Scoping An Agency Engagement So Both Sides Can Win

Most agency relationships in India do not fail on capability. They fail on scope. Two competent parties spend nine months quietly disagreeing about what was bought, and then one of them leaves.

Scope is not paperwork. It is the operating system of the engagement. Write it badly and every later conversation becomes a negotiation.

Choose the commercial model before you write the scope

There are three common shapes, and each buys something different.

A retainer buys capacity and continuity. It suits work that never ends, such as marketplace account management, catalogue hygiene, ads operations and support. Its failure mode is silent shrinkage. The fee stays flat while the attention drifts to whichever client shouted last. The fix is to name the team and the seniority in the document, and to review actual output against the assumed effort every quarter.

A project buys a defined output with an end date. It suits migrations, launches, catalogue rebuilds and one time integrations. Its failure mode is the orphan period after handover, when nobody owns the thing that was built. Decide who owns it on day one of the project, not on the last day.

Performance linked pricing buys alignment in theory. In practice it works only when the metric is clean, the agency genuinely controls it, and both sides read it from the same source. If the agency does not control pricing, stock availability or platform level decisions, a pure commission on revenue punishes them for your stockouts and rewards them for your festive calendar. A base fee plus a variable component tied to something the agency can actually move is usually the more honest structure. Where a share is quoted, the percentage varies widely by category and by how much the agency is expected to fund upfront.

What the scope must fix, and what it must leave loose

Fix the things that decide cost and accountability. The channels and marketplaces in scope, by name. The SKU count band, with a stated trigger for what happens above it. The named team, their roles and their seniority. The deliverable cadence. The metric definitions, written down, so that both sides compute return on ad spend and contribution the same way. Turnaround commitments in both directions, including yours. A single approval owner on the client side. The commercial, and the specific events that change it.

Leave loose the things that should improve with learning. Tactics. Campaign structures. Channel mix within an agreed budget. Creative direction inside brand guardrails. The order of work inside a sprint. If you fix these, you have hired a pair of hands and you will get the results of a pair of hands.

The most valuable line in any scope document is the one that defines what is not included. Buyers skip it because it feels adversarial. It is the opposite. An explicit exclusion list is what allows the included work to be delivered properly.

Deliverables versus outcomes

You need both, and they do different jobs.

Deliverables are countable and enforceable. Listings published. Campaigns live. Reports delivered. They protect you in the first ninety days, when there is not yet enough data to judge results, and they give the agency something concrete to be measured on while the account is still being repaired.

Outcomes are what you are actually buying. Share of category. Efficiency band. Availability. They protect you after month four, when the deliverables are all green and the business is flat.

A scope with only deliverables buys motion. You get twenty reports and no growth, and technically nobody breached anything. A scope with only outcomes buys arguments, because outcomes depend on inputs the agency does not own. Write both. State which deliverables are the leading indicators for which outcome, so a review can trace one to the other.

How scope creep actually starts

Not with a large request. It starts with a small favour on a message thread at nine in the evening. The agency says yes, because saying yes feels like service. Nobody logs it. Three weeks later it is a standing expectation, and it is now competing for hours with the work you are paying for.

It also starts on the client side, and this is the half buyers rarely examine. You add two marketplaces. You add a founder who wants a weekly deck. You add a new category with a different compliance profile. None of it was priced. All of it lands on the same team.

Here is the part that is unhelpful for agencies to say and true anyway. A vague scope damages the client more than the agency. The agency has defined revenue and an undefined obligation, so it rations effort, and the rationing is invisible to you until results slip. The client has defined cost and undefined delivery. That is a bad trade in every month of the contract.

Which leads to a warning sign worth taking seriously. If an agency accepts a loose scope without pushing back, that is not flexibility and it is not eagerness to help. It means they intend to interpret the gap in their favour later, or they have not costed the work at all. Either way you are the one carrying the risk. The bidder who argues with your brief before signing is usually the safer choice.

The review cadence that catches drift early

Three loops, and they are not the same meeting.

Weekly is operational. Blockers, escalations, what moved, what is stuck and who owns the unblock. Thirty minutes. No slides.

Monthly is performance. Outcomes against the agreed band, spend against plan, and an honest read of what did not work. Insist that the agency brings one thing that failed. A monthly review with no failures in it is not a review.

Quarterly is scope. This is the meeting almost every brand skips, and it is the one that prevents the eventual breakup. Open the document. Walk the list. Ask what is now being done that is not written down, and what is written down that nobody has asked for in three months. Trade them. Reprice if the trade is uneven.

Scope reviewed four times a year rarely produces a surprise. Scope opened once at renewal almost always does.