A Performance Max recovery example is only useful if it deals with the problem that matters: profit. A campaign can report an attractive return on ad spend while spending too heavily on low-margin products, discount-led sales or customers you would have acquired anyway. Recovery is not about making Google Ads look tidier. It is about restoring controlled, commercially viable growth.
For established eCommerce brands, Performance Max usually breaks down for predictable reasons. Product feeds become unfocused, conversion data becomes unreliable, campaign structures lose intent, and budget keeps flowing because the platform can still find revenue somewhere. The result is often more sales in the account and less confidence in the bank.
Performance Max Recovery Example: The Starting Position
Consider a representative online retailer spending £20,000 a month on Performance Max. The account reports £70,000 in revenue at a 3.5x ROAS. On the surface, that is not disastrous. But the business needs at least a 4.0x blended paid-media ROAS after accounting for product margin, fulfilment, returns and agency cost.
The founder sees another issue: best-selling, full-price products are frequently outperformed in the feed by clearance lines and low-value accessories. Average order value from Performance Max is falling. Brand search is also carrying a material share of the reported conversion value, making the campaign look stronger than its true incremental contribution.
This is a recovery job, not a reason to switch the campaign off blindly. Performance Max has access to Shopping inventory, remarketing signals, YouTube, Display, Discover and Search. Removing it overnight can create a revenue hole. Keeping it unchanged, however, means accepting unprofitable scale.
The objective is clear: move the account from £70,000 at 3.5x ROAS to a healthier contribution from paid media, while protecting revenue from products that genuinely deserve budget.
Start With Commercial Truth, Not the Google Ads Interface
Before changing a bid, establish what a conversion is worth. This means confirming the correct revenue figures, excluding VAT where appropriate, checking return rates, and identifying product-level gross margin. A £150 order is not automatically better than a £90 order if the larger order contains heavily discounted stock or high-return items.
In this example, the audit finds three material issues. First, purchase values are being passed correctly, but refunds are not reflected in the decision-making process. Secondly, the feed includes hundreds of products with poor availability, thin margins and weak imagery. Thirdly, the campaign has been optimising against a single broad purchase goal, with no guardrails around product profitability.
This is where many accounts go wrong. They treat target ROAS as a profit setting. It is not. Target ROAS is an efficiency instruction based on the conversion value Google receives. If the value sent to the platform does not reflect commercial value, the bidding strategy will efficiently pursue the wrong sales.
Check whether demand has changed
A recovery plan must separate campaign failure from business failure. If stock availability has dropped, prices are no longer competitive, delivery is slow, or the site conversion rate has deteriorated, Performance Max is not the only issue.
Compare the period before performance declined with the current one. Look at stock coverage, merchant approval rates, selling price, conversion rate, average order value and returning customer behaviour. Also check whether branded demand has risen or fallen. A weak campaign may be the symptom of a wider trading problem.
Rebuild the Feed Around What Can Scale
For most retailers, the product feed is the foundation of Performance Max recovery. Generic titles, incomplete attributes and one-size-fits-all product groupings force Google to make broad assumptions. That is not a strategy.
In this example, products are split according to margin, stock depth, price point, seasonality and historical demand. The retailer creates a priority group for proven, well-stocked products with acceptable margin. Clearance items move into a separate campaign with a tighter budget. Low-margin accessories are excluded from the core acquisition activity unless they are part of a profitable basket strategy.
Titles are rewritten to include the terms customers actually use, alongside critical attributes such as brand, product type, size, colour, material and compatibility where relevant. Images, GTINs and custom labels are reviewed rather than assumed to be adequate. This work sounds operational because it is. Yet it often has more commercial impact than another round of asset copy testing.
The trade-off is volume. Excluding products can reduce reported revenue initially. That is acceptable if those products were inflating sales while weakening contribution. A smaller, more profitable product set gives the algorithm a clearer route to finding customers worth acquiring.
Fix Campaign Structure Without Creating Chaos
Performance Max works best when the structure reflects a meaningful business decision. Splitting campaigns simply because a catalogue has several categories often fragments data and slows learning. Equally, one campaign covering every product, margin profile and promotional objective is rarely controllable enough for a serious retailer.
For this recovery, the account is restructured into two primary campaigns. The first focuses on core, profitable products with room to scale. The second contains promotional and clearance inventory with a fixed budget ceiling. Brand exclusions and account-level controls are assessed carefully so the core campaign is not given undue credit for demand generated elsewhere.
Customer acquisition settings are also reviewed. If the brand needs new customers, bidding solely for any purchase may cause Performance Max to lean heavily into previous buyers. New-customer value can help, but only when customer lists are accurate and the additional value is realistic. Setting an exaggerated new-customer value simply distorts reporting in a different direction.
Asset groups are then aligned with genuine product themes, not used as a cosmetic filing system. Audience signals provide useful starting points, particularly customer lists, high-value purchasers and category-specific search intent. They are signals, not targeting restrictions. Google can expand beyond them, which is why feed quality and conversion integrity remain decisive.
Reset Targets Gradually and Measure the Right Things
The original 3.5x ROAS was below the retailer’s viable level, but immediately setting a 5.0x target would risk restricting traffic too aggressively. The campaign needs enough conversion volume to relearn after the feed and product selection changes.
The target is raised in stages, with budget held steady during the first learning period. Performance is reviewed against contribution indicators rather than daily fluctuations. Daily ROAS is noisy, especially for higher-consideration products or brands with varied order values. A seven-day or 14-day view is more useful, provided it is considered alongside conversion lag.
After six weeks, spend falls slightly to £18,500, but attributable revenue settles at £78,000. Reported ROAS improves to 4.2x. More importantly, the sales mix shifts towards full-price, higher-margin products and average order value rises. The account has not been rescued by a clever setting. It has been redirected towards commercially sensible demand.
That outcome will not apply to every retailer. Some brands need to accept a lower first-purchase ROAS because repeat purchase economics justify it. Others have narrow margins and need stricter acquisition efficiency from the outset. The correct target comes from the business model, not an agency benchmark or a number copied from a competitor.
What This Performance Max Recovery Example Shows
A credible recovery does not begin with new creative or a bigger budget. It begins with deciding which sales are worth buying. Once tracking, feed quality, product economics and campaign controls point in the same direction, Performance Max has a better chance of doing what it is designed to do: find scalable demand across Google’s inventory.
If your campaign has become a black box, do not judge it by clicks, impressions or even headline ROAS alone. Ask which products it is pushing, which customers it is reaching and whether the revenue left after advertising is worth keeping. That is the standard a serious eCommerce account should meet.