How to write a SOW that does not start a fight in week three
A scope of work fails in week three or it never fails at all. The fights about what was promised, what is extra, and who is paying for which assumption all surface around then. Here is the shape that holds, and the section most agencies still underwrite.
Why SOWs blow up in week three
Not in week one, because the client is excited. Not in week two, because the team is still gathering inputs. Week three is the first time someone on the client side compares the work in flight to the page they signed, and the gap between "what we discussed" and "what we wrote" turns into a meeting.
The fix is almost always in the document, not in the conversation. A SOW that lists deliverables without phases, or phases without assumptions, or an "out of scope" section that says nothing concrete, is a document built to argue with later.
The shape that survives
A SOW that holds reads in roughly this order.
Engagement summary. Two paragraphs. Who is doing what for whom, and what success looks like at the end. Plain language. No buzzwords. If your account lead would not say this sentence out loud, do not write it.
Phases. Three to five, with a name, a goal, and a date range. Phases are how clients track progress. Skipping them in favor of a flat list of deliverables makes the work feel timeless, which is bad for both sides.
Deliverables, by phase. Specific. "Strategy doc, up to 20 pages, two rounds of revisions" beats "strategy work." If you cannot describe the size and shape of the artifact, the team will inflate it and the client will expect it.
Assumptions. The most undersung section in any SOW. Write the assumptions that, if wrong, would change the scope. Access to data, response times, stakeholders available for review, content provided by client. If an assumption breaks, the change order writes itself.
Out of scope. The short list of obvious things that look like they belong in this work but are not in this engagement. SEO when the project is paid media. Branding when the project is web. Without this list, "we thought that was included" is a fair argument.
Pricing and payment terms. Numbers, what they cover, and when they are due. If pricing is a single line, the proposal was clearer than the contract is.
Change orders. One short paragraph. The mechanism for adding scope and the rate it bills at. If the mechanism is "we will figure it out," that is how week six will feel.
Phases, not features
The biggest mistake in agency SOWs is listing deliverables without grouping them. A flat list of fourteen items feels like a shopping cart. The same fourteen items grouped into four phases feels like a plan. Clients buy plans, not carts.
A phase is also where the team protects itself from scope creep. If a new request is on the table, it either belongs to a phase or it does not. Without phases, every request looks like part of the work.
Assumptions are the contract
If a SOW only lists what your team will do, it is half a contract. The other half is what has to be true for that work to land. Stakeholders available in week one. Access to analytics by day three. Content drafts back inside seven business days. None of these are unreasonable. If they go unsaid, the team that misses them is the agency, not the client.
A serious SOW often has more assumptions than deliverables. That is not paranoia. That is reading week three before it happens.
The out of scope section
Two or three lines. Things that came up in conversation and could plausibly be expected. Naming them does two things at once: it protects the engagement, and it tells the client that you listened. Most SOWs skip this because it feels negative. The clients who notice it read it as competence.
Change orders without drama
The phrase to write into every SOW is something close to: changes outside the scope above will be quoted as a change order at the rates listed, and signed in the same way as this SOW. That is it. Once it is in there, change orders stop feeling like upsells and start feeling like the obvious next step.
A short sniff test
Read the SOW pretending to be the client. If you can name what you are getting, when, for how much, and what would change the price, the document is doing its job. If any of those is fuzzy, week three will fix it for you in less pleasant ways.
Where Docsiv fits
We built Docsiv so the SOW is a real document type, not a glorified Word file. Phases, deliverables, assumptions, out-of-scope, and pricing all have proper blocks. Templates compose, the AI draft fills in your voice, and the team finishes the parts that need judgment. Sign it inside the same workspace, then run the engagement from the same branded portal the client will live in for the next six months.
Frequently asked questions
Tap a question to expand the answer. The same content is in structured data on this page for search.
What goes into a strong agency SOW?
Engagement summary, three to five phases with goals and dates, deliverables grouped by phase, assumptions, an out-of-scope list, pricing and payment terms, and a short change-order paragraph. That order is what survives week three.
Why does an agency SOW need an assumptions section?
Because half of what kills scope is not what the team will do, it is what has to be true for that work to land. Naming the assumptions explicitly is what makes change orders feel obvious instead of confrontational.
What is the difference between deliverables and phases in a SOW?
A deliverable is an artifact. A phase is a chunk of time with a goal. A SOW with deliverables but no phases reads like a shopping cart. A SOW with phases reads like a plan, and clients buy plans.
How should change orders be handled in an agency SOW?
Define the mechanism in the original SOW: changes outside scope will be quoted as a change order at the rates listed, and signed in the same way. Once that paragraph exists, change orders stop feeling like upsells.
How long should an agency SOW be?
Long enough to be specific. For most engagements that is three to seven pages. Shorter SOWs underwrite assumptions; longer SOWs usually have padding nobody reads. The signal of a good SOW is density, not length.
Related posts
Reusable, not recycled: agency templates clients do not notice
A good template feels like a tailored document. A recycled one feels like a copy-paste with the names swapped. The difference is in how the template was built, not in how hard the writer tries.
Docsiv Team
Your pricing page is buried on page 9. Move it.
Buyers find pricing first anyway. Hiding it on page nine does not protect you, it just makes the proposal harder to read. A short case for putting numbers near the top.
Docsiv Team
PandaDoc alternative for agencies: when a doc tool is not enough
PandaDoc is good at what it does. The thing is, agencies rarely need just a doc tool. Here is the honest version of where the seams start showing up, and what to consider instead.
Docsiv Team