How to optimise your PMax campaigns

Per­form­ance Max (PMax) campaigns from Google Ads offer a fully automated way to run ads across all of Google’s channels. They use machine learning and strong signals to drive as many con­ver­sions as possible. To get the most out of PMax, it’s important to pair auto­ma­tion with smart, highly targeted op­tim­isa­tion.

Step 1: Set the right con­ver­sion goals

A suc­cess­ful PMax campaign starts with defining clear con­ver­sion goals and op­tim­ising for them. Be sure to dis­tin­guish between primary and secondary goals:

• Primary goals are the actions that drive business success. These might include completed purchases in e-commerce or lead form sub­mis­sions in B2B settings.
• Secondary goals are useful for tracking and improving campaign per­form­ance. These could be news­let­ter sign-ups or page visits.

Each business is likely to have different pri­or­it­ies when it comes to con­ver­sion actions. Retailers should focus on purchases and completed checkouts, as these directly increase revenue. Service providers, on the other hand, should pri­or­it­ise leads, contact requests or ap­point­ment bookings – actions that move potential customers closer to a sale.

Make sure each con­ver­sion action is meas­ur­able and generates enough data for machine learning to be able to optimise the campaign. Over­load­ing the primary goals with small or ir­rel­ev­ant actions can confuse the AI and slow op­tim­isa­tion. It’s also helpful to assign con­ver­sion values to each action to reflect their im­port­ance in achieving your campaign goals.

Don’t overlook the New Customer Ac­quis­i­tion goal. In the campaign settings, you can either bid more for new customers or target them ex­clus­ively. This approach is ideal if you’re aiming for growth and market expansion instead of re-engaging existing customers.

Step 2: Optimise asset groups

Asset groups are where your campaign’s creatives live. They bundle various ad assets (text, images, logos, videos) into cohesive units. Google uses gen­er­at­ive AI to auto­mat­ic­ally create or modify assets. By providing your landing page URL, you can have Google generate text sug­ges­tions and even imagery. This feature saves time but always review the results to ensure brand alignment.

Each asset group should focus on a specific theme, product or target audience. Google will then auto­mat­ic­ally combine the assets into ap­pro­pri­ate ad formats for YouTube, Gmail, Search, or Maps.

Use a variety of SEO-optimised headlines, high-quality images and short videos to cater to different channels. Headlines should be clear and concise, and tailored to the target audience. Use attention-grabbing visuals and ensure videos are short and impactful. Videos are par­tic­u­larly important for reach on YouTube, so don’t skip them – they improve campaign per­form­ance across multiple channels.

Step 3: Use audience signals

While PMax campaigns are highly automated, you can still fine-tune per­form­ance by providing audience signals. In Google Ads, audience signals help guide the AI in targeting the most relevant audience for your products or services, improving its decision-making. Adding demo­graph­ics, interests or custom audiences helps the AI focus on the right groups.

Here are some best practices to guide Google without limiting auto­ma­tion:

• Test different signals in separate asset groups: Assign a different audience signal to each asset group. This helps you see which com­bin­a­tions deliver the best results in terms of reach and con­ver­sion.
• Use your own data: Upload your customer data—like re­market­ing lists or Customer Match audiences. These signals are es­pe­cially strong because they target people who’ve already in­ter­ac­ted with your brand, like by signing up or making a purchase.
• Add interest and intent-based segments: Include audiences based on interests, purchase intent or search behaviour to reach users more likely to take action.
• Refresh signals regularly: Review and adjust your audience signals so they align with seasonal trends or new campaign goals.
• Let signals guide the AI without limiting it: Audience signals should steer the AI in the right direction, without re­strict­ing how it learns or adapts.
• Stick to specific search themes: Only include search themes that closely match your product or what you offer. Broad or generic terms can reduce per­form­ance.

Step 4: Check data feed quality

For online stores, success with PMax starts with a clean, well-struc­tured product feed in Google Merchant Center. This feed holds all the key product details Google needs to display your ads correctly and reach the right users. Make sure your product data is complete, accurate and up to date. Even small issues, liked outdated prices, incorrect avail­ab­il­ity or vague product de­scrip­tions, can hurt ad per­form­ance. Use clear, de­script­ive titles and keyword-rich de­scrip­tions to help Google cat­egor­ise your products correctly.

To take it a step further, consider using Custom Labels. These let you group products by criteria such as margin, sea­son­al­ity, best­seller status, or campaign priority. Grouping products this way makes it easier to fine-tune your budget and focus on high-value items.

Check your feed regularly to keep your data up to date and accurate. The Dia­gnostics tool in Merchant Center makes it easy to spot and fix issues before they cause problems. High-quality data ensures your products appear more often, perform better and ul­ti­mately drive sales.

Step 5: Refine bidding strategies

Bidding strategies are key to the success of your PMax campaign. With Smart Bidding, Google auto­mat­ic­ally adjusts your bids to get the best possible results within your budget. That means you no longer need to manually set bids for each keyword or product. The AI makes these decisions based on your campaign goals.

Google offers several Smart Bidding options:

• tROAS (Target Return on Ad Spend): tROAS tells Google to optimise your bids based on the return you want from your ad spend. To do so, it adjusts your CPC bids (cost per click) to reach the target return. Example: If your target ROAS is 400%, Google aims to generate £4 in revenue for every £1 you spend.
• tCPA (Target Cost per Ac­quis­i­tion): tCPA tells Google how much you’re willing to pay, on average, for a con­ver­sion – like a purchase or a lead. It then aims to drive as many con­ver­sions as possible without going over that target.

If you’re running multiple campaigns at once, allocate budgets stra­tegic­ally. A PMax campaign needs enough budget to learn which com­bin­a­tions perform best, but it shouldn’t pull too much budget away from well-per­form­ing Search or Shopping campaigns. Review your results regularly and adjust targets like tROAS or tCPA based on seasonal trends, market shifts or new business goals.

Step 6: Keep op­tim­ising and testing

Auto­ma­tion doesn’t mean your PMax campaigns auto­mat­ic­ally run at their best. You still need to track per­form­ance and make regular im­prove­ments to get the best results. Keep a close eye on key metrics (KPIs) like con­ver­sions (number of completed actions), ROAS (the ratio of revenue to ad spend), im­pres­sion share (your share of im­pres­sions compared to com­pet­it­ors), and cost per con­ver­sion.

Google uses your campaign data to refine its approach over time, but you can speed up the process through targeted testing of specific variables. Always roll out changes gradually so you can see what actually moves the needle. Here a few test ideas to get you started:

• Asset tests: Try out different headlines, de­scrip­tions, images, or videos. Don’t just tweak them! Make bold changes in tone, messaging, calls to action or visual style to see what resonates with your target audience.
• Audience tests: Compare different audience signals to figure out which groups convert best.
• Offer or pricing tests: Ex­per­i­ment with discounts, bundles or price points to see what drives con­ver­sions.
• Landing page tests: Even small updates, like a clearer layout, new product images or a more prominent call-to-action button, can improve the user ex­per­i­ence and boost your con­ver­sion rate.

After each test, review the results and write down what you learned. Over time, this gives you a solid base for making smarter decisions and running stronger campaigns.

Step 7: Identify and eliminate sources of error

Even automated PMax campaigns can run into issues. Common mistakes that hurt per­form­ance include:

• Too little con­ver­sion data: If your campaign doesn’t log enough con­ver­sions, the AI can’t learn which users are most likely to convert.
• Re­pet­it­ive assets: If you keep showing the same headlines, images or video, your ads lose impact and variety – the system needs both to optimise per­form­ance.
• Outdated or in­com­plete data feeds: If your product info is wrong or in­com­plete, your ads might not be displayed properly, or your products may not appear at all.

If your PMax campaign isn’t per­form­ing as it should be, take a sys­tem­at­ic approach to identify and fix what’s not working. Regular tweaks usually lead to better long-term results than big one-off changes. Track the ad­just­ments you make and assess their impact. Go step-by-step:

1. Check your data setup: First, make sure you’re getting enough con­ver­sion data. Then check if your assets are varied and up to date, and your audience signals are on target.
2. Add what’s missing: Give the AI more to work with. Add con­ver­sion data, refresh audience signals or include ad­di­tion­al products in the campaign.
3. Test asset com­bin­a­tions: Try different headlines, images or videos to boost the relevance and appeal of your ads.
4. Dig into your campaign reports: Look closely at key metrics like con­ver­sions, ROAS and im­pres­sion share. Identify patterns, spot weak points and track what’s improving.
5. Adjust budget and bids: Start with small changes as these often make a bigger dif­fer­ence than large overhauls. The AI also needs time to adapt to new settings.

Why a PMax campaign isn’t a set-it-and-forget-it

Pmax does a lot auto­mat­ic­ally, but it still needs clear input and smart oversight to perform at its best. That’s because success now depends less on manual bidding and more on giving the AI the right data to work with.

1. Goals: Define clear con­ver­sion values and apply new customer rules when they align with your strategy.
2. Assets: Use strong images, videos and ad copy. AI-generated content is fine too.
3. Controls: Set brand ex­clu­sions and URL rules to keep your budget on track.
4. Patience: After making changes, give the algorithm 2 to 3 weeks to adjust before making further ad­just­ments.

With the right setup and steady input, PMax can deliver real, meas­ur­able results.