Why does every deal end up amending our standard agreement?

Because the version sales sends out is the last deal's redlined copy, carried forward as the new baseline instead of the master file legal approved. Each new deal copies forward whatever concessions the previous one won, since that is faster than reopening the original template, and those concessions accumulate silently until the 'standard' document no longer matches what anyone thinks it says.

How do we track which customers got which contract exceptions?

With a short log per exception: the customer, the clause, the standard position, the granted deviation, the named approver, the reason, and when it gets reviewed again. The named-approver field is what keeps the log honest — a deviation nobody can attribute to a decision is one that gets rediscovered and re-granted by someone who assumes it was already policy.

When should a repeated exception become the new template?

Not automatically, and not at a fixed count treated as a formula. A workable rule to pre-commit to: when independent negotiators on unrelated deals concede the same clause a third time, treat that as a trigger to schedule a review, not as proof the template must change. The review then asks whether the opening position was wrong or the segment genuinely differs from the one the template was written for — those get different fixes.

Should the negotiator who granted an exception decide whether it becomes permanent?

No. Whoever granted the exception was under deal pressure and motivated to close, which makes them a poor judge of whether the concession should become standing policy. That decision belongs to whoever owns deal desk or contracts operations, reviewing the log on a fixed cadence separate from any single negotiation.

Why your standard MSA stopped being standard

Guide · Enterprise · 4 min read · last verified 2026-08-11

Reviewed before publication Editorial board Independent commercial review

In shortWhy MSA templates drift deal by deal, the logged-exception discipline that tracks departures back to a master file, when a repeated exception should trigger a template review, and where the order-form boundary fits.

Your standard MSA stopped being standard because the document your team sends out today is not the master service agreement legal originally approved — it is deal forty's redlined copy, carried forward as the new starting point because copying forward is faster than reopening the master file. Each concession made to close one deal quietly becomes the baseline offered to the next, and nobody decided that on purpose. None of the framing below is legal advice; it assumes your legal team has already set the policy positions that the exception log is built to hold onto.

How a standard template stops being standard

Template drift has one mechanical cause. Producing a new MSA for deal forty-one means opening deal forty's version and changing the customer name — that file already reflects a real negotiation, and starting fresh feels like re-litigating settled ground. Each subsequent deal repeats the shortcut. Enough cycles later, the file sales sends out as "our standard MSA" is not standard at all — it is an accumulation of one-off concessions, still labeled with the name of a template that no longer describes what it contains. The fix is not refusing to reuse language; reuse is efficient and mostly correct. The fix is reusing from one place: every deal should redline a fresh copy of the actual master file, and every accepted departure should be tracked back to that master rather than folded silently into the copy that becomes the next starting point. The same shortcut shows up one layer down, in a statement of work template customized for enough services deals that it stops being a template too, for the identical reason.

The logged-exception discipline

An exception log answers a question that becomes unanswerable once drift sets in: which customer holds which deviation from the master, who approved it, and why. Without one, whoever is negotiating deal forty-one has no way to know that deal thirty-eight already has an eighteen-month liability cap instead of the twelve months the master file sets, so they either re-fight a position that was already lost once or unknowingly hand out a better term than deal thirty-eight received. A workable log is short and specific, not a running commentary on every redline:

Customer: [name]
Clause: Limitation of liability
Standard position: 12 months of fees paid
Granted exception: 18 months of fees paid
Approved by: [named approver, title - not "legal"]
Reason: [multi-year commitment, strategic account, etc.]
Review at: [renewal date, or next template revision]

Skip the named-approver field and the log records the deviation but loses the decision behind it. A deviation nobody can attribute to a decision-maker is a deviation that will be granted again by someone who assumes it must already be standard, because it is sitting right there in a prior contract.

When a repeated exception should become the template

Not every exception deserves to graduate, and a log that promotes every third request into the master file just moves the drift problem up one level. The useful trigger is not a fixed count treated as a rule; it is a scheduled look, prompted by a pattern. A reasonable habit to pre-commit to: when the same clause draws the same exception from independent negotiators on unrelated deals, treat the third instance as a signal to review the template, not as proof the position must change. The review asks a different question than any single negotiation did — not "should we grant this to close this deal" but "is our opening position wrong, or is this customer segment genuinely different from the one the template was written for." Those have different fixes: a wrong opening position gets rewritten; a genuine segment difference gets a second template, not a weaker version of the first one.

The order-form boundary

Drift is not only an MSA problem. It also shows up when a negotiator patches a relationship-level term into an order form because that document was already open and the amendment route felt slower. Which terms belong on an order form, and which belong one document up, is a boundary already drawn; the drift question is what happens to the term afterwards. A concession that belongs in the relationship terms but lives only in one order form disappears the moment that order form is superseded, which either loses a customer's real commitment or resurrects an old fight the next time they buy something. So the log needs a second field beyond the clause: which of the two documents the exception belongs in.

Who owns the log

An exception log with no named owner decays exactly the way the template it is meant to protect already did. The person best placed to grant an exception during a live negotiation — under deadline pressure, motivated to close — is the worst-placed person to later judge whether that exception should become permanent policy. Ownership belongs with whoever runs deal desk or contracts operations, on a standing review cadence, separate from any single deal's timeline. None of this replaces the actual redline exchange — the cover notes, the fallback positions, and the front-loaded reviews described in what a contract redline is still govern how any one negotiation runs. This is the layer above that exchange: what happens to a concession after the deal that won it has already closed.