How to Audit Your Amazon Ads Account: The 7-Stage Method

Amazon is its own world. Most product searches in the United States now start on Amazon, not Google. eMarketer reports that 57% of US shoppers start their product searches on Amazon, based on Jungle Scout survey data. That one fact changes what an audit should do.

On Meta and Google, the ad reaches out and finds demand. On Amazon, the demand is already standing in the store. The ad just buys you a better spot on the shelf. So a real Amazon audit does not start with bids. It starts with the shelf itself: the listing, the reviews, the price. If those are broken, no bid change will save the account.

This post is the full order I run when an Amazon account lands on my desk. Seven stages. The check at each step. The fix order. The patterns I see most often. You can run it yourself in about 4 to 5 hours.

One honest note up front. The core of BTB Audits is Meta and Google, where the $150M+ spend marker comes from. The Amazon method below comes from the same diagnostic-first worldview, applied to direct-to-consumer (DTC) brands selling on Amazon. The structure is what carries the value, not a claim of equal spend depth on Amazon.

Why most Amazon ads audits fix the wrong layer

Search "amazon ads audit" and the first page is full of the same advice. Check your match types. Clean up your negative keywords. Lower your ACOS. Most of these come from one of two places.

The first is the agency sales page. The free audit is the bait. The retainer is the product. The findings get shaped to create just enough worry to close the deal.

The second is the SaaS tool blog. SaaS means Software as a Service. The post pitches the tool as the audit. Run the scan, buy the seat, problem solved. Except a tool reads the ad layer. It does not read the listing, the reviews, or the price.

Both skip the thing that decides everything: the order.

On Amazon, ad performance rides on top of listing health. If your product page converts at 4% and the category average is 8%, your ads are fighting a losing battle. You can lower bids all day. The page is still the leak.

So the method below runs in a strict order. The first two stages are the foundation. The next four are the structure. The last one is the delivery layer. Fix the foundation first. Then optimize the delivery.

For the broader case on why most audits work this way, see the BTB audit worldview. The same logic runs through the Meta ads audit method and the Google ads audit method. The order changes by platform. The habit of fixing the cause, not the symptom, does not.

Stage 1: Listing health and product-page conversion

Before I judge a single ad, I check the page the ad sends people to.

This is the gate. The detail page is where the sale happens. If the page cannot turn a click into a buyer, more clicks just cost more money. The ad is doing the job the listing should do.

What I check:

• Is the main image clean, with a white background and the product centered?
• Are 6 to 7 images uploaded, including lifestyle, infographic, and close-up shots?
• Is the title built in a clear order (brand, then product, then key feature, then size or color)?
• Are there 5 bullet points, each benefit-led, with the strongest benefit first?
• Is A+ Content live? A+ Content is Amazon's richer product description with brand story and comparison tables.
• Is the price in line with the top 3 organic results for the main keyword?

The fix order is simple. Main image first. Then title. Then bullets. Then A+ Content. Then a price review. The main image alone often moves click-through rate (CTR) by 20% to 40%. CTR is the share of people who click after they see the ad.

The pattern I see most often: a brand runs a clever ad strategy on a detail page that has not been touched in 18 months. The listing is stale. The ads paper over it. The fix is the page, not the bid.

Stage 2: Review velocity and rating health

Next I look at social proof. Reviews are the trust layer. On Amazon, they decide whether a click becomes a sale.

Review velocity means how fast new reviews come in. A product with strong velocity keeps its trust fresh. A product with stale reviews slowly loses it.

What I check:

• Does the product have 50+ reviews in a normal category, or 100+ in a crowded one?
• Is the average rating 4.3 or higher? Below 4.0, the conversion rate drops fast.
• Is review velocity 5+ per month on steady products, or 15+ per month on new launches?
• Are the 1-star and 2-star reviews being handled through Amazon Vine or follow-up email?

The fix order: deal with the 1-star clusters first, since they often point to a real product flaw. Then turn on post-purchase email so happy buyers leave reviews. Then read the worst 20% of reviews to spot patterns.

The pattern I see most often: a brand spends $20K+ a month on ads but sits under 80 reviews. They are using ad budget to make up for missing trust. The fix is review velocity, not more spend.

Stage 3: Campaign structure

Now I open the campaigns. Amazon has three ad types: Sponsored Products, Sponsored Brands, and Sponsored Display. A clean structure is what makes every later stage readable.

What I check:

• Are campaigns split by match type (exact, phrase, broad, and auto)?
• Are auto campaigns set aside as harvesters that mine new keywords, not as main performers?
• Does each top keyword sit in its own ad group so you can control its bid?
• Are Sponsored Brands campaigns kept separate from Sponsored Products?
• Are Sponsored Display campaigns split by audience (remarketing, product targeting by ASIN, and view-based)? ASIN means Amazon Standard Identification Number, the unique code for each product.

The fix order: clean up the messy campaigns first. Then split auto away from your exact-match performers. Then move your top 10 keywords by spend into their own ad groups.

The pattern I see most often: 60% to 70% of accounts at $20K+ a month run a dirty structure. Auto and exact fight over the same keyword. Broad match bleeds into brand traffic. The reports cannot tell you which keyword actually drives the sale.

Stage 4: ACOS, TACOS, and the metric trap

This stage checks whether the operator is even watching the right number.

Three metrics get mixed up all the time. Let me split them cleanly. ACOS means Advertising Cost of Sales. It is ad spend divided by ad-driven sales. TACOS means Total Advertising Cost of Sales. It is ad spend divided by total sales, which includes organic. ROAS means return on ad spend. It is just ACOS flipped over.

What I check:

• Does the operator know the difference between ACOS and TACOS?
• Is the target ACOS built from real break-even margin math, not a generic 25% to 30% rule?
• Is TACOS tracked every month to see if ads support or eat organic sales?
• Are branded keywords broken out on their own? Brand ACOS is a different animal from non-brand ACOS.

The fix order: work out the break-even ACOS from real margin. Then measure current TACOS. Then check whether brand campaigns are flattering the blended number. The break-even ROAS calculator does the margin math in about 30 seconds.

The pattern I see most often: operators chase a lower ACOS as the goal. They hit 15% ACOS while TACOS climbs to 30% because organic sales fell off a cliff. The right target is a healthy TACOS, not the lowest possible ACOS. For the full breakdown, see what is ACOS and the honest math behind Amazon ad profitability.

This stage also explains the scale of the whole game. In its Q4 2025 results, Amazon reported advertising growing 22% year over year. That is a lot of sellers paying to be seen. Most of them are steering by the wrong metric.

Stage 5: Keyword harvesting and negative keywords

This stage checks whether the account learns. Auto campaigns are a discovery engine. They show you which real searches convert. A good operator harvests those wins and cuts the waste.

Harvesting means moving a search term that converted in an auto campaign into its own exact-match keyword, where you control the bid. A negative keyword is a term you block so your ad stops showing for it.

What I check:

• Are auto campaigns running and reviewed every month for converting search terms?
• Is there a clear process to promote winning terms from auto into exact match?
• Are negative keywords being added to auto campaigns to stop waste?
• Is the negative keyword list at 100+ entries for an account at $20K+ a month?

The fix order: set a monthly harvesting habit. Build the negative list from the search term report. Promote the top 5 converting terms each week into exact match.

The pattern I see most often: brands that never harvest waste 20% to 30% of auto spend on terms that will never convert. At the same time, they miss the strong terms that should be moved to exact match for tighter bid control. This mirrors search term hygiene in the Google ads audit method. Same leak, different store.

Stage 6: Sponsored Brands and brand defense

This stage asks one question. When someone searches your brand name, who shows up?

If competitors are bidding on your brand keyword, they sit on your own search page. They catch buyers who were looking for you. Brand defense is the campaign that keeps that spot yours.

What I check:

• Does the brand run a Sponsored Brands campaign on its own brand keyword?
• Is brand defense impression share above 90% on branded searches?
• Are competitors bidding on your branded keywords?
• Is the brand running Sponsored Display with ASIN targeting to win back traffic from competitor pages?

The fix order: launch the brand defense Sponsored Brands campaign first. Then measure impression share. Then add Sponsored Display ASIN targeting on the top competitors.

The pattern I see most often: brands leave their own name undefended and lose 15% to 30% of brand-search sales to competitors. This is the Amazon twin of the Self-Defense Campaign methodology on Google. Defending your brand search is not optional once you pass $10K a month in spend.

Stage 7: Bid optimization and budget pacing

Bids come last for a reason. A bid change on a broken foundation just spends money faster. Once Stages 1 through 6 are clean, the bid layer finally pays off.

Budget pacing means how the daily budget burns through the day. If a campaign hits its cap by 2 PM, you lose every evening shopper.

What I check:

• Are auto-campaign bids kept low, in harvesting mode, not scaling mode?
• Are exact-match performers bid high enough to hold top-of-page placement?
• Is the budget pacing through the full day, not capping out by midday?
• Does the dayparting match when the category actually buys? Dayparting means setting bids by time of day.

The fix order: find campaigns that hit the cap by 2 PM, since they lose evening traffic. Bid up your exact-match winners to hold placement. Bid down auto campaigns to a harvesting level.

The pattern I see most often: brands run one flat bid across every campaign, no matter its job. Auto campaigns get scaled like performers and waste spend on noise. Exact-match winners get under-bid and lose their spot. The fix is to bid by the campaign's role, not by one account-wide rule.

Why Amazon ads are a visibility tax, not advertising

Here is the position the rest of the industry will not take.

Amazon ads are not really "advertising" in the Meta and Google sense. On Meta, the ad creates demand. It shows a product to someone who was not looking for it. On Google, the ad captures intent that lives out in the open web. On Amazon, the demand is already inside the store. The shopper typed the search. They are holding their wallet.

So what does the ad buy? Placement. A better spot on a shelf where most of the category traffic flows through one search box. You are not creating demand. You are paying a visibility tax to stand in front of demand that already exists.

This reframe is not a clever line. It decides what you audit. The Meta audit method leads with creative, audience, and attribution. The Google audit method leads with match types, bidding, and brand defense. The Amazon audit method has to lead with listing health, review velocity, and product-page conversion. If those are weak, no bid work will fix the account. That is why Stages 1 and 2 come before everything else.

It also reframes the work itself. On Amazon, the ad is maybe 30% of the job. The listing, the reviews, the price, and the operations underneath are the other 70%. The brands that win in 2026 treat it that way. The agencies that sell "Amazon ad management" as a stand-alone service are often selling bid tweaks dressed up as strategy.

There is a deeper reason the line between paid and organic blurs. Amazon is a closed marketplace. Paid sales lift your organic rank, and organic rank lowers your ad cost. The two feed each other. So the question "are my ads profitable?" quietly turns into "is my Amazon business profitable?" You cannot answer one without the other.

This is the same honest worldview behind the Meta ads audit method and the wider BTB audit worldview. The job is always the same. Find the real cause, name it plainly, and fix it in order. The patterns repeat across the Amazon accounts I have audited. The 7-stage method is the tool that surfaces them.

A real account, run through the 7 stages

Here is what the method looks like on a real account. This is a composite, blended from real audits to protect the brands.

Picture a DTC supplements brand. They spend $30K a month on Amazon ads. They run a 60% gross margin. The account is 18 months old and runs Sponsored Products only. Running the 7 stages surfaced 5 clear findings.

Add it up. Roughly $4,200 a month of ad spend was recoverable, mostly from Stages 3 and 5. On top of that, the listing fixes from Stage 1 pointed to an 18% to 25% lift in conversion rate. Notice the order. The biggest single lever was the page, not the bids. The bids were never the problem.

That is the whole point of the method. Run the stages in order. The leaks show up where the order says they will.