Picture a mid-sized Pune software firm that has just signed a fixed-price deal to build a dealer portal for a two-wheeler manufacturer. Everyone is pleased. Four months later the client’s operations head asks why the portal does not show live inventory from the plant’s ERP system. The sales team remembers talking about it. The delivery team says it was never in scope. The client says it was “obviously” part of a dealer portal. Nobody is lying. The problem is that the Statement of Work never said either way.

That scenario, in one form or another, is the most common way services projects lose money. The Statement of Work (SOW) is the document meant to prevent it.

Where the SOW sits in the contract stack

Most enterprise services relationships use two layers of paper. A Master Services Agreement (MSA) carries the long-lived legal terms: liability caps, confidentiality, intellectual property, payment terms, termination rights and governing law. It is negotiated once and can run for years. Each individual piece of work then gets its own SOW, which references the MSA and describes the specific project: what, when, how much, and on what basis.

This split is practical. Legal teams do not have to renegotiate indemnities for every new project, and the commercial and delivery teams can focus each SOW on the work itself. The contract usually states which document wins if the two conflict, and it is worth knowing which one that is before you write anything into an SOW that touches legal territory.

What a good SOW actually contains

Templates vary by company, but a solid SOW covers these areas:

  • Objectives and background. A short statement of the business problem, so both sides interpret detail in the same spirit.
  • In-scope items. Specific deliverables, modules, integrations, environments, user volumes and languages.
  • Out-of-scope items. Things a reasonable reader might assume are included but are not. In the dealer portal example, “integration with plant ERP” belongs here if it was not priced.
  • Assumptions. The conditions the price depends on.
  • Client responsibilities. Access, data, decisions, test users, third-party coordination.
  • Timeline and milestones. Usually linked to payment.
  • Acceptance criteria and process. How a deliverable is judged complete and who signs it off.
  • Commercials. Pricing model, rates or fixed fees, payment schedule, expenses, taxes such as GST.
  • Governance and change control. Who meets when, how issues escalate, and how changes are raised and approved.

Assumptions do the heavy lifting

Experienced presales people will tell you the assumptions section matters more than the scope list. Scope says what you will build. Assumptions say what has to be true for your price and timeline to hold. Examples: the client will provide test data within two weeks of kick-off; there will be no more than two rounds of UI review; the existing APIs are documented and stable; the client will make design decisions within three working days; production deployment will be done by the client’s infrastructure team.

Each of these is a lever. If the client takes six weeks to provide test data, the assumption has been broken and the vendor has a documented basis to revisit timeline or cost. Without the assumption written down, the delay simply becomes the vendor’s problem.

The common mistake is writing assumptions so vague that they protect nobody, such as “client will cooperate in a timely manner”. A useful assumption is specific, testable and tied to something the client controls.

Acceptance: deciding in advance what “done” means

Disputes at the end of a project are almost always disputes about acceptance. A good SOW defines what is tested, against which requirements, by whom, within how many days, and what happens if the client neither accepts nor rejects in that window. Many contracts include deemed acceptance: if no defects of a stated severity are reported within, say, ten working days, the deliverable is treated as accepted. Without that, a milestone payment can sit unpaid for months while the client “reviews”.

Acceptance criteria should also separate defects from new wishes. A screen that crashes is a defect. A screen that works as specified but that the business now wants laid out differently is a change.

Change requests: the pressure valve

No SOW anticipates everything, so every services contract needs a change control process. A change request (CR) is the formal mechanism for altering scope, timeline or price after signing. The usual flow is:

  1. Either side raises the request in writing, describing what should change.
  2. The vendor performs an impact analysis: extra effort, effect on schedule, effect on other deliverables, new risks.
  3. The vendor submits a priced proposal.
  4. The client approves, rejects or negotiates.
  5. Once signed by authorised people on both sides, the CR amends the SOW.

Back to the dealer portal. If the ERP integration was listed as out of scope, the vendor can raise a CR. Suppose the impact analysis estimates 120 hours of additional work at ₹2,500 an hour: a ₹3 lakh change, plus three extra weeks. The client can now make an informed choice. Perhaps they approve it, perhaps they defer it to a phase two. Either way the conversation is commercial, not accusatory.

A few habits keep CRs healthy. Small requests should still be logged even if the vendor decides to absorb them free of charge, because a record of goodwill is useful in later negotiations. Work on a change should not begin before approval, or the vendor loses its bargaining position. And under T&M contracts, CRs still matter: they keep the budget visible even if the vendor is not carrying the cost risk.

Why salespeople, specifically, must get it right

The SOW is usually drafted while the deal is still being won, under time pressure, by people motivated to close. That is exactly when scope gets loose: a feature mentioned in a demo becomes an implied promise, a client’s assumption goes unchallenged to keep the mood positive. Delivery teams inherit whatever was signed. A vague SOW does not save the deal; it moves the argument from before signature, when the vendor has leverage, to after signature, when the vendor has a team already deployed and a reputation to protect.

The best practice is simple to state and hard to follow: involve the delivery lead in drafting, read the SOW as if you were the client’s most sceptical project manager, and treat every sentence that could be read two ways as a future dispute.

Written by the RP Strativa Team. The companies and figures in the examples are illustrative.

Leave a Reply

Trending