Normalise scope before you normalise price
The most common mistake is comparing monthly fees before checking that the proposals cover the same work. One agency's retainer might include content and listing optimisation while another quotes them as separate line items. Build a simple table listing every workstream (advertising, content, operations, strategy, reporting) and mark, for each proposal, whether it sits inside the base fee or is an add-on.
Once scope is normalised, the price comparison starts to mean something. A higher monthly fee that includes operations and content is often better value than a lower fee that covers only advertising and leaves the rest for you to handle or pay for separately.
Look closely at month one compared with the months after it. Some proposals fold a substantial setup or audit fee into the first month's rate, which makes the ongoing rate look higher than it really is once spread out. Ask each agency to show onboarding or audit costs separately from the steady-state retainer.
What the scope table should cover
A scope table works best when each row is specific enough that two agencies can't both tick it while meaning different things. Rows we would include:
- Sponsored Products, Sponsored Brands and Sponsored Display management, and whether Amazon DSP is in or out.
- Keyword research and search term harvesting, and how often it happens.
- Listing copy and images: how many ASINs a month, and whether new images or video are included or only briefs.
- A+ content and Brand Store work.
- Catalogue and account health issues, such as suppressed listings, cases with Seller Support and reimbursement claims.
- Inventory and replenishment advice.
- Reporting: format, frequency, and whether a change log is included.
- Meetings: how often, and with whom.
- Additional marketplaces: included, extra cost, or not offered.
For each row, record "included", "extra at £X / $X", or "not offered". Any blank cells are questions to put to the agency before you go further.
Read the pricing model for its incentive as well as its number
Percentage of ad spend rewards an agency for spending more. That is fine while you are scaling, and it can pull in the wrong direction the moment the right move is to cut spend and fix conversion first. Percentage of revenue aligns better in principle, but it can reward an agency for revenue it had little to do with, particularly around seasonal peaks. Flat retainers put the burden on you to define scope precisely, and they survive a decision to cut spend without the agency's incentive working against it.
No model is right for everyone. The question is whether the incentive it creates matches what you need from the engagement right now. A brand in aggressive growth mode may be comfortable with a spend-based model. A brand protecting margin on an already large account usually isn't.
Model each fee at three spend levels
Headline fees are easiest to compare when you run each one through the same scenarios. Take your current monthly ad spend, a figure about a third lower (the "we need to cut back" case) and a figure about a third higher (the "we are growing" case), and calculate what each proposal would cost at all three.
For example, with illustrative numbers: at $30,000 of monthly ad spend, a proposal charging 10% of spend costs $3,000; a flat retainer of $3,500 costs $3,500; and a hybrid of $1,500 plus 5% of spend costs $3,000. At $20,000 of spend, those become $2,000, $3,500 and $2,500. At $40,000, they become $4,000, $3,500 and $3,500. The flat retainer is the most expensive when you cut back and the cheapest once you grow. Laying it out like this shows which model suits the year you expect to have.
Pressure test the case studies the same way across every proposal
Apply one standard to every case study you are shown. What was the baseline revenue? Was the category growing on its own, regardless of the agency's work? Did new products launch during the period? How much did ad spend rise alongside the growth being claimed? Agencies with real results give you this context without resistance. Vague or defensive answers tell you more than the growth percentage does.
Ask each agency for one reference client in a category and revenue range comparable to yours, and ask that reference about the worst month of the engagement, not the best. How an agency communicated during a bad month tells you more about the working relationship than any growth number.
Check whether the case study reflects the team that would work on your account. Case studies often feature an agency's most senior people while your account would be run by someone more junior. Ask directly who would be assigned and look at that person's track record.
It also helps to ask how many accounts that person currently manages. An answer that is specific, even if higher than you would like, is a better sign than a vague one. Our article on why Amazon PPC agencies fail explains why account load matters so much.
Contract terms that decide whether you regret it later
Compare notice periods across every proposal. A long lock-in with a short notice window is the clause most likely to cause regret if the relationship doesn't work out, and it is worth negotiating even with an agency you feel confident about.
Confirm who owns advertising data, creative assets and account access if you leave. The answer should be that you keep full ownership and the agency has scoped access that you can revoke. If leaving would mean starting from zero, push back before you sign.
Get any exclusivity clauses in writing. Some agencies require exclusivity across all your Amazon marketplaces or all your advertising channels, which can be reasonable or can box you in unnecessarily depending on the wording. Read this section closely.
In practice, account access should work through your own logins. You stay the primary owner of Seller Central and the Amazon Ads account, and you add the agency as a user with the permissions it needs. Ask each agency whether it expects to work this way. One that asks for your main login, or wants to open the ad account under its own name, is creating a problem for the day you part ways. Our guide to account and data ownership covers what to check.
A simple scorecard for the final decision
After normalising scope and price, score each proposal out of five on the same criteria and write one sentence of evidence for each score. The evidence matters more than the number, because it stops one impressive pitch from colouring every row.
- Scope fit: does it cover what you need without paying for what you don't?
- Cost at your three spend scenarios.
- How well the pricing incentive fits your goals for the next year.
- Quality of the case study answers and the reference call.
- The named person who would run your account, and their current workload.
- Contract terms: notice period, lock-in, exclusivity, ownership.
- Reporting: would you see what changed and why, or only top-line charts?
Weight the rows to suit your situation. A brand worried about lock-in should weight contract terms heavily, while a brand with a weak catalogue should weight scope fit. If two proposals end up close, pick the one with the clearer reporting and the easier exit. Those are the two things that make a wrong choice cheap to fix.
Red flags in a proposal
- A guaranteed ACoS, sales figure or ranking. Nobody controls Amazon's auction, competitors or seasonality closely enough to guarantee them.
- A proposal written before the agency has seen any of your account data, with specific promises in it.
- Pricing that is only available if you sign within a few days.
- No named account manager, or a refusal to say how many accounts that person runs.
- Reporting described only as "monthly reports", with no sample offered.
- Contract terms that make it hard to take back account access or data.
For a fuller list of questions to put to each agency on your shortlist, see our guide to questions to ask an Amazon agency.
Frequently Asked Questions
Three is usually enough to see real variation in approach without the comparison becoming unmanageable. With fewer than three you have no real basis for comparison. More than four or five tends to give diminishing returns and mostly costs you time in repeated discovery calls.
Not until you have normalised scope. A cheaper proposal that leaves out operations, content or reporting often costs more once you count what you will have to source separately or do yourself. Compare total cost for the same scope instead of the headline monthly figure.
It depends on account complexity, catalogue size and how much structural work the account needs, so there is no universal benchmark. What matters is that it is quoted separately from the ongoing retainer, so you can see what you are paying for in month one compared with month six onward.
Some agencies offer one and some don't, and it is reasonable to ask either way. If a full trial isn't offered, a shorter initial term with clearly defined success criteria can work as a lower-risk alternative.
Ask for the baseline revenue, the category growth trend, whether new products launched in the period, and the change in ad spend alongside the claimed growth. Without that context the number can't be verified, and an agency confident in its results will provide it readily.
Giving shortlisted agencies limited, read-only information, such as exported reports or a screen share, helps them write proposals based on your account instead of guesses. Avoid handing over full account access before you have signed anything. If an agency needs access for an audit, add it as a user with restricted permissions and remove it if you don't go ahead.