Amazon Marketing Consultants: What They Actually Do, and How to Judge Their Numbers

What an Amazon marketing consultant actually does

The title covers two very different jobs. An advisory consultant audits your account, hands you a written strategy, and leaves execution to you or your team. A managed consultant does the advisory work and then runs the campaigns, writes the listings, and reports the results themselves. Most agencies that call themselves consultants are actually the second kind, even when the word on the page is "consulting."

Inside that, the scope varies by shop. Common tasks include:

• Account health audit and suppression/compliance checks
• Sponsored Products, Sponsored Brands, and Sponsored Display campaign management
• Listing and A+ content optimization for keywords and conversion
• Inventory forecasting, FBA prep, and storage limit management
• Brand store build and enhanced content
• DSP and off-Amazon media, where the consultant holds it

Before you sign anything, ask which of these you're actually getting. A firm strong on compliance and operations (FBA shipments, CPSC documentation, storage appeals) is solving a different problem than one strong on media buying. Both call themselves "Amazon consultants." Neither is wrong to. But they're not interchangeable, and a brand with a listing suppression problem doesn't need a DSP specialist, and a brand with flat ad ROAS doesn't need a shipping-label expert.

How consultants price their time — and the number to actually ask for

Four structures cover almost every engagement: hourly or per-project (audits, one-off listing rewrites), monthly retainer (a fixed fee for a defined scope of work), percentage-of-ad-spend (the fee scales with your budget), and full managed service (a base fee plus a variable rate, often blended with spend-based pricing for media). None of these is inherently better. A retainer suits a brand that needs steady hands-on-keyboard work. Percentage-of-spend suits a brand that's scaling budget fast and wants the consultant's incentives aligned with growth.

The number that matters is the one most listicles and even vendor sites don't print clearly: the variable rate on top of any base fee. A percentage that looks trivial at $10,000 a month in ad spend is not trivial at $100,000 a month — one percentage point there is roughly $12,000 a year, on top of whatever the base fee already covers. Ask for the full structure, not just the headline number, before you compare two proposals side by side. If a firm won't say how the percentage is calculated or at what spend tier it changes, that's information too.

A worked example: reading a consultant's performance report

Say a consultant sends you a monthly report that leads with return on ad spend. ROAS alone tells you almost nothing about whether the spend grew your business or just harvested demand that was already there. Here's what a fuller report looks like, using one real portfolio as the example: across 30 advertisers in July 2026, one group's managed book ran 78.4 million impressions at a $4.00 CPM and a blended $1.42 cost-per-click, producing a 6.04x return on ad spend measured across the whole book — not the best-performing line item pulled out to look good in a deck.

Underneath that ROAS, the same book converted 57,137 attributed purchases at a blended $5.49 cost per acquisition, and 20.1% of those purchases came from a shopper new to the brand. That last figure is the one most reports leave out entirely, and it's the one that tells you whether the media is finding new customers or just closing out people who were going to buy anyway.

When you get a report from your own consultant, ask for these four numbers together — impressions and CPM, blended CPC (not the cherry-picked cheapest format), blended CPA across the whole account, and new-to-brand share. A single ROAS number without the other three is a headline, not an audit.

When the numbers don't add up

The most common way a report looks better than the business actually performed is double-counting. If a shopper sees a display ad and then clicks a sponsored ad before buying, last-click attribution gives the whole sale to the sponsored ad, and a separate DSP report may claim the same sale too. Add both reports together and you've invented a sale that only happened once. This isn't a rare rookie error — it's the default outcome of running two ad products through two separate dashboards without reconciling them, and plenty of experienced teams, including ones that should know better, have reported a combined number that was quietly double-counted.

The fix isn't a better dashboard. It's reconciling DSP and sponsored ads in a shared measurement layer so a sale only gets counted once, and testing incrementality directly with holdout groups or matched controls rather than trusting last-click alone to say what caused a purchase. Last-click can tell you what happened right before a sale. It cannot tell you whether the sale would have happened anyway. If a consultant's report can't answer "would this have converted without the ad," ask how they're checking — and if the answer is "we don't," that's useful to know before you renew.

The common mistakes brands make hiring one

Three show up over and over. First, hiring a media-buying specialist to solve an operational problem — an out-of-stock or suppressed listing needs compliance and catalog work, not a smarter bid strategy, and no amount of ad optimization fixes a listing that isn't live. Second, judging a pitch on a screenshot of one client's best month instead of asking for whole-portfolio numbers across a defined period — three case studies out of a hundred accounts tell you what's possible, not what's typical. Third, not asking how new-to-brand purchases are measured at all, which means you can't tell if six months of ad spend actually grew your customer base or recycled the same buyers at a higher cost.

A quieter mistake, on the agency side: reporting DSP and sponsored ads as if they're independent channels when a shopper often touches both before buying. That inflates the combined ROAS and makes display look more effective than it was. It's an easy number to get wrong, and worth asking any consultant — including us — how they avoid it.