Amazon PPC strategy: structure, bids and budgets built around profit

What Amazon PPC is

Amazon PPC (pay-per-click) is advertising on Amazon where you bid for ad placements and pay only when a shopper clicks. The core format, Sponsored Products, promotes individual listings at the top of, alongside or within shopping results and on product pages; there are no monthly or upfront fees, you set a daily budget, and you bid the most you are willing to pay for a click¹. The more competitive the bid, the more likely the ad is to show when it matches a shopper’s search¹. To advertise, a product must be in an eligible category and be eligible for the Featured Offer¹, which our guide to the Amazon Buy Box explains.

An Amazon PPC strategy is a set of decisions about which products to advertise, to whom, at what price per click and with how much money. Most guides judge those decisions on ACoS, the share of ad-attributed sales spent on ads¹³. We judge them on a simpler question: after paying for the ads, did the business keep more money? This guide shows how to build each part of an account around that question.

Start from each product’s break-even

Every product has a break-even ACoS: its margin before advertising, as a share of the selling price excluding VAT, because that is how Amazon Ads reports attributed sales. A product that keeps 28p of every £1 of ex-VAT sales after cost of goods, Amazon’s fees and fulfilment can spend up to 28% of its ad-attributed sales on ads before each sale starts losing money. Our break-even ACoS guide shows the full calculation.

Two consequences shape everything that follows:

• One target ACoS for the whole account is wrong for most products. It is too tight for the high-margin ones and too loose for the thin ones.
• Break-even also sets the most you can pay per click. Because ACoS is cost per click ÷ (conversion rate × order value), the highest click cost a product can carry is its order value × conversion rate × break-even ACoS. A £30 product that converts 12% of clicks with a 25% break-even can afford £0.90 a click, and no more.

Keep that click ceiling in mind when you choose a bidding strategy, because some settings can double a bid.

Campaign architecture

Amazon now groups its ads into sponsored ads, Display, Video and Audio, and Full-Funnel Campaigns, all inside a platform it renamed Amazon Ads Agent on 29 September 2026³. For most brands selling on Amazon.co.uk, the decisions are still these:

Amazon says DVA+, which combines Sponsored Display, Sponsored TV and programmatic campaigns, starts rolling out from late October 2026, while Full-Funnel Campaigns is available only in the United States so far³. Existing DSP campaigns keep running without any action on your part⁵. Check your own console for UK timing.

Inside Sponsored Products, we structure campaigns so that each has one job and one budget: discovery (automatic and broad match), proven terms (exact match), competitor and category targeting (product targeting), and brand defence (your own brand terms). The reason is control. When one campaign mixes a hero product with a slow seller, or two variations that convert very differently, the budget tends to flow to whichever gets the most impressions rather than whichever makes the most money. In one audit we found the better-converting colour of a hero product receiving a fraction of the spend because both colours shared a campaign. Splitting them by variation was the fix.

Match types and negative keywords

Amazon offers three keyword match types⁷:

• Broad: the search can contain your words in any order, plus synonyms and related terms. The keyword itself may not appear in the search at all.
• Phrase: the search contains your phrase, or close variations such as plurals and misspellings.
• Exact: the search matches your keyword or a close variation. The most restrictive, and usually the most relevant.

Automatic campaigns add four targeting groups: close match, loose match, substitutes (product pages similar to yours) and complements (product pages for items bought alongside yours)⁷.

The loop we run is simple. Discovery campaigns find searches that convert, alongside the sources in our guide to Amazon keyword research; those searches move into exact-match campaigns with their own bids; and the discovery campaign gets a negative for each one moved, so the two do not compete. Searches that spend without converting become negatives too. Amazon allows negative phrase (up to 4 words) and negative exact (up to 10 words), each up to 80 characters, plus negative products to stop ads on specific product pages⁸. Amazon says new negative keywords can take 72 hours to take effect, and negative products 96 hours⁸, so do not judge a change the next morning.

Bidding strategies and placements

Sponsored Products offers four ways to bid⁹:

• Dynamic bids, down only. Amazon may lower your bid, by up to 100%, when a click looks less likely to convert.
• Dynamic bids, up and down. Amazon may raise or lower your bid by up to 100% in all placements. Amazon’s own example: a $1.00 bid can become $2.00.
• Fixed bids. Amazon uses your exact bid every time.
• Rule-based bidding. For eligible campaigns, you set a return-on-ad-spend guardrail and Amazon adjusts base bids to increase conversions within it.

Placement adjustments sit on top of the strategy. You can raise bids for top of search (the first row of the first results page), rest of search, product pages and Amazon Business placements, by up to 900%¹⁰. The effects stack: a £1.00 fixed bid with a 50% top-of-search adjustment becomes £1.50 there, and with dynamic bids up and down it can reach £3.00¹⁰.

That is why the click ceiling matters. On the £30 product above, a £0.60 bid is safe on fixed bidding. Under up and down with a 50% placement adjustment, the same bid can reach £1.80 on top of search, twice what the product can afford at its usual conversion rate. Amazon raises bids for clicks it expects to convert better, so the maths is not always that bleak, but check the placement report before you trust it. Our practice is down only for new and unproven campaigns, up and down for proven exact-match campaigns with margin to spare, and fixed bids for tests.

Split budget by margin, not by habit

The most common problem we find in large accounts is not a bad bid. It is budget spread across everything, with no link to which products make money. We saw this in one of our biggest turnarounds, a brand we work with that was advertising almost its whole catalogue without priorities. The fix was not complicated: focus budget on the best products, rebuild the structure for control, and combine automated bidding with human review. We cannot publish that client’s figures, so here is the same pattern with illustrative numbers.

The hero products get more than twice the budget and their ACoS rises, as it should when you push into less efficient searches, but it stays well below their 35% break-even. The core range loses the searches that were costing more than its 25% margin. The long tail keeps only the exact-match terms that have proved themselves. Spend falls by £5,000, ad-attributed sales rise from £133,333 to £153,333, and profit from advertising moves from a £4,000 monthly loss to an £11,100 gain: £15,100 a month.

Two groups need their own budget outside this logic: launches, which may run above break-even for a set period, and brand defence, where the cost of losing your own search results can exceed the ACoS. Give both a budget and an end date or a review date.

The weekly and monthly rhythm

This is the routine we run. It is our practice, not an Amazon rule.

Every week

1. Review the search term report: move converting searches into exact match, add negatives for searches that spend without converting.
2. Check which profitable campaigns ran out of budget early, and which unprofitable ones did not.
3. Adjust bids only on targets with enough clicks to judge.
4. Check stock. Pause or reduce ads on products close to running out.

Every month

1. Work out ACoS against break-even for every product group, and TACoS (ad spend ÷ total sales) for the account and your biggest products. Our TACoS guide shows how to set a ceiling.
2. Review placement performance and bidding strategies against the click ceiling.
3. Download and keep the search term report. It only reaches back 65 days¹², so without your own copies you lose the history.
4. Recalculate break-even for any product whose price, cost or Amazon fees changed.

Signs an account is drifting: the same campaigns and keywords as three months ago, no new negatives in a month, and TACoS rising with no planned reason.

Measure what advertising adds

Amazon’s reports tell you what it credits to ads, which is not the same as what ads add.

• Attribution windows. For sellers, a Sponsored Products sale counts only if the shopper buys within 7 days of clicking, on a last-click basis; for vendors the window is 14 days, including other products from the same brand¹¹. Brand ads such as Sponsored Brands, display and DSP compete with each other for credit, generally over 14 days, and some can also claim sales after an ad was only seen¹¹.
• Recent data is incomplete. Conversions can take up to 12 hours to appear, and figures for a date stay incomplete until the attribution window has closed¹¹. Do not cut a campaign on last week’s ACoS.
• VAT. Amazon says attributed sales exclude VAT¹¹. Compare ACoS with a margin worked out on the ex-VAT price, and check one attributed order against what the customer paid.
• Total sales. Take Ordered Product Sales from your Business Reports¹⁴, divide ad spend by it, and watch the trend. Profit after advertising is total sales × (margin before advertising − TACoS).

For a bigger test, pause one group of campaigns for two weeks and watch total sales, not ad sales. If total sales barely move, the ads were mostly buying sales you would have made anyway.

The strongest PPC accounts we see are not the ones with the cleverest settings. They are the ones where every pound of budget can be traced to a product that makes money, and where someone checks every week that it still does.