Performance Max Campaigns Guide 2026

Every time you open Google Ads and click “new campaign,” there it is. Performance Max. Sitting right at the top, the option Google clearly wants you to pick. Since 2022, Google has been steering advertisers toward this campaign type, and by 2026, the majority of Google Ads accounts in both the UK and the US have at least one PMax campaign running. The question worth asking is whether this campaign type genuinely works for your business, or whether it primarily works for Google’s bottom line. The honest answer sits somewhere between the two.

Performance Max bundles every Google-owned channel into a single campaign: Search, YouTube, Display Network, Discover, Gmail, Maps and Shopping. It uses machine learning to handle bidding, audience targeting and creative combinations automatically. For a busy marketer managing multiple client accounts or a business owner who doesn’t have time to build separate campaigns for every channel, that sounds appealing. But the convenience comes with trade-offs that aren’t always obvious until you’ve spent real money.

This guide covers how PMax campaigns actually work in practice, not just what Google’s help documentation says they do. We’ll look at the asset group structure, audience signals, channel distribution, e-commerce versus lead generation differences, and the optimisation steps that separate profitable campaigns from expensive experiments. All of this draws on campaign management experience across UK and US accounts spending anywhere from $2,000/month to six-figure budgets.

What Performance Max Is (and What It Isn’t)

The traditional Google Ads structure gives you separate campaign types for each channel. A Search campaign for text ads on Google Search. A Display campaign for banner ads across the web. A Video campaign for YouTube. A Shopping campaign for product listings. Each one has its own targeting settings, its own budget, its own ad formats, and its own reporting. You control everything, but you also manage everything.

Performance Max replaces that fragmented structure with a single campaign that spans all channels simultaneously. You provide a goal, a budget, creative assets (text, images, video) and audience signals. Google’s AI then decides how to allocate the budget across channels, which users to target, which creative combinations to show, and how much to bid. All in real time, all without asking you first.

Think of it this way. With standard campaigns, you’re the driver. You choose the route, the speed, the stops. With Performance Max, you’re a passenger who told the driver the destination and handed over the wallet. The driver might take a brilliant shortcut you never knew about, or might take the scenic route because it suits them.

The table above paints a fairly clear picture. PMax makes campaign management simpler, but it does so by taking control away from the advertiser. You can’t see exactly which search terms triggered your ads (though Google has improved this with the Insights tab). You can’t shut off an underperforming channel. You can’t manually adjust bids for specific times of day. For experienced paid media managers, this can feel like being asked to cook dinner while wearing oven mitts.

Where PMax Delivers (and Where It Doesn’t)

There’s a lot of noise around Performance Max. Google’s own case studies paint consistently rosy pictures. Independent advertisers tell more mixed stories. Cutting through both requires separating what genuinely works from what merely looks good on a dashboard.

The Genuine Advantages

Cross-channel reach from a single campaign. For a small business or lean marketing team, building and managing separate Search, Display, YouTube, Shopping, Discovery and Gmail campaigns is a full-time job. PMax compresses all of that into one campaign with one budget. The time savings alone can justify using it, especially if you don’t have a dedicated paid media specialist on staff.

Machine learning at scale. Google’s bidding algorithms process billions of signals that no human could evaluate: device type, browser, time of day, location, previous browsing history, weather patterns, even macroeconomic trends. When the system has enough conversion data to learn from, it can find pockets of high-intent users that manual targeting would miss entirely. We’ve seen PMax campaigns discover converting audiences on Discover and Gmail that nobody would have thought to target manually.

New customer acquisition. Standard Search campaigns only reach people who are actively searching for your keywords. PMax can identify users who show purchase intent through their broader online behaviour, even if they haven’t searched yet. Someone watching product review videos on YouTube, browsing competitor websites, or reading industry articles on Discover can be reached before they ever type a query into Google.

Unified Shopping and Search. For e-commerce advertisers, PMax replaced Smart Shopping campaigns and now serves as the primary way to run product listing ads alongside text ads, YouTube placements and Display inventory. Having all of these in one campaign means Google can coordinate the touchpoints rather than having them compete against each other in separate campaigns.

The Real Weaknesses

The black box problem. This remains PMax’s most significant drawback. While Google added channel-level reporting in mid-2025 and expanded the Insights tab with search category data in 2026, the level of visibility is still far below what a standard Search campaign provides. You can see broad search categories driving traffic, but you can’t pull a full search terms report and identify every query your ads appeared for. For advertisers spending thousands per month, this lack of transparency is a genuine concern.

Brand cannibalisation. This is the issue that skews more PMax performance reports than any other factor. When someone searches for your brand name, they were already going to find you. If PMax shows them an ad and they click it, PMax gets credit for a conversion that would have happened anyway. Your campaign dashboard shows brilliant ROAS figures, but the actual incremental value is questionable. Google introduced brand exclusions for PMax in 2025. Use them. Always run branded search in a separate campaign and keep PMax focused on non-branded acquisition.

Underperformance on small budgets. PMax’s machine learning engine needs data to function properly. Running a campaign on $20-30 per day doesn’t generate enough clicks or conversions for the algorithm to learn which users convert. The result is a prolonged learning period, erratic performance, and the common conclusion that “PMax doesn’t work.” It often does work. It just needs a budget large enough to produce statistically meaningful data within a reasonable timeframe.

Loss of granular control. No keyword-level bidding. No ability to pause a specific channel. Limited negative keyword support (account-level only, added late 2025). No dayparting controls. No device bid adjustments. For experienced PPC managers who are used to pulling individual levers, PMax can feel frustratingly opaque. You’re giving directions from the back seat, not driving the car.

Asset Groups and Creative Structure

The core building block of any PMax campaign is the asset group. Each asset group contains a collection of creative materials that Google mixes and matches to produce hundreds of different ad combinations. You can have multiple asset groups within a single PMax campaign, each targeting different products, services or audience segments.

Getting the asset group structure right matters more than most advertisers realise. A poorly constructed asset group forces Google to work with weak materials, and even the best algorithm can’t compensate for bad creative.

What Goes into an Asset Group

Text assets. Up to 15 headlines (30 characters each), 5 long headlines (90 characters), 5 descriptions (90 characters) and your business name. Include a mix of brand-focused, benefit-focused and action-focused headlines. Google will test different combinations across channels, so variety matters. A headline that works on Search might fall flat on Display. Give the system options to work with.

Image assets. Minimum 3 landscape images (1200×628), 3 square images (1200×1200) and 1 portrait image (960×1200). Upload the maximum of 15 in each orientation if possible. Logos are required: 1 square (1200×1200) and 1 arena (1200×300). Keep text overlay below 20% of the image area. Use real product photography and genuine workplace images rather than stock photos. Users have become remarkably good at spotting generic stock imagery, and it damages credibility.

Video assets. Upload at least one video. If you don’t, Google will auto-generate videos from your images, and the results are almost always poor. Short videos of 10-30 seconds perform best for PMax placements on YouTube and Discover. Ideally, provide videos in three formats: vertical (9:16) for Shorts and mobile placements, square (1:1) for Discovery and social-style feeds, and field (16:9) for standard YouTube.

Final URLs. Each asset group needs a destination page. Here’s the catch: PMax enables “Final URL expansion” by default. This means Google can override your chosen landing page and send traffic to a different page on your site that it considers more relevant. For e-commerce stores with product pages, this can work well. For service businesses, it can mean your carefully designed landing page gets bypassed in favour of a blog post or about page. Turn this off for lead generation campaigns, or use URL rules to restrict which pages Google can redirect to.

How Many Asset Groups Do You Need?

There’s no magic number. The principle is straightforward: products or services that have meaningfully different audiences, different imagery and different messaging should be in separate asset groups. A furniture retailer would create separate asset groups for sofas, bedroom furniture and dining tables because each has a different buyer profile and visual language.

As a general guideline, 3-7 asset groups works well for e-commerce accounts and 2-4 for service-based businesses. Don’t over-segment. Each asset group needs enough budget and conversion data to optimise properly. Twenty asset groups competing for a $100/day budget means none of them get enough data to learn.

Google assigns performance labels to each asset: “Low,” “Good” and “Best.” Replace underperforming assets, but do it gradually. Swapping out everything at once resets the learning process. Change 2-3 assets at a time, wait 1-2 weeks, then evaluate again.

Setting Up a PMax Campaign

Creating a PMax campaign inside the Google Ads interface takes a few minutes. Setting it up properly takes considerably longer, and the work that matters most happens before you ever open the campaign builder.

Prerequisites

Conversion tracking is non-negotiable. PMax’s entire optimisation engine depends on knowing which users convert and which don’t. If your tracking is missing, broken, or misconfigured, PMax has no foundation to build on. It will still spend your budget. It just won’t spend it intelligently.

For UK advertisers, make sure your conversion tracking setup complies with ICO cookie consent requirements. For US advertisers, state-level privacy regulations (particularly in California and Colorado) may affect how conversion data is collected. Without proper consent mechanisms, your conversion data will have gaps that undermine PMax’s learning.

E-commerce advertisers also need an active, up-to-date Google Merchant Center account. Your product feed quality directly determines Shopping performance within PMax. We’ll cover feed optimisation in the e-commerce section below.

Bidding Strategy Options

PMax offers two automated bidding strategies, and picking the right one matters more than most guides suggest:

• Maximise conversions: Aims to generate the highest number of conversions within your daily budget. You can optionally set a target CPA (cost per acquisition). Setting a target CPA of $50 or £40 tells the system to aim for that average cost per conversion. If you set it too aggressively low, the system restricts delivery and you’ll see minimal spend.
• Maximise conversion value: Aims to generate the highest total conversion value (revenue). The right choice for e-commerce. You can optionally set a target ROAS. A target ROAS of 500% means you want $5 (or £5) in revenue for every $1 (or £1) spent on ads.

A practical recommendation: launch without a target CPA or ROAS for the first 2-4 weeks. Let the system collect data and establish baseline performance. Once you have at least 30-50 conversions, look at your actual CPA or ROAS figures, then set a realistic target. Starting with an aggressive target from day one starves the algorithm of data and produces worse results than no target at all.

Budget Considerations for UK and US Advertisers

The minimum budget question comes up constantly. Google doesn’t set a hard minimum, but from practical experience, PMax needs enough daily budget to generate 10-15 clicks at a minimum. Given that average CPCs vary significantly by industry:

If your monthly Google Ads budget is under $1,500 (or roughly £1,200), allocating the entire amount to PMax is risky. A better approach at that budget level: put 60-70% into a focused Search campaign targeting your highest-intent keywords, and allocate 30-40% to PMax for broader discovery. Once your monthly budget crosses $3,000-5,000, PMax can take a larger share because there’s enough data flowing through the system for meaningful learning.

Audience Signals That Actually Work

One of the most misunderstood aspects of PMax is how audience signals function. They’re not targeting parameters in the traditional sense. They’re suggestions. You’re telling Google: “Start by looking at people like this.” Google uses those signals as a starting point, then expands to other audiences it identifies as likely to convert. Over time, the algorithm may drift considerably from your original signals if it finds better-performing segments elsewhere.

This means the quality of your initial signals matters most during the first few weeks, when the algorithm is still learning. After 4-6 weeks, PMax will largely be following its own data regardless of what you originally provided. But strong signals accelerate that learning period and reduce wasted spend during the critical early days.

High-Value Signal Sources

Customer match lists. Upload your existing customer email addresses. Google matches them against its user database and builds lookalike profiles. The more customers on your list, the better. A list of 1,000+ matched emails gives the system a meaningful foundation. If you have purchase value data, segment your list: upload your highest-value customers as a separate signal and your full customer list as another. This teaches PMax what your best customers look like, not just your average ones.

Website visitor audiences. Your remarketing lists are extremely valuable signals. Users who visited your site but didn’t convert represent warm prospects that PMax can re-engage across channels. Users who did convert represent the profile PMax should look for more of. Both lists serve different but complementary functions.

Custom segments. You can build segments based on search behaviour (people who search for specific terms), website activity (people who browse specific competitor sites or industry sites), or app usage. Custom segments built around competitor brand names and high-intent search queries tend to produce the strongest results. For instance, a London-based accounting firm might create a custom segment of users who recently searched for “accountant near me,” “small business tax advice” or “limited company formation.”

Google’s pre-built segments. In-market audiences (people actively shopping for products in your category), life events (people who recently moved, got married, started a new job) and affinity audiences (people with strong long-term interests) are all available. In-market audiences are the most directly actionable for performance campaigns.

Running PMax without any audience signals is technically possible, but the learning period will be longer and the first few weeks will burn more budget on exploratory traffic. Providing strong signals up front is like giving a new employee a thorough briefing on day one rather than letting them figure everything out by trial and error.

E-Commerce PMax Strategy

Performance Max is at its strongest in e-commerce. Google has said so publicly, and the data from our own campaigns supports it. E-commerce PMax campaigns report 12-18% more conversions on average compared to standard Shopping campaigns running in the same account. But that uplift depends heavily on two things: feed quality and proper brand separation.

Your Product Feed Makes or Breaks PMax

The Merchant Center feed is the foundation of e-commerce PMax. If your feed is poorly optimised, no amount of asset quality or audience signal work will compensate. Feed optimisation is the single highest-leverage activity you can do for e-commerce PMax performance.

Product titles need to match how people actually search. “Men’s Black Leather Wallet” is fine for a catalogue, but “Genuine Leather Bifold Wallet for Men – Black – RFID Blocking” mirrors the way users search and helps Google match your product to relevant queries. Include key attributes: material, colour, size, target gender and unique selling points.

Product descriptions should be at least 500-750 characters. Images should be high-resolution, on a clean background, shot from multiple angles. Google’s Shopping algorithm rewards feeds with complete data: GTINs (Global Trade Item Numbers), correct product types, properly mapped Google product categories, shipping details and sale price annotations.

Most UK e-commerce platforms (Shopify, WooCommerce, BigCommerce) and US-popular platforms (Shopify, Magento, WooCommerce) generate product feeds automatically via plugins or apps. But the default output from these tools is rarely sufficient. Budget time for manual feed enhancement. It’s tedious work, but the performance gap between an optimised feed and a default one is substantial.

Segmenting Asset Groups by Margin

The standard approach is to create asset groups based on product categories. That’s a reasonable starting point. But a more commercially minded approach is to segment by profit margin. Group your highest-margin products into one asset group with a more aggressive ROAS target. Group your lower-margin products separately with a higher ROAS floor to protect profitability.

For seasonal products, create dedicated asset groups that can be paused and reactivated as seasons change. Pausing and restarting an entire PMax campaign resets the learning period. But pausing an asset group within a running campaign is less disruptive. Keep the campaign running year-round and manage seasonality at the asset group level.

The Brand Cannibalisation Problem

This deserves its own section because it distorts so many PMax performance reports. When a user searches for your brand name, clicks a PMax ad, and then purchases, PMax takes credit for that sale. Your campaign metrics look fantastic. But that customer was already looking for you. They would have clicked your organic listing or typed your URL directly. The PMax ad didn’t create that demand; it just inserted itself into an existing journey and charged you for the privilege.

Google added brand exclusions to PMax in 2025. Activate them immediately. Route all branded search traffic through a separate Search campaign where you have full keyword-level control and pay a fraction of the CPC. Let PMax focus on non-branded discovery and new customer acquisition, which is where its cross-channel capabilities actually add incremental value.

When brand traffic is properly separated, PMax ROAS will typically drop compared to the blended figures you were seeing before. That’s not a problem. It means you’re now seeing real performance instead of inflated numbers. The campaigns that survive this reality check are the ones genuinely working.

Lead Generation PMax

E-commerce gets most of the PMax attention, but service businesses and B2B companies use it too. The dynamics are quite different, though, and the pitfalls are sharper.

In e-commerce, a conversion is binary: someone bought or they didn’t. In lead generation, a conversion is typically a form submission or phone call. The problem is that not every form fill represents a qualified lead. Someone researching “what is Google Ads” might land on your blog, fill in a contact form out of curiosity, and have no intention of becoming a paying client. PMax records that as a conversion and starts looking for similar users. The result: your lead volume goes up, but lead quality goes down. You end up fielding more enquiries that go nowhere.

Improving Lead Quality Within PMax

Three approaches work well here, and the best campaigns combine all three:

Offline conversion import. Connect your CRM to Google Ads and feed back actual sales outcomes. If 100 people fill your form but only 8 become paying clients, Google needs to know which 8. Upload that data through Google Ads offline conversion tracking or the Google Ads API. The system then optimises for users who resemble your actual clients, not just anyone willing to fill in a form. This requires regular data uploads (weekly is ideal, fortnightly at minimum), but it’s the single most effective improvement you can make to lead gen PMax.

Layered conversion actions. Don’t rely on a single conversion event. Set up primary conversions (form submission, phone call) alongside secondary conversions (time on site above 3 minutes, key page views, WhatsApp clicks). Giving Google multiple signals about user quality improves the optimisation surface area.

Disable final URL expansion. This is essential for lead gen. You don’t want traffic landing on blog posts, about pages, careers pages or privacy policies. Restrict PMax to your service pages and landing pages only. Use URL rules if you want to allow a set of specific pages rather than just one.

PMax Plus Search: The Combined Approach

For lead generation, PMax should complement your Search campaigns, not replace them. A standard Search campaign targeting high-intent keywords (“google ads agency london”, “ppc management services UK”, “adwords management cost”) gives you precise control over exactly which queries trigger your ads. PMax then expands your reach to users who haven’t searched yet but match the profile of your converting audience.

The budget split for lead gen accounts running both: 60-70% to Search, 30-40% to PMax. As PMax matures and starts generating quality leads (verified through offline conversion data), you can gradually shift the balance. But don’t abandon Search entirely. It remains the highest-intent, most controllable channel in the Google Ads ecosystem.

Channel Distribution and Reporting

One of the most common questions from advertisers running PMax is: “Where is my money actually going?” For years, this was effectively unanswerable. Google treated PMax as a single opaque unit with no breakdown of spend or performance by channel. That changed in mid-2025 when channel-level reporting was added.

Reading the Channel Report

You can now see what percentage of your PMax budget is being allocated to Search, Display, YouTube, Shopping, Discover, Gmail and Maps. This data is revealing, and sometimes alarming. A pattern we’ve seen across multiple accounts: PMax campaigns with weak video assets and strong image assets tend to push 50-60% of budget into Display Network. The ROAS looks acceptable in aggregate, but when you isolate the Display spend, it’s mostly brand awareness impressions with very low direct conversion contribution.

If you see this pattern in your account, it usually means one of two things: your audience signals are too broad (so PMax defaults to cheaper Display inventory to stay within ROAS targets), or your Search and Shopping assets aren’t competitive enough to win meaningful auction volume.

The fix varies by situation. Strengthening video assets often pulls budget back toward YouTube and Discover. Tightening audience signals with customer match lists can shift allocation toward Search. Adding negative keywords at the account level removes low-quality search queries that were degrading the Search channel’s performance.

The Insights Tab

Google’s Insights section within PMax has improved considerably in 2026. It now shows search category performance (which groups of search terms are driving conversions), audience segment breakdown (which demographic and interest groups are most active), and trending signals (emerging opportunities based on seasonal or market changes).

Check this tab weekly. The search categories report, in particular, often reveals surprises. You may find that PMax is generating conversions from search categories you hadn’t considered targeting, or conversely, that a significant portion of your budget is going toward irrelevant categories that need to be excluded via negative keywords.

Ongoing Optimisation

There’s a persistent myth that PMax is a “set and forget” campaign type. It’s not. The nature of optimisation is different from traditional campaigns, but the need for regular attention is no less real.

Weeks 1-4: The Learning Period

After launch, PMax enters a learning phase of 2-4 weeks. During this period, the system is testing creative combinations, exploring audience segments and calibrating bids. Performance will be volatile. Some days will look brilliant, others will look wasteful. The critical rule during this period: don’t intervene. Don’t change the budget. Don’t change the target CPA/ROAS. Don’t swap assets. Doing any of these restarts the learning clock.

The exception: if something is clearly broken (conversions not tracking, landing page down, incorrect product feed), fix it immediately. But don’t make strategic changes during the learning phase. Patience is genuinely part of the process here.

Weeks 5-8: First Optimisation Pass

Once the learning period ends, begin with asset performance review. Check the performance labels on every text headline, image and video. Replace assets marked “Low” with fresh alternatives, but limit changes to 2-3 assets at a time. Run the updated configuration for another 1-2 weeks before making further changes.

Set a target CPA or ROAS based on your actual performance data from the learning period. If your average CPA during weeks 1-4 was $65, start with a target of $60-65. Don’t immediately try to cut it to $40. The algorithm needs headroom. Tightening targets too fast restricts delivery and can actually increase your average CPA as the system chases fewer, more expensive conversions.

Negative Keyword Management

Account-level negative keyword lists became available for PMax in late 2025. Use them aggressively. Review the Insights tab’s search categories report and identify irrelevant categories. Add the associated terms to your negative list.

A practical example: a dental clinic running PMax was spending significant budget on informational queries like “home remedies for toothache” and “can I pull my own tooth.” These users had no intention of booking an appointment. After adding these terms and related variations as negatives, the campaign’s CPA dropped by roughly a third. The spend that had been going to information-seekers was reallocated to commercial-intent queries where users actually wanted to book a dentist.

Experiment and Iterate

Google Ads offers an experiments framework for PMax campaigns. You can duplicate your campaign, modify one variable (different asset groups, different audience signals, different ROAS target), split traffic between the original and experiment, and measure which performs better with statistical confidence.

This is one of the most underused features in PMax management. Rather than making changes to your live campaign and hoping for the best, run a controlled experiment. Test whether a lower ROAS target actually increases total profit. Test whether removing Display-heavy assets shifts budget to higher-converting channels. Test whether a customer match list outperforms a custom segment signal. Make decisions based on data rather than instinct.

Seasonality Adjustments

For UK advertisers, key seasonal peaks include January sales, Easter, summer holidays, Black Friday and Christmas. For US advertisers: Presidents’ Day sales, Memorial Day, Back-to-School (July-August), Black Friday, Cyber Monday and the December holiday season. PMax takes 3-5 days to adapt to budget changes, so plan ahead.

Increase budgets a full week before the seasonal peak, not on the day itself. Use Google’s seasonality adjustment feature to tell the system that conversion rates will temporarily increase by a specific percentage during your peak period. This overrides the machine learning model temporarily, allowing the system to bid more aggressively during high-conversion windows without waiting for real-time data to confirm the trend.

PMax in 2026: What’s Changed

Google continues to invest heavily in PMax, and the 2026 updates address several long-standing complaints from the advertising community. The most significant changes:

Greater transparency. Channel-level reporting, search term categories, asset combination performance data and audience segment breakdowns are all more detailed than they were a year ago. Google yielded on transparency because large advertisers (the ones spending hundreds of thousands per month) were threatening to pull PMax budget entirely without better visibility into where their money was going.

Enhanced first-party data integration. Google Ads Data Manager now makes it easier to import CRM data, app analytics data and offline sales data directly into Google Ads. For lead gen advertisers, this is the most impactful update. The easier it is to feed back real conversion outcomes, the better PMax performs. The technical barrier that previously stopped many businesses from implementing offline conversion tracking has been significantly reduced.

AI-generated assets. Google now offers AI-powered headline suggestions, image variations and even basic video generation within PMax. These tools are powered by Gemini and draw on your existing website content, product feed and historical campaign data. They’re genuinely useful as a starting point or for filling gaps in your asset library. But don’t rely on them exclusively. AI-generated copy tends to sound generic, and AI-generated images still carry a recognisable artificiality. Use them as supplements, then refine with your own brand voice and visual identity.

Improved new customer acquisition controls. PMax now lets you bid higher specifically for new customer acquisition versus returning customers. You can set a premium for new customers, encouraging the system to prioritise expanding your customer base rather than re-converting existing buyers. For growing businesses, this feature aligns PMax more closely with actual business objectives.

Frequently Asked Questions