A Profit-First Structure for Amazon PPC: Campaigns, Bids and Budgets

Most Amazon ad accounts are built by accretion. A campaign gets added when someone has an idea, a bid goes up because a keyword feels important, and the budget is divided by whatever seemed fair on the day. After a year, nobody can say what each campaign is for. A profit-first structure replaces habit with three decisions: the job each campaign does, the most you can afford to pay for a click, and how to divide the budget among your products. A fixed review rhythm then keeps those decisions current.

Give every campaign one job

A campaign that tries to discover new shoppers, defend your pages and convert loyal buyers all at once cannot be judged fairly. Separate campaigns by job, so each one is measured against a standard that fits it.

Keep branded campaigns out of the blended numbers. Shoppers who type your brand name already want your product, so these campaigns show a low ACoS almost by default. Mixed into an account average, they make weak discovery or proven-term campaigns look healthy. Defense matters because Sponsored Products ads can appear on other brands' detail pages. Targeting your own ASINs is how you stand on that ground yourself.

Discovery feeds the other jobs. Each week, read the search term report: terms that produced orders get promoted to exact match in the proven campaign, and terms that keep spending without orders become negatives in discovery, so you do not pay twice for the same lesson. Put the job, product and match type in every campaign name so a report can be filtered in seconds.

Watch for the same term running in two places. If a keyword is promoted to the proven campaign but still sits in discovery, your own campaigns compete against each other and the report shows two partial results instead of one clear one. Adding the promoted term as a negative exact match in discovery keeps each term in a single home.

Work out the most you can pay for a click

A bid is a decision about what a click is worth. A click is worth the share of clicks that turn into orders, multiplied by the profit each order earns before ad spend. That profit-before-ads figure is the same one used to find break-even ACoS: price minus landed cost, Amazon fees and an allowance for returns and storage.

Check it with round numbers. At $1.40 a click, 100 clicks cost $140. At a 10% conversion rate those clicks bring 10 orders, and 10 orders at $14 of profit before ads each is $140. Spend and profit cancel out, which is what break-even means.

To keep some profit, subtract it first. If you want $4 per unit left after advertising, the ceiling is 10% times $10, which is $1.00. Treat this as the top of the range, not the opening bid. The figure you set is a maximum, and actual costs per click are often lower.

One more adjustment. Dynamic bidding and placement bid modifiers can raise your effective bid above the base bid you typed in. If you plan to add a 50% modifier for top-of-search placement, the base bid has to leave room for it: $1.00 divided by 1.5 is about $0.67.

Be careful with small samples. A keyword with five clicks and one order shows a 20% conversion rate, and a single click more or one order fewer changes the picture completely. Until a term has gathered plenty of clicks of its own, use the product's overall ad conversion rate for the calculation.

Split budget by profit per click, not by habit

Dividing the budget equally ignores profit. Dividing it by revenue rewards expensive products that may keep little. A better starting ranking is profit per click, the same break-even cost per click you just calculated. It shows how much profit before ads one average click earns on each product.

Take three hypothetical products. Product A: $40 price, $14 profit before ads, 10% conversion. Product B: $25 price, $5 profit before ads, 12% conversion. Product C: $60 price, $24 profit before ads, 6% conversion. Multiply profit by conversion and you get $1.40 for A, $0.60 for B and $1.44 for C.

Product B has the best conversion rate and the weakest economics, because it keeps so little per sale. Give it a smaller share of the budget, or fix its margin before pushing it. Product C converts least often yet still ranks first, because each sale is large.

Then apply two limits to the ranking. First, inventory: do not push spend on a product that will run out of stock before the next shipment lands. Second, demand: a product can only absorb as much budget as there are relevant searches to buy. If campaigns are hitting their daily budget early, which the console flags, you are missing impressions later in the day and may have room to add budget. If they are not, more money will not help.

A product with weak profit per click may still deserve a floor, for example because it opens the door to a better-margin item. Make that a conscious choice with a written reason, so it does not survive on habit. Decide the discovery share in advance. Setting aside, say, a fifth of a product's ad budget as a research cost is an illustrative choice, not a rule. What matters is that the share is deliberate and that you check what it bought.

A weekly and monthly rhythm

A structure only stays useful if someone maintains it, and the schedule should match how fast the data becomes meaningful. Daily work is about alarms. Weekly work is about terms and bids. Monthly work is about money.

At the weekly check, compare each bid with the ceiling you calculated. A term that converts well and sits far below its ceiling may be missing impressions, so it can take a higher bid. A term that sits above its ceiling needs a lower bid, a negative or a pause, and the product page or price may deserve a look if conversion is the weak link. Resist daily bid changes. A single day of clicks is too small a sample for most terms, and chasing it means reacting to noise. Make one type of change at a time, such as lowering bids on a group of terms, and leave it long enough to see what it did. Changing bids, budgets and targeting together makes it impossible to say which one worked.

Where to start

• Calculate profit before ads for each advertised product from current costs and fees.
• Work out the break-even cost per click for each one, and write the ceiling next to the product.
• Sort campaigns into discovery, proven terms, defense and branded, and rename them to say so.
• Report branded campaigns separately from everything else.
• Rank products by profit per click and reset budgets, limited by stock and by demand.
• Put the daily, weekly and monthly checks in a calendar and keep to them.