By Mako Murazvu
If you’re a junior lawyer, your first encounter with a Master Services Agreement (MSA) probably went something like this. A partner dropped a message asking for a “quick red-flag review,” and you opened a 20-page document packed with defined terms you’d never seen before. This guide walks through how to approach that task, and more importantly, how early drafts typically get reshaped by the senior lawyers who review them.
The goal here isn’t just to tell you what an MSA usually says. It’s to give you a sense of how senior lawyers actually think about these documents. Spotting issues? You’ll pick that up pretty quickly. Knowing which ones are worth raising, and how to frame them so a client can actually do something with your advice? That’s the part that takes longer.
Of course, I have to make it clear that nothing in this guide constitutes legal advice, and it should not be treated as such. No solicitor-client, attorney-client or other professional relationship is created by your use of or reliance on this material.
The content is provided for general educational and informational purposes only, may not reflect current law or regulatory requirements, and may not be applicable in all jurisdictions or to all fact patterns. You should not act or refrain from acting on the basis of anything contained in this guide without first obtaining independent professional advice tailored to your specific circumstances.
Step one: get the lay of the land
Before you start marking anything up, take a step back and figure out how the document actually fits together. There are two structural things that trip people up in almost every MSA.
First, an MSA is almost never the whole picture. It usually sits on top of one or more Statements of Work (SOWs), and sometimes a licence agreement too. Find the order of precedence clause early on, because that’s what tells you which document wins if they contradict each other. The natural instinct on a first review is to just mark up the MSA on its own. What becomes clear pretty quickly is that a more experienced lawyer flags right at the start that any SOW or licence terms also need a look, and asks whether drafts of those are floating around yet.
Second, read the definitions before you dive into the operative clauses. I cannot stress this enough. If you get a key definition wrong early on, like what counts as a “Deliverable” or who actually falls within the definition of “Customer,” you’ll end up misreading every substantive clause that relies on it.
Step two: the clauses that really matter
Now work through the clauses that carry the most risk, particularly liability, IP, termination, payment, confidentiality and indemnities. These are the areas that come up again and again across pretty much every commercial services agreement I’ve seen.
Liability caps and exclusions
Find the limitation of liability clause and read it carefully. Honestly, read it twice. It’s usually the single most important clause in the whole agreement. There are three questions you want to answer. What is the cap and what does it apply to? What’s carved out of the cap? And what kinds of loss are excluded altogether, regardless of the cap?
One pattern you’ll see a lot is a cap tied to fees paid for the specific deliverable that caused the loss, rather than to the contract as a whole. That sounds reasonable until you realise a low-value component could cause a high-value loss, and the cap barely covers it. One-way exclusions of consequential loss and data loss in the supplier’s favour are also pretty standard.
A common junior mistake is flagging every single element of this clause with the same level of urgency. A senior lawyer’s version will look very different. Instead of pushing for a complete renegotiation of the cap, a more experienced reviewer might ask the client to consider whether the cap is commercially acceptable in context, and concentrate the firm pushback on a short list of loss categories that shouldn’t be capped, things like wilful misconduct, fraud, breach of law, breach of confidence, and third-party IP or property claims.
One more thing to watch for here is asymmetry. A cap that only protects the supplier, with no reciprocal limit on the customer’s exposure, is surprisingly easy to miss when you’re focused on uncapping the other side.
Intellectual property
IP clauses need to answer three questions. Who owns background IP and any customisations? What rights does the customer actually get? And what happens if a third party turns up claiming infringement?
A structure you’ll see all the time keeps all IP with the supplier and gives the customer only a non-exclusive, non-transferable licence, with nothing confirming the customer’s ownership of its own data. That’s a real gap in any hosted or SaaS arrangement. Infringement remedies limited to whatever fix the supplier chooses, sitting inside the standard liability cap, are also common and worth flagging.
The typical junior instinct is to push for broad rights over customisations, like a perpetual licence, as a default ask. The feedback that usually comes back is to think about what the client actually needs. If they’re planning to give contractors access, then sure, push for a sub-licence right. But pushing hard for a broader right the client will never use just burns negotiating capital for no real benefit.
On infringement remedies, though, being more measured doesn’t always mean going softer. Some categories of loss, like third-party IP claims, actually warrant a stronger ask (for example, an uncapped indemnity) precisely because they shouldn’t be limited by the standard cap.
Termination
Look at the termination triggers, both with and without cause, the notice periods, any minimum term, and what actually happens when the contract ends for each party, including cost recovery and data return.
A risk pattern that comes up more than you’d expect is a contract that allows termination without cause on short notice and with no minimum term. That means a supplier could walk away shortly after a customer has committed serious budget and resources to implementation.
This is the kind of point that often gets raised in a junior’s early draft and then dropped entirely from the client-facing advice. The first reaction might be to assume the analysis was wrong, but that’s not necessarily what’s going on. Sometimes a term is just standard for this type of contract, or raising it would spend credibility that’s better used on higher-value points. When you see one of your comments get dropped in review, ask why rather than assuming you made a mistake. There’s usually a good reason.
Cost pass-through clauses on termination or delay are another area to watch. The typical junior take is to treat these as open-ended and push for a cap with a reasonableness and mitigation requirement. The more experienced approach is to first check whether the clause actually does what its heading suggests, and then narrow the ask. For example, limiting reimbursement to termination for the customer’s default, or tying extensions of time to genuine critical-path impact.
Confidentiality and data handling
Confidentiality clauses are usually less contentious than liability or IP, and honestly, standard drafting often doesn’t need a comment. Learning to recognise which clauses are conventional, so your comments focus on real risk, is itself a skill worth developing.
One related gap that is worth checking for in any hosted arrangement is the absence of dedicated data handling provisions, things like data location, subprocessor use, deletion on termination, security standards and breach notification. A brief general privacy clause doesn’t necessarily cover this ground.
Indemnities
For any indemnity clause, figure out whether it’s mutual or one-directional, what it actually covers, and how it interacts with the liability cap.
A general mutual indemnity for breach can look fair on its face because it applies to both parties. But mutuality doesn’t automatically mean balanced. A general indemnity for contractual breach isn’t necessarily standard market practice; indemnities are more typically confined to specific categories of loss, like IP infringement, property damage or personal injury.
This matters because an indemnity claim usually escapes the remoteness, causation and mitigation limits that apply to an ordinary damages claim. So if one party’s obligations are more extensive or more easily breached, a mutual indemnity may actually expose that party more than it protects it.
This is a really good example of the difference between spotting an issue and exercising judgment about it. Spotting it is noticing the indemnity and checking the cap. Judgment is recognising that “mutual” doesn’t automatically mean “fair.”
Learning to calibrate, not just flag
Here’s something worth hearing early on. The biggest gap between a junior review and a senior one is rarely a missed issue. Junior lawyers are usually pretty good at finding problems. The gap is judgment. It’s about which problems matter enough to raise, how hard to push, and how to frame a point so the client can actually act on it.
Learn to distinguish between issues that are genuinely unusual and issues that are simply favourable to the drafting party. Every MSA favours whoever wrote it on things like liability, IP and payment terms. That alone isn’t a red flag. The question is whether a term goes beyond what’s typical for this kind of contract between parties of this size. Flagging every drafter-favourable term with the same level of urgency just dilutes your list.
Think about what the client can actually do with each point you raise. A comment that just describes a risk is less useful than one that frames a decision, with enough commercial context for the client to make the call themselves. A large corporate client is usually better placed than its lawyers to judge how much a given risk matters to its business, once you’ve explained the risk clearly.
There’s also a difference between a point that’s legally correct and one that’s worth spending negotiating capital on. A negotiation only has so many points either side will meaningfully push, which is why an experienced reviewer sometimes drops a perfectly valid point from client-facing advice. That doesn’t mean you should self-censor your own working drafts. It means treating the step from markup to advice as a deliberate one.
And please, resist the urge to treat every issue as critical. A markup where everything is flagged as serious reads as one that hasn’t been triaged. Get into the habit of sorting your own findings into significant risks, standard market terms, and modest points before you send them up the chain.
One last thing
Your first few reviews will be over-inclusive, and honestly, that’s exactly where you should be. It’s much better to flag too much and have a senior colleague help you prioritise than to miss something significant.
After every review cycle, take the time to study not just what changed in your draft, but why. When a point gets softened, removed, or made more assertive, try to work out the commercial reasoning behind it. That habit, more than anything else, is what turns contract review into commercial judgment.
