Scope creep happens when the deliverable list is vague and the fee is fixed. Fix it with a written scope, a named change-order process, and a monthly review where new work is absorbed, scheduled, or priced. At Flapen every tier includes all 50+ services, which removes most of the argument before it starts.
The short version
- Vagueness is the cause, not greed. Nobody writes "unlimited creative" intending to fight about it later.
- A change order needs three fields. What changed, what it costs, and who approved it. Anything shorter becomes a dispute.
- Price by product count, not by task count. Task-based scopes make every request a negotiation.
- Review scope monthly, not annually. Twelve small conversations beat one angry one.
- Watch what happens when targets move. A shifting advertising goal is the most common disguise for extra work.
Why scope creep is structural
An Amazon engagement is not a fixed project. Inventory arrives late, a competitor drops price, a listing gets suppressed, a variation needs splitting. Every one of those creates work nobody scoped, because nobody could have. So the scope document is not really a list of tasks. It is a set of rules for what happens when the list turns out to be wrong.
Two structures cause almost all of the friction I see. The first is a fee attached to a task count: ten listings, four creative refreshes, two campaigns per month. The eleventh listing then becomes a commercial event. The second is a fee attached to a percentage of ad spend, where the agency's income moves with the budget and "extra work" mysteriously always points toward spending more.
We price by product count instead, from $800 a month for one product up to $2,400 for five, with everything included at every tier. Six or more products gets scoped on a call. I am not claiming this is the only fair structure, but it does mean nobody on my side profits from redefining what is in scope.
The six-stage scope process, with a gate at each stage
- Write the scope before the contract. List the functions, not the volumes. Listing optimization, creative, advertising, sourcing support, reporting. Gate: if a function is not named, it is not included, and both sides agree to that in writing.
- Name the exclusions explicitly. Inventory, Amazon's own seller fees, trademark filing, freight, and ad spend sit outside any management fee worth paying. Gate: the proposal separates fee from pass-through cost on its own line.
- Set the standing cadence. Ours is a written update each week and a live review every two weeks, with Slack open in between. Gate: the cadence is in the agreement, not the sales deck.
- Define the change-order trigger. Decide the threshold at which new work stops being absorbed. A useful default is anything that consumes more than a day of a specialist's time. Gate: the threshold is a number, not a judgment call.
- Run a monthly scope review. Fifteen minutes. What did we do that was not in the original list, and what should we do about it. Gate: the outcome is written down before the call ends.
- Re-price only at renewal points. Month-to-month terms with 30 days' notice mean either side can act on the review rather than argue inside a locked annual contract. Gate: no mid-term fee change without a signed change order.
Three kinds of change, and what each should cost
| Change type | Example | Fair treatment |
|---|---|---|
| Within scope, unplanned | A listing suppression that needs same-week fixing | Absorbed. This is the job |
| New volume, same function | You add two products to the catalog | Move to the next pricing tier, no separate charge |
| New function entirely | Sourcing a new supplier when the deal was marketing only | Written change order, priced before work starts |
The middle row is where most agencies quietly overcharge and most clients quietly overpay. If the fee is tiered by product count, adding a product should move you one tier, full stop. If a proposal charges a per-product setup fee on top of the tier move, ask what the setup fee buys that the tier does not.
The moving-target problem
The most common disguise for scope creep is a target that changes without an agreement. Advertising cost of sale is the usual vehicle. The right target does change by product stage, aggressive during a launch and efficient at maturity, and any competent operator will tell you that. The problem is when the number moves without anyone saying so, because a moved target quietly reframes work that failed as work that was never in scope.
Fix it in one line. Ask, before signing, for the launch target and the maturity target as two separate numbers, and for the criteria that move a product from one to the other. Then any change to those numbers becomes a visible decision rather than a retroactive explanation.
What most agencies will not tell you about change orders
The change-order clause is usually written by the party that benefits from ambiguity, and it is almost never negotiated because it feels procedural next to price and term. It is not procedural. It is where the real price of the engagement gets decided over twelve months.
The other thing worth saying: some scope creep runs the other way. Clients send requests at 11pm on Slack, ask for decks for their investors, and add products without mentioning it. A good agreement protects both sides, which is why I prefer a written monthly review to a rigid task list. It surfaces drift in both directions while it is still small.
Related answers
- How to negotiate Amazon brand management contracts
- Contract terms to negotiate with Amazon agencies
- Month-to-month vs annual Amazon contracts
- What services do I need to start selling on Amazon
- Hiring an Amazon agency: the complete guide
Our full service list and tier pricing are published, so you can copy the scope language straight from Flapen.
