How to Audit an Amazon PPC Account, Step by Step

A PPC audit is not a hunt for the one keyword that is wasting money. Most accounts leak through their structure, and the leaks are rarely where people look first. This audit follows the order in which problems cost money: the page that receives the clicks, how campaigns are built and funded, whether they compete with each other and with your organic listing, what shoppers actually typed, and how bids and placements are managed. The order matters because an upstream problem, such as a weak listing, makes every number downstream misleading.

Why the order of the audit matters

If you start with bids, you will tune a campaign around a product that is out of stock, priced above its rivals or carrying a poor rating. You will lower bids to hide the symptom and learn nothing. The sequence below puts the cheapest, highest-leverage checks first and uses what you learn to interpret the rest.

Before the first pass, do two pieces of arithmetic per product. Calculate the break-even ACoS (profit before ads divided by price) and note the conversion rate. They are the yardsticks for every later step.

Pass one: the listing

For each product that carries meaningful spend, confirm the basics. It is in stock and holds the Featured Offer. The price sits in a sensible range against the closest alternatives. The main image and first bullets answer the question a shopper arrives with. The rating and the themes in recent reviews do not undermine the sale. Delivery promise and mobile display are fine.

Anything you cannot fix this week should lower the spend on that product until you can. Paying for clicks to a page that will not convert is the most expensive kind of waste, and no bid change cures it.

Pass two: structure, names and budgets

Open Campaign Manager and read campaign names as a stranger would. Can you tell the product, the job (discover, defend or brand), the targeting type and the match type? If not, nobody can analyze results, including you in six months. A simple convention such as Product | Job | Targeting | Match is enough.

Then look inside the ad groups. A group should hold products and targets that share a bid logic. Warning signs include unrelated products sharing one group, so the bid fits none of them, a single group stuffed with keywords of mixed intent, and auto campaigns left on their defaults for months with no review.

Budgets come last in this pass. List each campaign with spend, ACoS against the product's break-even and its budget status. You are looking for two patterns. One is a profitable campaign that runs out of budget early in the day, which means it turns away shoppers it could convert. The other is a weak campaign that never hits its cap and quietly spends, where the cause is bids and targets, not budget. Budget should follow margin: a product with a fatter margin can carry a higher ACoS than a thin one.

Pass three: overlap with other campaigns and with organic

Cannibalization comes in two forms. The internal form is the same search term targeted by several campaigns or match types. The campaign that serves for a given search tends to be the one with the strongest effective bid, which is not necessarily the one you meant to use, so each campaign's report is distorted. To find it, filter the search term report across campaigns for one term and count how many campaigns it appears in. Give each term one home and add negatives elsewhere.

The organic form is subtler. When your product already ranks at the top for a term, such as your own brand name, some of the clicks you pay for would have arrived free. Amazon does not report how many, so test it. Lower the bid on that term for a set period and watch total sales for the product (ads plus organic), not ad sales alone.

Do not read this as a case for dropping branded campaigns. If competitors bid on your brand name, a branded campaign protects the result shoppers see first. The audit question is whether you are paying more than the protection is worth.

Pass four: search terms

Pull the search term report over a window long enough to hold meaningful data. Sort the terms with a rule built on money, not click counts. First work out what one order is worth: price times break-even ACoS, which equals the profit before ads on a unit. At a $30 price and a 30% break-even ACoS, that is $9.

Terms with a lot of spend and no sale are the obvious wins. Do not stop there: also check the terms that convert, because they are the raw material for tighter campaigns.

When you negate, match the match type to the problem. A negative exact blocks only that precise search. A negative phrase blocks any search that contains the words in order, which suits a whole irrelevant theme, such as a different product type that shoppers confuse with yours. Record every negative with the date and the reason, so a future audit can tell a deliberate block from an accident.

Pass five: bids, placements and cadence

Look at the change history. Daily bid changes on thin data mean the account is reacting to noise. No changes in months mean it is ignoring information. Neither is healthy. Then open the placement report and compare conversion rate and ACoS across top of search, rest of search and product pages.

Check the dynamic bidding setting on each campaign as well, and make sure it was chosen on purpose. A setting that lets Amazon raise bids when a click looks likely to convert behaves very differently from one that only lowers them, so the same base bid can mean quite different spend. If you cannot say why a campaign uses its current setting, treat that as a finding.

Finally, check what your bid modifiers add up to. Placement modifiers and dynamic bidding strategies raise the bid you actually pay above the base bid you typed. The ceiling is your maximum affordable cost per click: price times break-even ACoS times conversion rate. At $30, 30% and a 10% conversion rate, that is $0.90. If a base bid plus its modifiers can climb past that figure on a placement that converts poorly, the account is losing money by design.

Where to start

• Calculate break-even ACoS and conversion rate for each product that carries spend.
• Check stock, Featured Offer, price and rating before opening a campaign.
• Rename campaigns so product, job and targeting are visible, and regroup ad groups.
• Flag campaigns that run out of budget early and campaigns that spend freely above break-even.
• Find terms that appear in several campaigns and give each one a single home.
• Run the search term triage, then check effective bids against your maximum cost per click.