Performance Max Spending But Not Converting? 14 Fixes (UK 2026)

Last updated: 28 September 2026 · Updated as Google changes Performance Max bidding, reporting and controls

Your Performance Max campaign is spending every pound you give it, and the conversions column says zero, or close to it. In almost every UK account we audit, PMax is not failing because Google’s AI is broken. It is failing because it has been given too little signal to find buyers, so it does the only thing a budget-pacing system can do: it spends where spending is easiest. This guide explains how the Performance Max algorithm actually allocates your money, the mechanism we call the Starvation Spiral, the 2026 platform changes quietly raising costs, and a fix plan you can run today.

Short answer: A Performance Max campaign spends without converting for one of four reasons: (1) conversion tracking is broken or too thin for Smart Bidding to learn from, (2) the campaign is still in, or was knocked back into, the learning period, (3) budget and targets stop it affording high-intent Search and Shopping auctions, so spend drifts to cheap Display and YouTube inventory, or (4) the feed, landing page or offer cannot convert the traffic it sends. Check in that order: tracking, then signal volume (30+ conversions in 30 days), then the Channel performance and Search terms reports, then the offer.

How the Performance Max algorithm actually spends your budget

To fix a Performance Max campaign that is not converting, you need to stop thinking of it as one campaign. It is two separate systems running on the same money, and most “PMax not converting” problems come from the gap between them.

System one: budget pacing

The pacing system’s job is to spend your budget. Google can spend up to twice your average daily budget on a given day and caps the month at roughly 30.4 times the daily figure. Pacing does not care whether the money converts. If there is budget left and eligible inventory exists, it will find a way to spend it.

System two: Smart Bidding

Smart Bidding decides what each impression is worth. In simplified form, bid = your target CPA × the predicted conversion rate for that auction. With Target ROAS the same logic runs on predicted value. With Maximise Conversions and no target, it tries to buy as many conversions as it can while spending the full budget. Every prediction comes from one place: the conversions you feed back to Google Ads.

What sits between them

• Channels are not budget lines. Performance Max has no per-channel budget. Search, Shopping, YouTube, Display, Discover, Gmail and Maps all compete for the same pound, and the model buys whatever its predictions say is worth it.
• You get the leftovers. When a search query is identical to a keyword in one of your Search campaigns, Google prioritises the Search campaign over PMax. Since late 2024, Standard Shopping and PMax compete on Ad Rank rather than PMax automatically winning. In a well-built account, PMax is working the harder, broader traffic your other campaigns didn’t claim. Its conversion rate will naturally sit lower.
• Audience signals are hints, not targeting. Customer lists and interests tell the model where to start looking. They do not fence it in. If your “signals” convert poorly, the model will leave them.
• Your conversion goals are the definition of success. PMax optimises to whatever is marked as a primary conversion action for that campaign. If that action is broken, rare or meaningless, the most powerful bidding system in advertising will optimise, perfectly, towards the wrong thing.

The Starvation Spiral: why PMax spends on junk when it can’t find buyers

This is the pattern we see behind most “PMax spending but no conversions” accounts. It is not a Google term; it is how we describe what happens when the two systems above collide on a thin data set.

Here is the mechanism step by step. With only a handful of conversions, Smart Bidding’s predicted conversion rate is low and uncertain for every auction. Plug that into bid = target × predicted rate and high-intent Search and Shopping clicks, which cost several pounds each in competitive UK categories, look unaffordable. But pacing still has budget to spend. The cheapest eligible impressions sit on Display, YouTube, Discover and Gmail, so that is where the money goes. Those placements convert at a fraction of Search’s rate, so conversions fall further, predictions fall further, and the campaign drifts deeper into cheap inventory.

From the outside it looks like PMax has “decided” to waste money on YouTube. It hasn’t decided anything. It has been starved of the one input it needs to tell a buyer from a browser.

How to tell if you are in the spiral

• The Channel performance report shows most spend on Display, YouTube or Discover with near-zero conversions.
• Impressions are very high, click-through rate is low and average CPC is far below your Search campaigns.
• The campaign spends its full budget every day, yet sits on fewer than 30 conversions in the last 30 days.
• Your Search terms report for PMax shows little volume, or mostly brand terms.

How to break the spiral

1. Fix the signal before the settings. Verify every primary conversion action fires, add enhanced conversions, and import offline sales if you generate leads. More honest signal is the only thing that raises predicted conversion rates.
2. Raise budget relative to CPA, not in absolute terms. Our operating default: daily budget at least 3× your target CPA, ideally 5× or more. A £40 budget chasing an £80 CPA cannot produce a daily rhythm of conversions for the model to learn from.
3. Loosen or remove the target. A Target CPA set far below reality makes every Search auction “too expensive”. Start near your real CPA and tighten by 10 to 15% every two weeks once conversions flow.
4. Narrow what PMax is allowed to sell. Fewer asset groups, your best-selling products or services, strong search themes and clean final URLs concentrate thin data instead of spreading it.
5. Close the cheap exits. Account-level placement exclusions, negative keywords, brand controls and customer-list exclusions remove inventory that soaks up budget without converting.

The LEAKS audit: five places Performance Max spend escapes

When we run a Performance Max audit for UK clients, we check five areas in a fixed order. The order matters: there is no point writing negative keywords for a campaign whose conversion tag stopped firing three weeks ago.

PMax not converting: symptom, cause and fix

Match what you see in the account to the most likely cause. Each row is something we have found in live UK accounts.

Menu names reflect the Google Ads interface as of September 2026 and may move as Google updates the UI.

2026 Performance Max changes that affect whether you convert

PMax has changed more in the last 18 months than in the three years before. Several of those changes directly explain accounts that were converting in spring and aren’t now. For the full account-wide picture, see our Google Ads updates tracker.

The UK tracking gap: why your conversions may be invisible to PMax

This is the most under-discussed reason UK Performance Max campaigns look like they are not converting. Consent Mode v2 has been required for advertisers serving users in the UK and EEA since March 2024. When a visitor rejects cookies, the Google Ads tag cannot tie their purchase back to the ad click. Google fills the gap with conversion modelling, but only once you clear a threshold most small UK accounts never reach.

Google’s eligibility rule for consent-mode modelling is 700 ad clicks over 7 days, per country and domain. That is about 100 UK ad clicks a day to one website. At a £1.50 average CPC, it takes roughly £150 a day in total Google Ads spend just to qualify. Below that, the conversions from people who clicked “reject” never reach your PMax campaign, modelled or otherwise.

Put that next to the Starvation Spiral and the picture becomes clear. A UK business spending £2,000 a month can have real sales coming from PMax while the campaign itself sees only the consented fraction, too few to learn from. It then spirals into cheap inventory, and the owner concludes that “PMax doesn’t work for us”.

How to close the UK tracking gap

• Use advanced Consent Mode if your legal and privacy advice allows it. It sends cookieless pings when consent is denied and enables an advertiser-specific model rather than a general one.
• Turn on enhanced conversions. Hashed first-party data such as email addresses recovers conversions lost to cookie restrictions and cross-device journeys. It is the single highest-return tracking fix we implement.
• Move tagging server-side or use Google tag gateway, so browser blocking does not silently drop your purchase and lead events.
• Import offline conversions from your CRM for sales, booked appointments and signed contracts. These are consent-independent and are the best signal PMax can receive.
• Consolidate campaigns so the conversions you do capture feed one model rather than four thin ones.

The same measurement logic runs through paid social. Our guide to Meta Ads optimisation strategies that actually work covers the Conversions API equivalent.

Performance Max not converting for ecommerce: the fixes

For online shops, the product feed does more targeting work than any setting in Google Ads. Most ecommerce PMax failures we see trace back to the feed or the offer, not the bidding.

• Clear Merchant Center diagnostics first. Disapproved or limited products cannot serve, and price or availability mismatches between feed and site suppress whole product sets.
• Show VAT-inclusive prices and UK policies. UK consumer prices must include VAT, and missing shipping and returns information weakens how Shopping ads compete.
• Separate heroes from zombies. Use custom labels to split best-sellers, low-margin lines and products with no impressions. One PMax asked to sell 5,000 products on a small budget will spend on a few and ignore the rest.
• Run Standard Shopping alongside PMax where you need control. Since late 2024 they compete on Ad Rank, so a hybrid structure is viable again for long-tail products and margin protection.
• Optimise to value, not volume. Once you have 50+ conversions a month, Target ROAS on real order values stops PMax chasing cheap orders that look good on a conversion count.
• Check price competitiveness. If you are 20% dearer than the three retailers beside you in Shopping results, no bid strategy fixes that.
• Supply proper video. If you give PMax no video, Google may auto-generate one from your images. Short, product-led videos of your own usually perform better on YouTube placements.

Creative still matters inside PMax asset groups. Our framework for UGC ads that actually convert works just as well for PMax video and image assets.

Performance Max not converting for lead generation: the fixes

Lead-gen PMax has a different failure mode. It often does convert, on paper, while sales stay flat. That is still “not converting”, because the conversions it reports are not the ones you get paid for.

• Stop optimising to raw form fills. Display and YouTube inventory produces accidental and bot submissions that count as conversions. PMax learns that those placements “work” and buys more of them.
• Import qualified leads and sales from your CRM. Offline conversion imports, and Google’s Data Manager with its 2026 connectors for tools such as Mailchimp, ActiveCampaign and Klaviyo, let you bid on the outcomes that matter.
• Add friction on purpose. A required phone number, a postcode check or a qualifying question lowers junk volume and sharpens the signal.
• Assign values. Give a booked consultation a higher value than a brochure download, then move to value-based bidding once volume allows.
• Exclude existing customers with the 2026 first-party audience exclusions so budget goes to new pipeline, not re-engaging clients you already have.

How long should you wait before judging Performance Max?

Google’s guidance is to allow at least two weeks without changes after launch or a significant edit, and to evaluate Target CPA over a period containing at least 30 conversions, or 50 for Target ROAS. Here is how we turn that into a decision rule.

Day ranges and the 15-conversion cut-off are Global PPC operating defaults. Long sales cycles need longer windows because conversions are credited to the click date.

When Performance Max is the wrong tool

• New accounts with no conversion history. Build signal with tightly managed Search first, then add PMax.
• Very small budgets. If your daily budget is below your target CPA, you are asking PMax to learn from almost nothing.
• Long B2B sales cycles with no offline import. PMax will optimise to whatever it can see, usually low-quality form fills.
• A single product or service with tiny search demand. There may not be enough high-intent inventory for PMax to do anything but spill into Display.

If Google Ads is struggling and you are weighing other channels, read our ChatGPT Ads UK 2026 guide and the TIEBREAK framework for AI assistant ads before moving budget: the minimums and mechanics are very different.

The 60-minute Performance Max rescue plan

1. Confirm the goal (10 min). Open the campaign’s goal settings. Note which conversion actions are primary and check each one is “Active” with recent conversions. Remove anything that is not a real business outcome.
2. Test the tag (10 min). Complete a test purchase or lead with Tag Assistant running, once with cookies accepted and once rejected. Confirm the conversion and consent signals fire as expected.
3. Check signal volume (5 min). Count primary conversions in the last 30 days for this campaign. Under 30 means your first job is more signal, not more optimisation.
4. Read the channel split (5 min). Insights → Channel performance. If more than half the spend is on Display and YouTube with no conversions, you are in the spiral.
5. Check bidding (5 min). Is the campaign “Limited by budget” on a target strategy? Compare target to actual CPA in the Bid Target Adjustment Tool and set a target you can afford.
6. Mine the search terms (10 min). Export the Search terms report. Add obvious waste as campaign-level negatives and decide your brand exclusion policy.
7. Audit landing pages and locations (10 min). Review final URLs, add URL exclusions, and set location options to “Presence”.
8. Freeze and schedule (5 min). Make your changes in one batch, note the date, and don’t touch targets for 14 days. Put a review in the calendar.

Frequently asked questions

Hassan Tahir

Founder & Lead PPC Strategist · Global Performance Propel Company · Google Partner & Meta Business Partner