline?
- "materially" — material to whose business? Quantify it.
- "including but not limited to" — the examples don't bound the list. If the bounds matter, make them explicit.
The fix is to replace fuzzy phrases with specifics: amounts, dates, named parties, and concrete examples. "With reasonable notice" becomes "with at least 30 days' written notice". The contract gets longer; the risk gets smaller.
These three dimensions together describe almost every clause worth worrying about. If you'd rather not run this analysis by hand, paste your contract into our free AI reviewer and you'll get a structured breakdown that highlights asymmetry, uncapped exposure, and ambiguous language — clause by clause. The review is informational, not legal advice, but it gives you the same first-pass triage a careful reader would do.
How to weigh risk against deal value
Spotting risk is half the work. The other half is deciding whether to accept it. A risky clause in a $5,000 project is not the same as a risky clause in a $500,000 contract — the appropriate amount of scrutiny scales with the stakes.
A simple way to weigh it:
- Stakes. What's the total deal value, and what would you lose if this contract turned hostile?
- Probability. How likely is the worst-case scenario to actually happen? An indemnification clause matters more if you're doing risky work; an IP assignment matters more if you have a portfolio you care about.
- Reversibility. Can you exit the contract if things go badly, or are you locked in? Termination-for-convenience clauses are the most underrated risk-reducers.
For low-stakes, short-duration, easily reversible contracts, you can accept a lot of one-sided language without losing sleep. For high-stakes, multi-year, locked-in contracts, every clause deserves scrutiny — and an hour with a qualified lawyer is well worth the cost.
When to walk away versus when to negotiate
Most risky clauses can be negotiated. Counterparties expect a marked-up redline back from a careful reader; refusal to negotiate is itself a signal about the relationship.
That said, some contracts are genuinely take-it-or-leave-it: SaaS terms of service, platform agreements, employment offers from large companies. For those, you have two real choices — accept the standard terms with eyes open, or walk away. The intermediate move is to limit your exposure by other means: keep the engagement small, document everything, and don't rely on a single contract for outsized economic outcomes.
A short checklist before you sign anything
The three dimensions condense into about ten questions. Run through them before you sign:
- Which clauses apply to me but not the other party?
- Are my obligations financially capped, and at what amount?
- What is the worst-case dollar exposure under each clause?
- Is any language vague where the dollars are large?
- How can I exit the contract if I need to?
- What happens to my IP, data, and customers if I leave?
- Who decides disputes, and in what venue?
- How long does this contract last, and how does it renew?
- Are there standard carve-outs (already-known, publicly-known)?
- If I sued under this contract, would I want to be on my side?
Run your contract through the reviewer
The fastest way to get a structured view of asymmetry, exposure, and ambiguity in your specific contract is to paste it into our free AI reviewer. You'll get a risk score, a list of one-sided and uncapped clauses, missing protections to add, and questions to bring back to the other party. No signup. We don't store the contract text.
For deeper dives on specific contract types, see our targeted reviewers and guides: how to review an NDA, 10 contract red flags, and the landing pages for freelance, employment, lease, and service contracts.