A £10,000 monthly Amazon ad budget can look productive on the surface, generate plenty of attributed sales and still leave very little profit behind. If you are asking, why is my Amazon ROAS low, the answer is rarely simply that bids are too high. Low ROAS is usually a symptom of a wider problem across targeting, conversion, retail readiness, attribution or the commercial target being used to judge performance.

The fix is not to cut every expensive keyword. It is to identify where spend stops producing incremental, profitable sales, then rebuild control around the account.

First, make sure low ROAS is actually the problem

ROAS is ad-attributed sales divided by ad spend. Spend £1,000 and generate £4,000 in attributed sales, and your ROAS is 4.0. It is useful, but it is not a profit metric.

A 4.0 ROAS can be excellent for a high-margin repeat-purchase product and unacceptable for a bulky, low-margin item with high fulfilment fees. Before changing campaigns, calculate your break-even ACoS and its ROAS equivalent for each meaningful product group. Include Amazon fees, landed product cost, promotions, VAT treatment, returns and any contribution you need left after advertising.

This matters because one blended target across the whole catalogue creates bad decisions. A hero SKU may justify aggressive acquisition spend. A low-margin variation may need strict protection. The commercial target should lead the media strategy, not the other way round.

Why is my Amazon ROAS low? Start with traffic quality

The most common cause is paying for searches that are relevant enough to win clicks but not specific enough to win orders. Amazon will happily spend against broad search terms unless the account architecture gives it clear boundaries.

Broad match is doing discovery and harvesting in one place

Broad match can be valuable for uncovering demand, especially when launching or expanding a catalogue. It becomes inefficient when the same campaign is expected to discover new terms, defend profitable terms and hit a hard ROAS target.

Separate discovery from control. Use controlled research campaigns with defined budgets and bid limits, then move proven customer search terms into exact-match campaigns where bids, placement multipliers and budget can be managed deliberately. Add negatives promptly so the discovery layer does not keep buying traffic that has already failed the test.

Do not judge a search term after two clicks. Equally, do not let a term accumulate 25 or 30 clicks without a sale simply because it sounds relevant. The right threshold depends on conversion rate, price point and margin, but it must be set before optimisation begins.

Product targeting is too loose

Sponsored Products product targeting often leaks spend through broad category targets, weak refinements and ASIN targets selected because they are competitors rather than because they offer a realistic conversion opportunity.

A competitor with a lower price, stronger review count and Prime advantage is not automatically a good target. You may gain impressions, but you are buying a difficult comparison. Prioritise ASINs where your product has a credible reason to win: a better pack size, stronger proposition, price advantage, distinctive feature or improved review position.

Category targeting needs the same discipline. Use refinements around price, ratings, brand and Prime eligibility where appropriate. Then review performance at target level, not just campaign level. A category campaign can conceal both your best opportunities and the waste draining its ROAS.

Brand and non-brand spend are being judged together

Branded searches normally convert more efficiently because shoppers already know the brand. Non-brand activity is where Amazon growth is won, but it usually carries more risk and requires better retail execution.

If brand defence and generic acquisition sit in the same reporting line, the account can look healthier than it is. Split them. Measure branded efficiency separately, protect your own terms where competitor pressure warrants it, and assess non-brand campaigns against a realistic acquisition target.

Your listing may be the real reason ROAS is low

Advertising gets shoppers to the detail page. It cannot make an uncompetitive listing convert. When click-through rate is acceptable but conversion rate is weak, bidding is not the first lever to pull.

Check the retail basics before increasing spend: stock availability, Prime eligibility, Buy Box status, main image quality, price position, review volume and rating, variation structure, title clarity, supporting images and A+ content. For many brands, the issue is not that the advert is being shown to the wrong customer. It is that the customer arrives and sees a weaker offer than the alternatives.

Price is especially easy to mishandle. A discount may lift conversion enough to improve ROAS, even though margin per unit falls. Conversely, holding price can be the right decision where profit protection matters more than share growth. Model both outcomes rather than treating lower price as an automatic answer.

Stockouts also distort decision-making. If a best-selling variation runs out, ads may continue to direct traffic towards less compelling alternatives or the campaign may lose its conversion history. Advertising, inventory and catalogue management need to operate as one commercial system.

Bids and placements are misaligned with intent

Low ROAS is often blamed on bids, and bids do matter. But a blanket 20% reduction across the account is usually a blunt response that reduces sales before it solves waste.

Analyse performance by search term, target, match type, campaign and placement. Top of Search can justify a premium when conversion rate is materially stronger and the product is competitive in the results. It is wasteful when the placement multiplier pushes cost per click beyond what the product can recover.

The same applies to dynamic bidding. “Up and down” bidding can work well for tightly controlled exact-match campaigns with established conversion history. Used across loose discovery campaigns, it can accelerate poor spend. Set the bidding mode according to the job of the campaign, not as a default account setting.

Budget caps create another hidden problem. When profitable campaigns run out of budget early while lower-quality campaigns continue spending, blended ROAS falls for the wrong reason. Reallocate budget towards proven profitable demand before adding more campaigns or raising total spend.

Attribution can make performance look worse – or better – than reality

Amazon ROAS reflects Amazon-attributed sales within its reporting window. It does not tell you whether the customer would have bought anyway, whether they later purchased through another route, or whether the order generated a repeat customer.

This does not make ROAS irrelevant. It means it should sit beside total sales, organic rank, branded search share, new-to-brand metrics where available, contribution margin and TACoS. TACoS – total ad spend divided by total Amazon sales – is particularly useful for judging whether advertising is supporting organic growth or simply becoming more expensive to maintain the same revenue.

Be careful with attribution changes as well. A campaign can appear to improve after reducing bids because spend falls faster than attributed sales. That may be a good outcome, or it may signal that Amazon is taking credit for sales that would have occurred organically. Look at unit volume and total sales trends before declaring a win.

A practical recovery plan for low Amazon ROAS

Start with a 30-day diagnostic rather than daily bid changes. Pull campaign, search-term, placement and advertised-SKU data, then map it against product margin and retail readiness. The aim is to establish where the account is losing efficiency, not to produce a prettier dashboard.

First, stop avoidable waste. Pause or negate search terms and targets that have spent beyond their agreed threshold without converting. Reduce bids where cost per click and conversion rate make the target impossible. Protect profitable exact terms and high-converting ASIN targets from budget starvation.

Next, rebuild campaign roles. Create distinct structures for brand defence, exact-match proven terms, phrase and broad discovery, competitor ASINs, complementary product targeting and category exploration. Each should have a clear budget, bid logic and success measure.

Then fix the product pages receiving meaningful paid traffic but converting below expectation. Improving conversion on an already active listing is often faster and more profitable than forcing more traffic into a weak page.

Finally, set a pacing routine. Review budgets and major performance shifts weekly, search-term waste and harvesting at least fortnightly, and product-level profitability monthly. Daily intervention is justified during a launch, promotion or stock issue. For established accounts, it often creates noise rather than control.

Low ROAS is not solved by chasing a universal benchmark. It is solved when every pound of spend has a defined job, every advertised SKU has a viable margin, and the account is built to distinguish profitable demand from expensive activity. That is the point where Amazon advertising becomes a controlled growth lever rather than a cost line that needs constant explaining.