The Deal Room, No.1: How to Read a Contract You Have Never Seen Before

This series is written for non-lawyers. It is general education, not legal advice, and it does not replace a lawyer. If you have a question about a specific transaction, get a legal opinion.

The best guidance I have received about reviewing a contract, and the best I can give: read it three times.

Pass one is structure. What kind of deal is this, how is it built, and where did the drafter choose to put things? Pass two is substance, line by line, where you work out what each provision actually says and does. Pass three is critical review, where you stop reading and start revising, testing the agreement against your best case and your worst case and marking up everything that fails either one. This piece is about pass one.

I run Word Witch Consulting, where most of my work is entertainment and interactive media transactions, on the developer side and the publisher side both. What follows is roughly what I do in the first ten minutes with an agreement I have never seen.

Shape tells you what the other side cares about

A contract's shape will tell you more about a deal, and about the parties' priorities, than any single substantive term.

People put the things they care about up front, where they cannot be glossed over or waved away. Lawyers and companies are no different. They want to be sure you see what they want you to see, and they are content to let you skim past the rest.

So let's say you open a publishing agreement. Chances are the first page is commercial terms, by the numbers: scope, timing, deliverables, milestones, money, and what each side is signing up to do. It reads clean, because that page was written to be read. Then come the standard terms and conditions, and that is where most of the risk lives. It is also the part the other side is least willing to move on. Funny how that works.

Now open a work-for-hire or independent contractor agreement instead. The front of that document usually describes how the parties intend to work together in the abstract: roles, process, IP ownership, confidentiality. The actual deal, meaning what you are building and what you are being paid for it, is frequently sitting in a Statement of Work stapled to the back. A creator who reads front to back and loses steam around page nine will never reach the section that governs their week.

Two layouts, opposite construction, same lesson. Find the money, find the risk, and notice how far apart the drafter decided to put them.

Read the headings before you read a single word of substance

Once you have the layout, go through the headings and subheadings in order and read nothing else. You are not trying to understand the deal yet. You are building a map, so that when a provision on page four sends you to Section 11.3, you already know roughly where you are going and why.

Then ask the question that does the real work here: why is this section in this place? Which comes first, compensation or your obligations? Is there a section you expected to find and cannot? Ask it at every heading. The answers accumulate into a fairly honest picture of the deal before you have read a term of it.

The introduction is simple. The recitals are not.

A contract's introduction usually does four things: it identifies the parties, names the agreement, states the effective date, and acknowledges that both sides agree to be bound. That should be the whole of it, and it is the least interesting page in the document. Two exceptions are worth a second look: a parent signing on behalf of a subsidiary, and a third party meant to have rights under a deal it is not a party to.

The recitals are the paragraphs that follow, often labeled Background or beginning with “Whereas.” They describe what the parties do, what they want to do, and how they intend to do it. No promises are made here, and nothing in this section obligates anybody to anything.

So you can skip it, right?

Not quite. When an operative term turns out to be ambiguous, and terms turn out to be ambiguous more often than anyone would like, recitals are one of the first places a court or an arbitrator will look to work out what the parties meant. A background paragraph that casually describes your studio as a “service provider,” or describes the project as “work commissioned by Publisher,” can quietly color how the rest of the agreement gets read. Recitals cost nothing to write, which is exactly why people write them carelessly. Read them like they matter.

Defined terms are where the deal is actually made

This is the part of a first pass I care most about. Defined terms are where the parties shape the meaning of everything else in the agreement, and ambiguity here poisons the entire document.

There is a concept in Tibetan Buddhist debate called the eight doors of logical pervasion, which sets out what has to be true before a term and its definition can be said to actually match. I have found it to be a much better test for contract definitions than anything I was taught in law school. Call the term A and the definition B:

1.  If A, then B

2.  If B, then A

3.  If not A, then not B

4.  If not B, then not A

5.  If there is A, there is B

6.  If there is B, there is A

7.  If there is no A, there is no B

8.  If there is no B, there is no A

Eight doors, and every one of them has to be open. Satisfy seven and you do not have a definition. You have a description that holds up most of the time, which in a contract is another way of saying it holds up until somebody has a reason to argue with it.

What the doors are doing is testing one relationship from every available side: forward and backward, in the affirmative and in the negative, as a matter of meaning and as a matter of what actually exists in the world. That looks redundant, and the redundancy is the entire point. It is genuinely difficult to write something that walks through all eight doors and still fails to mean what you intended.

So run a real one through. A publishing agreement defines Net Revenue as gross receipts less returns, chargebacks, platform fees, and marketing costs.

Start at door two, which asks whether everything the definition captures is actually the thing. Marketing costs, with no cap and no approval right, means the other side’s spending decisions are now inside your revenue term. They can spend your royalty and hand you the receipt. Things are falling within B that you never meant to be A, and that door is shut.

Now try door six, which asks whether the thing existing in the world means the definition reaches it. What about revenue from a bundle, a subscription service, a regional storefront, or a sale of your game as part of a catalog? If the clause says gross receipts without saying whose receipts or from where, revenue can exist that your definition never touches. Shut.

Two doors down and you can stop, because one failure is fatal. Walk the other six anyway, since each tends to surface a different problem, and the negative doors are good at catching terms that quietly overlap with each other.

That is one definition, in one clause, in a deal that might carry sixty of them. This is why the definitions section is not the boring part. I have built this same test into the AI tools I use to draft and review, grading every definition on which doors it actually clears, and it catches things every single time.

Scope tells you what you are not getting paid for

After the definitions, the provision that shapes a deal most is its scope. Scope tells you what the contract is supposed to do and what the parties are actually agreeing to.

Scope is deal dependent, so the specifics vary. What does not vary is the useful way to read it, which is backwards. A scope provision draws a boundary, and the interesting question is what falls outside it. Which platforms are covered, and what happens when the publisher wants a Switch port? Does scope include updates, patches, live operations, or a sequel? Can you reuse your own engine on the next project?

Anything outside that boundary is either work you will be asked to do for free, or a second negotiation you will be conducting from a much weaker position than this one. Read scope for its silences.

What you should have by the end of pass one

You have not read the contract yet. That is fine, and it is the point.

What you should be able to answer after a first pass is this: what kind of deal is this, where did the other side put the money and where did they put the risk, and what are the three or four defined terms that everything else is going to hang on? If you can answer those, you now have somewhere to stand when you go back in for pass two and start reading line by line.

That is the next piece in this series. Bring your notes.