Legal Teams// definition

Third-party paper: why their template is a different job from your own

In short

Third-party paper means the draft under review is the counterparty's template, so none of its structure is known in advance. On your own paper, review is a diff. On theirs, the reviewer first locates the concept, normalises it to their own clause taxonomy, resolves the defined terms, and only then judges — and the hardest step is finding what is not there at all.

Key takeaways

  • On your own template review is a diff; on theirs it is reconstruction, and the two need different pipelines.
  • Four steps precede any judgement: locate the concept, normalise it, resolve the defined terms, detect absence.
  • Absence detection is the highest-value step, because a missing clause has no text for anything to retrieve.
  • A concept can be split across 2 sections, buried in a schedule, or expressed without ever using its usual name.

Third-party paper is any draft built on the counterparty's template rather than yours. The distinction matters because it changes what review is. On your own paper the reviewer is comparing text against known positions in known places. On theirs, nothing about the layout is known, and the comparison cannot start until the reviewer has rebuilt a map of what the document contains.

That is the whole asymmetry, and it explains why review times on incoming paper are not a multiple of review times on your own. They are a different distribution: mostly fine, occasionally very long, because the long ones are documents where a concept was somewhere nobody looked.

What you lose the moment the draft is not yours

  • Position. Your limitation of liability is section 11 in every document you send. Theirs may be in the middle of a warranty section, or in a schedule of commercial terms.
  • Vocabulary. The same obligation appears as indemnity, hold harmless, or a bare covenant to reimburse. Matching on the heading finds none of them.
  • The definition map. Your template's defined terms are stable and known; theirs are set by the document, and may be reset in a schedule — see defined terms and where their meanings move.
  • The completeness guarantee. Your template contains every clause you require because you wrote it that way. Theirs contains what suits them.

Four things the reviewer rebuilds before judging anything

  1. Locate. Find every span of text that carries a concept you care about, regardless of heading, section number or where in the file it sits. One concept can occupy 2 non-adjacent spans, and an obligation in a schedule is still an obligation.
  2. Normalise. Map each located span onto your own clause taxonomy, so downstream logic can ask about limitation of liability rather than about section 11.4. Without this step every later rule has to be written against their headings, which is a rule set that lasts one counterparty.
  3. Resolve. Substitute the definitions this document actually assigns to its capitalised terms before scoring anything, because the same sentence can be harmless or dangerous depending on a definition 40 pages away.
  4. Detect absence. Compare the taxonomy list against what was located, and report what is missing. This is the step nothing does for free.

The order is not negotiable. Scoring before normalisation produces findings expressed in their vocabulary, which no playbook can consume. Scoring before resolution produces confident judgements about words whose meaning has not been established.

The finding with no text to retrieve

Every other finding has something to point at. Absence does not. If the counterparty's template simply has no data-protection clause, there is no passage to flag, no low-confidence extraction, no odd wording — the document is silent, and silence looks identical to a clean bill of health.

So absence has to be checked against the list, not against the document. The system iterates your clause taxonomy and asks, for each entry, whether anything in this contract was mapped to it — and a taxonomy entry that no reviewer agreed to is an entry that will never be missed. That is why the list is settled before build, not after.

One pipeline with a switch, not two products

StepYour templateTheir paper
Find the clauseKnown position, stable numberingSearch by concept across body, schedules and annexes
Name itAlready your taxonomyNormalise their heading and wording onto your taxonomy
Read the termsDefinitions known and stableResolve per document, including redefinitions in schedules
Judge itDiff against approved languageCompare a reconstructed clause against a position, with a stated basis
Report gapsRare: the template is completeThe main event, and it comes from the taxonomy list
The same review step, on your template and on theirs

Two consequences follow. The first is calibration: a playbook tuned on your own drafts will treat almost everything on incoming paper as a deviation, which is how every clause comes back flagged and reviewers stop reading flags. The second is the standard you judge against — with no template to diff, the reference set becomes what you have signed before, and each precedent needs to arrive with the provenance a clause carries.

On your own paper the hard question is whether a clause is acceptable. On theirs, the hard question is whether you have found it.

One more trap: incoming paper rarely arrives alone. An order form on their template usually sits under a master agreement, and the term you are reading may not be the one that governs — see which document wins when the MSA and the SOW disagree. The locate-and-normalise layer is the part worth building once and reusing, which is the shape of work we describe under internal tools and ops. Sibling definitions sit in contract review and redlining, inside our work with legal teams.

Frequently asked questions

Short answers to the follow-ups this page tends to raise.

What is third-party paper in contract review?

It is a contract drafted on the counterparty's template rather than your own. The practical difference is that clause positions, headings and defined terms are all unknown, so the reviewer has to locate and name each concept before any assessment of it can begin.

Why is reviewing on the other side's paper slower?

Because the work starts earlier. On your own template the clause is where it always is; on theirs, finding the concept, mapping it to your own clause names and resolving the definitions all happen before the first judgement. The variance is the real cost — most documents are quick, and the slow ones are slow because something was not where anyone looked.

How does a review system detect a clause that is missing?

By iterating the clause taxonomy rather than the document. For each concept the firm requires, the system asks whether any span in the contract was mapped to it, and reports the ones with no match. Nothing in the text itself can raise the alarm, because a missing clause leaves no text behind.

  • contract review
  • third-party paper
  • clause taxonomy
  • legal ops
// shipped work

The work behind this page

Builds from our portfolio that this page draws on.

Working on something in this space?

Tell us where you are in a sentence or two. We'll tell you honestly whether we're the right team, and what a sensible first slice of the work looks like.

Start the conversation