A ROAS improvement case study should not be a victory lap for a headline number. For an established eCommerce brand, the real question is whether paid media is generating profitable orders that can be scaled without handing more margin to Google and Meta. A 5x ROAS can still be poor trading if it is driven by low-margin products, repeat customers you would have acquired anyway, or an attribution model that flatters the platform.
This representative case study reflects the patterns seen in mature eCommerce ad accounts: decent revenue, rising spend and an uncomfortable feeling that the advertising is working harder without improving the bottom line. The figures are illustrative, but the process is the same one a serious retailer should expect from a specialist PPC partner.
The starting point: revenue was growing, profit was not
The brand was spending approximately £28,000 a month across Google Shopping, Performance Max, Search and Meta. Platform reporting showed £84,000 in attributed revenue, producing a 3.0x blended ROAS. On the surface, that looked workable.
It was not. After product cost, packaging, payment fees, fulfilment and a realistic allowance for returns, the business needed at least a 3.6x blended ROAS to make cold acquisition commercially worthwhile. The account was therefore buying revenue below its profit threshold.
The problem was not a lack of traffic. It was a lack of control. Performance Max was absorbing budget into a broad product mix, Shopping campaigns were allowing mediocre products to spend alongside proven winners, and Meta was optimised for purchase volume without enough regard for order value. Branded search was also making the account look healthier than it was, because people already looking for the brand were being credited as incremental paid-media wins.
This is where many agencies reach for more creative, more audience tests and a larger budget. That is backwards. When the economics are wrong, scaling simply amplifies the problem.
ROAS improvement case study: finding the wasted spend
The first job was to establish the commercial rules. Not every product deserved the same bid, and not every sale deserved the same value.
The retailer had strong gross margins on a group of higher-priced bundles and accessories, but thin margins on several discounted entry products. Yet the existing campaigns treated them as equal. The low-margin lines were generating lots of conversions, which made automated bidding happy, while quietly dragging blended profitability down.
A product-level review exposed three clear issues. First, a small group of SKUs was consuming a disproportionate share of Shopping spend while producing ROAS below the brand’s breakeven target. Secondly, best-selling bundled products had poor titles and incomplete attributes in the feed, reducing their visibility for high-intent searches. Thirdly, budget was spread too evenly across products with radically different stock levels, margins and conversion rates.
The account also needed a measurement reality check. Platform ROAS is useful for steering campaigns, but it is not a profit and loss statement. Google Ads and Meta both claim credit under their own attribution rules. The management team therefore tracked platform data alongside total paid spend, ecommerce revenue, new-customer performance, average order value and contribution margin. That prevented a cosmetic ROAS improvement from being mistaken for genuine commercial progress.
The changes that moved the number
1. Set targets from margin, not platform averages
The agency and brand calculated a breakeven cost of sale by product group rather than settling for one broad account target. Higher-margin bundles could support more aggressive acquisition. Discounted products required stricter efficiency, while certain loss-leading products were retained only where there was evidence of worthwhile repeat purchase behaviour.
This matters because a single ROAS target can be lazy management. A 4x return may be excellent for one category and damaging for another. The right target depends on gross margin, fulfilment costs, returns, average order value, customer lifetime value and the brand’s cash-flow tolerance.
2. Rebuild campaign structure around commercial intent
Campaigns were reorganised so that profitable product groups could receive their own budgets, targets and search-query control. High-margin bundles and proven bestsellers were separated from lower-value products. Poor performers were either restricted, tested with a lower bid ceiling or removed from paid acquisition altogether.
Branded search was also separated from non-brand activity. This did not mean switching brand campaigns off. Defending branded demand can be sensible, particularly in competitive categories. It meant reporting it honestly, rather than allowing low-cost branded conversions to disguise weak prospecting performance.
For Performance Max, the priority was not endlessly changing settings. It was improving the inputs: product segmentation, asset relevance, audience signals and a cleaner feed. Performance Max can scale efficiently, but it is not a substitute for a clear product strategy. Give it mixed economics and it will often find the easiest conversions, not the most profitable ones.
3. Fix the feed before demanding more from Shopping
Feed optimisation was a material part of the recovery. Product titles were rewritten to better reflect how people searched, including category, key material, size or use case where relevant. Missing product identifiers and attributes were corrected, images were reviewed, and promotional messaging was aligned with landing pages.
The objective was not to stuff titles with every possible keyword. It was to make the catalogue legible to Google and persuasive to shoppers. Better feed quality improves eligibility and query matching, but it also reduces the gap between the searcher’s expectation and the product page. That can raise conversion rate without increasing the cost of a click.
4. Stop paying for weak signals on Meta
Meta activity had been optimised around purchase events, but the account was over-serving warm audiences and discount-led products. The revised structure protected remarketing without allowing it to consume acquisition budget, then tested prospecting creative against the brand’s stronger-margin ranges.
Creative was treated as a commercial input, not a branding exercise. Ads led with the product, price logic, proof and use case. Results were judged by contribution to new-customer revenue and sustainable cost of acquisition, not by cheap clicks or flattering engagement rates.
The result after eight weeks
After the account had enough time to stabilise, monthly paid-media spend fell from roughly £28,000 to £25,000. At the same time, attributed revenue increased from around £84,000 to £105,000. Blended platform ROAS rose from 3.0x to 4.2x.
The 40% ROAS increase is useful, but it is not the most valuable outcome. Revenue rose while spend reduced, and a greater proportion of that revenue came from product groups capable of carrying acquisition costs. The brand was no longer using profitable bestsellers to subsidise low-margin, low-quality volume.
There were trade-offs. Restricting poor-margin products initially reduced conversion volume in some campaigns. Separating brand reporting made prospecting look weaker than the previous headline figures suggested. And feed improvements took time to be crawled, approved and reflected in performance. None of that was a failure. It was the account becoming more truthful and more controllable.
What this case study means for your account
If your ROAS has stalled, the cause may not be bid strategy. It may be that your account is optimising against the wrong commercial objective. Automated campaigns are highly capable of finding conversions. They cannot decide which products are worth selling, what level of acquisition cost your margins support, or whether reported revenue is contributing to profitable growth.
A credible ROAS improvement plan starts with the numbers your finance team cares about: margin by product, breakeven cost of sale, returns, stock availability and cash flow. It then translates those constraints into campaign structure, feed quality, budget allocation and reporting that cannot hide wasted spend.
Oxedent works exclusively with eCommerce brands because that level of detail is not optional when serious ad budgets are at stake. Before asking how quickly you can scale, ask whether every extra pound is being directed towards the products and customers that make scaling worth it.