Master Service
Agreements
Everything you need to know before you sign, draft, or negotiate one
An MSA is not a single contract. It is the framework inside which all future contracts live.
ost companies sign their first Master Service Agreement without fully understanding what they are agreeing to. The document is long, the language is dense, and the pressure to start the project is real. So they sign, file it away, and hope for the best.
That hope is often misplaced. The clauses buried inside an MSA determine who owns the intellectual property when a project ends, how much either party can recover when something goes wrong, whether the relationship can be terminated overnight or requires six months of notice, and whose courts will decide a dispute. These are not small matters. They are the architecture of every commercial relationship built on top of that agreement.
This guide takes you through every dimension of the MSA — what it is, why it exists, how to draft it well, how to negotiate it strategically, and how to spot the clauses that can destroy a business relationship even when both parties intended to behave in good faith.
What is a Master Service Agreement — and what it is not
A Master Service Agreement is a contract that establishes the standard terms and conditions that will govern all future work between two parties. It is not a project contract. It does not, by itself, authorise any work to be performed or any money to change hands. It is the rulebook — the framework inside which all specific engagements operate.
The actual work is defined in a separate document: a Statement of Work (SOW), a Purchase Order (PO), or a Work Order (WO). These documents describe the specific deliverables, timelines, pricing, and acceptance criteria for a particular project. The MSA sits above them, providing the legal infrastructure that all SOWs inherit automatically.
The MSA is not the contract for the work. It is the contract about how all contracts for work will be conducted.
Fundamental distinction that most first-time signatories missThe relationship between an MSA and a Statement of Work
Think of the MSA as a constitutional document and the SOW as legislation passed under it. The SOW can specify anything not already covered by the MSA — project scope, milestones, pricing, team composition — but it cannot contradict the MSA unless there is an explicit override clause. When there is a conflict between the two, most MSAs specify that the MSA prevails. Some specify the opposite — that the most recent SOW governs. This hierarchy must be explicit.
This structure has a powerful practical benefit: once an MSA is in place, launching a new project requires only a short SOW — sometimes a single page — rather than a full contract negotiation. For companies that work together frequently, this dramatically reduces legal friction and accelerates project starts.
Who uses MSAs?
MSAs are standard in technology services, management consulting, marketing agencies, staffing and outsourcing, financial services, construction subcontracting, and any industry where a service provider and a client expect to engage in multiple projects over time. They are less common — though not unheard of — in one-off transactions.
Both parties benefit from having one in place early. Service providers benefit from standard payment terms and liability limits across all engagements. Clients benefit from consistent IP ownership, confidentiality, and performance standards. Neither party should be reluctant to propose an MSA before the first project begins — it signals professionalism and seriousness of intent.
The ten clauses that define every MSA
An MSA can run anywhere from four pages to forty. Regardless of length, the same ten core provisions appear in every meaningful agreement. Understanding each one — and the interests it protects — is the foundation of competent negotiation.
Scope of services
Defines what category of services the MSA covers. Broad definitions give flexibility; narrow ones prevent disputes about whether new types of work are covered.
Payment terms
Net-30 is standard. Net-60 or Net-90 strongly favours the client. Late payment interest, invoice dispute procedures, and set-off rights all belong here.
Intellectual property
The single highest-stakes clause. Determines whether the service provider or client owns what is created. Default rules vary by jurisdiction — never leave this to implication.
Confidentiality
Mutual NDAs embedded in the MSA. Duration, definition of confidential information, permitted disclosures, and obligations on termination are all critical sub-elements.
Limitation of liability
Caps the maximum financial exposure of each party. Often set at 12 months' fees. The exclusion of consequential damages is the most frequently litigated provision in tech MSAs.
Indemnification
Who defends and compensates the other party when a third-party claim arises. IP infringement indemnity by the service provider is standard. Cross-indemnities require careful balance.
Term and termination
How long the MSA lasts, auto-renewal provisions, notice periods for termination for convenience, and triggers for immediate termination for cause.
Warranties
What the service provider promises about the quality and compliance of its services. Disclaimers of implied warranties ("as-is" delivery) often appear here and reduce the service provider's exposure significantly.
Governing law and jurisdiction
Which country's or state's laws apply, and which courts or arbitration bodies have jurisdiction. For cross-border relationships, this clause can be more valuable than any other.
Non-solicitation
Prevents each party from hiring the other's employees or contractors during and after the engagement. Duration, scope (all staff vs key personnel), and remedies for breach must be specified.
Intellectual property: the clause neither party fully understands
Intellectual property ownership in an MSA context is genuinely complex because service delivery almost always involves three categories of IP simultaneously: pre-existing IP that the service provider brings to the engagement (its proprietary tools, frameworks, methodologies, code libraries), custom work created specifically for the client under the SOW, and derivative works that combine elements of both.
The naive approach — "the client owns everything we create for them" — is commercially catastrophic for service providers. A development agency that assigns full ownership of every line of code it writes will eventually give away its core technology stack to a client. The sophisticated approach separates ownership by category.
Background IP (pre-existing) remains with the service provider permanently. Foreground IP (custom deliverables) typically transfers to the client upon full payment. Derivative IP — where the deliverable is built on top of the provider's background IP — is handled through a perpetual, irrevocable licence rather than an assignment, so the client can use and modify the deliverable without owning the underlying engine.
Clients, for their part, should ensure that the licence granted over background IP embedded in deliverables is truly perpetual and survives termination of the MSA. A licence that expires when the MSA ends leaves a client unable to use a system it paid millions to build, should the relationship end.
Open-source software embedded in deliverables creates a separate IP obligation. If a service provider uses GPL-licensed libraries in custom work, the client may be required to open-source any modifications under the same licence. MSAs should require the service provider to disclose all open-source components and identify their licences before delivery.
The moral rights problem
In jurisdictions that recognise moral rights (most of Europe, India, and many other common law countries), creators retain certain rights over their work even after assigning economic ownership — including the right to be identified as the author and the right to object to derogatory treatment of the work. IP assignment clauses in MSAs governed by these jurisdictions should include an explicit waiver of moral rights to the extent permitted by law.
Limitation of liability: the most negotiated, most misunderstood clause
The limitation of liability clause does two distinct things that are often conflated. First, it caps the total financial exposure of each party — typically at an amount equal to the fees paid in the 12 months preceding the event giving rise to the claim. Second, it excludes certain types of damages entirely — most commonly, indirect and consequential damages, lost profits, and loss of data.
The second part is often more significant than the first. A service provider that delivers a faulty software system to a logistics company could, in theory, be liable not just for the fees paid for the project, but for every shipment the logistics company failed to fulfil during the outage — lost revenue that could dwarf the original project fee by a factor of 100. The consequential damages exclusion eliminates this exposure entirely.
| Clause variation | Favours | Practical effect |
|---|---|---|
| Cap = 12 months' fees | Balanced | Standard market position. Predictable exposure for both sides. |
| Cap = total fees paid | Service provider | For long engagements, cap grows. Reasonable for high-value, long-term work. |
| Cap = 1 month's fees | Service provider | Often seen in provider-drafted agreements. Provides minimal protection to clients. |
| No cap on gross negligence | Client | Carve-out for wilful misconduct. Fair — no party should be able to escape liability for intentional harm. |
| No cap on IP indemnity | Client | If provider's work infringes third-party IP, client exposure is uncapped. Reasonable given asymmetry of knowledge. |
| Full consequential damages exclusion | Service provider | Standard in tech. But for critical infrastructure, clients should seek carve-outs for data breach and prolonged outage scenarios. |
A liability cap of one month's fees on a ten-million-dollar project is not a limitation of liability. It is a near-complete transfer of risk to the client.
Mutual vs asymmetric caps
Many MSAs drafted by large enterprises apply the liability cap only to the service provider, while leaving the client's exposure (e.g. for non-payment or breach of confidentiality) uncapped. This asymmetry is commercially aggressive and should be resisted. A well-negotiated MSA applies the same cap bilaterally, with agreed carve-outs for specific high-risk scenarios on both sides.
Termination: exits, triggers, and the consequences of getting this wrong
Termination provisions govern both the routine end of a commercial relationship and its emergency termination when things go badly wrong. A poorly drafted termination clause can trap a party in a failing relationship for months, or expose them to significant financial liability for a reasonable exit.
Termination for convenience
This allows either party to end the MSA (and typically all active SOWs) without cause, simply by giving adequate notice — usually 30 to 90 days. This is a standard and reasonable provision. Clients value it because it prevents lock-in. Service providers value it because it allows them to exit problematic relationships without having to prove a breach.
The key negotiation point is the notice period and its effect on in-flight work. Does a 30-day notice period require the service provider to complete work in progress? Are milestone payments accelerated on termination? Is there a kill fee for projects that are terminated mid-delivery?
Termination for cause
This allows immediate or expedited termination when a specific trigger occurs: material breach, insolvency, change of control, failure to cure a breach within a defined period, or regulatory violation. The definition of "material breach" is critical — and almost always contested when invoked. Courts have generally held that a breach is material if it defeats the purpose of the contract; smaller failures typically require notice and a cure period.
- Convenience termination: 60 days' written notice. All active SOWs continue to completion or wind down per their own terms unless otherwise agreed.
- Cause termination (curable breach): Written notice specifying the breach. 30 days' cure period. If uncured, immediate termination.
- Cause termination (incurable breach): Immediate written notice. Triggers include insolvency, criminal conviction of key personnel, material data breach, and wilful misconduct.
- Effect of termination: All outstanding payments become immediately due. IP transfers (if earned) are confirmed. Confidentiality obligations survive for the agreed term. Non-solicitation survives for 12 months.
Governing law, jurisdiction, and dispute resolution
For domestic contracts between parties in the same country, the choice of governing law is rarely controversial. For cross-border engagements — a US client and an Indian service provider, for example — it is one of the most commercially significant decisions in the entire agreement.
Governing law determines which jurisdiction's legal principles are used to interpret and enforce the contract. Jurisdiction determines which courts or arbitration bodies have authority to hear disputes. These are two distinct choices, and they need not be the same — a contract can be governed by English law but submit disputes to Singapore arbitration.
Why arbitration is usually preferred for B2B disputes
International commercial arbitration offers several advantages over litigation in cross-border disputes: confidentiality (court proceedings are public), enforceability (arbitral awards are enforceable in over 160 countries under the New York Convention, while court judgments often are not), speed relative to most court systems, and the ability to choose arbitrators with domain expertise.
Common arbitration seats for India-related international contracts include Singapore (SIAC), London (LCIA), and Hong Kong (HKIAC). For purely domestic Indian contracts, arbitration under the Indian Arbitration and Conciliation Act, with a seat in a major commercial city, is standard.
A governing law clause that says "the laws of [State/Country]" without specifying which dispute resolution mechanism applies defaults to litigation in that jurisdiction's courts. For international contracts, this can mean spending more on enforcement than the dispute is worth. Always specify the mechanism explicitly.
Common MSA mistakes — and how to avoid them
After reviewing hundreds of MSAs across industries, the same errors appear repeatedly. They are rarely about bad intent. They are almost always about insufficient attention during drafting, or misplaced confidence that "standard terms" are adequate for every situation.
- Using a template without adapting it. Template MSAs are starting points, not finished agreements. A technology consulting agreement and a marketing retainer agreement have different risk profiles, different IP considerations, and different SLA requirements. Using the same template for both is a mistake.
- Leaving IP ownership to implication. The single most expensive mistake. In most jurisdictions, absent an explicit assignment, the creator owns the IP. Clients who pay for custom software development and receive only a licence — not ownership — are often genuinely shocked when they discover this later.
- Not defining acceptance criteria. A service provider who delivers work "to industry standard" and a client who expects "perfection" are both reading the same MSA and seeing different promises. Acceptance criteria must be objective, measurable, and agreed in the SOW.
- Ignoring the effect of termination on active SOWs. What happens to a six-month project when the MSA is terminated after three months? Most MSAs are silent on this. The answer should be explicit: either the SOW survives until completion, or it terminates with the MSA, with clear financial consequences either way.
- Asymmetric confidentiality obligations. A mutual NDA in the MSA should impose the same standard of care on both parties. MSAs drafted by large enterprises sometimes impose stricter obligations on the service provider (including certifications, audits, and specific data handling requirements) without reciprocal obligations on the client — even when the client receives equally sensitive information.
- No change request mechanism. Every long-term service relationship experiences scope changes. Without a formal change request process in the MSA — requiring written agreement and pricing before any out-of-scope work begins — providers do the work informally and then cannot collect payment for it.
- Failing to review after material changes in the relationship. An MSA signed when a company was paying ₹10 lakh a year should be revisited when the relationship grows to ₹2 crore. Liability caps, insurance requirements, and data handling obligations appropriate for a small engagement may be dangerously inadequate for a large one.
Negotiation strategy: how to approach the MSA process
MSA negotiation is, at its core, a risk allocation exercise. Each clause transfers risk from one party to the other. The goal is not to "win" — it is to reach an allocation that reflects the actual risk profile of the relationship, the bargaining leverage of each party, and the commercial reality of the industry.
Identify your non-negotiables before you sit down
Before entering any negotiation, know the three to five clauses where you will not move and the ten to fifteen where you are willing to give ground. Service providers typically hold firm on liability caps, consequential damage exclusions, and payment terms. Clients typically hold firm on IP ownership, data security standards, and audit rights. Knowing your own hierarchy prevents you from spending negotiation capital on provisions that do not matter while conceding on those that do.
The "standard terms" gambit
Almost every counterparty will describe their MSA as "standard terms." This is almost never true — every MSA is drafted to favour the party that drafted it. The appropriate response to "these are our standard terms" is "we are happy to use your template as a starting point and mark it up." A refusal to accept any markups is a red flag about how disputes will be handled in the relationship.
Sequencing matters
Address high-stakes clauses — IP, liability, termination — early in the negotiation when both parties are motivated to close the deal. Leaving these for last, after the commercial terms are agreed, creates asymmetric pressure: the party that needs the relationship more has less leverage to hold out on legal provisions.
The best MSA negotiation is one where both parties leave the table believing the agreement is fair. A contract that one party resents is a liability, not a protection.
A final word: the MSA as relationship infrastructure
The best commercial relationships rarely invoke their MSA. The agreement sits in a drawer, its provisions rarely tested, because both parties are working in good faith toward shared goals. But the existence of a well-drafted MSA changes the character of the relationship in subtle and important ways.
It creates clarity. When everyone knows who owns what, who pays when, and what happens if things go wrong, there is less ambiguity for resentment to grow in. It creates accountability — performance standards are written down and cannot be revised retrospectively. And it creates trust, paradoxically, because the willingness to negotiate a fair agreement signals that each party takes the other seriously as a commercial partner.
Invest in the MSA. Get legal advice if the relationship is substantial. Negotiate it honestly. And then, with any luck, put it in a drawer and never think about it again.
Comments
Post a Comment