Performance Max campaigns are Google’s AI-driven campaign type that runs ads across all Google properties simultaneously: Search, Display, YouTube, Gmail, Discover, and Maps. They use machine learning to allocate budget and generate ad combinations based on the assets you provide and the conversion goals you set. They work exceptionally well for e-commerce brands with strong product feeds and clear conversion tracking, but they can waste significant budget for lead generation businesses without proper guardrails.
“Performance Max is the most polarizing campaign type in Google Ads right now. Half the advertisers I talk to swear by it, the other half are burning money. The difference isn’t the tool, it’s whether the business model fits the campaign type,” says Hardik Shah, Founder of ScaleGrowth.Digital.
What Exactly Is a Performance Max Campaign?
Performance Max (PMax) is a goal-based campaign type where you provide creative assets (headlines, descriptions, images, videos), set a conversion goal, and let Google’s machine learning decide where, when, and how to show your ads. You don’t choose placements, write individual ads for each network, or set keyword-level bids. Google handles all of that.
At a technical level, PMax uses Google’s auction-time signals (device, location, time of day, audience segments, search context) to make real-time decisions about which combination of your assets to show to which user on which platform. It’s a single campaign that replaces what used to require 4-5 separate campaign types.
Google launched PMax in November 2021 and has been steadily expanding its capabilities since. As of early 2025, it accounts for roughly 40% of all Google Ads spend globally, according to data from Optmyzr’s analysis of their advertiser base. For Shopping advertisers specifically, that number is closer to 65% since Smart Shopping was automatically migrated to PMax in 2022.
When Do Performance Max Campaigns Work Best?
PMax excels in specific conditions. It’s not universally good or universally bad. Understanding where it fits saves you from either missing an opportunity or wasting budget on the wrong campaign type.
E-commerce with a strong product feed. This is PMax’s sweet spot. If you have a Google Merchant Center feed with hundreds or thousands of products, proper titles, descriptions, images, and pricing, PMax will outperform standard Shopping campaigns in most cases. Google’s algorithm can test millions of product-audience-placement combinations that no human could manage manually. We’ve seen e-commerce brands improve ROAS by 15-30% after switching from standard Shopping to PMax, provided their feed quality was solid.
High conversion volume. PMax needs data to learn. Google’s recommendation is 30+ conversions in the past 30 days at minimum. Below that, the algorithm doesn’t have enough signal and will make poor allocation decisions. Accounts generating 50-100+ conversions per month see the best results.
Clear, measurable conversion actions. If your conversions are well-defined (purchases, qualified form submissions with value assignment, app installs), PMax can optimize effectively. If your conversion tracking is messy, the algorithm optimizes toward garbage.
Multiple Google properties matter to your audience. If your customers engage with YouTube, browse Discovery feeds, and search on Google, PMax can follow them across these touchpoints. Single-channel businesses (search-only, for example) may not benefit from the cross-network distribution.
When Should You Avoid Performance Max?
Here’s where it gets uncomfortable, because Google would prefer you use PMax for everything. But the data from our accounts tells a different story in certain situations.
Lead generation with long sales cycles. If your conversion is a form fill and the actual sale happens 30-90 days later through a sales team, PMax optimizes for the form fill, not the sale. We’ve seen PMax flood lead-gen accounts with low-quality leads because it found cheap conversions (spam submissions, tire-kickers) on Display and Discover. The CPA looked great, but the sales team was drowning in junk.
The fix exists (offline conversion tracking, where you feed actual sales data back to Google), but many businesses don’t have their CRM-to-Google integration set up. Without that feedback loop, PMax will always optimize for quantity over quality in lead gen.
Low conversion volume. Accounts with fewer than 15-20 conversions per month shouldn’t use PMax. The algorithm will make erratic decisions, and you’ll have no idea where your money went because PMax’s reporting is notoriously opaque.
Brand-sensitive businesses. PMax generates ad combinations from your assets automatically. You can’t control exactly which headline pairs with which description or which image appears on which network. If your brand has strict creative guidelines, this lack of control is a problem. We’ve seen ads show up with awkward headline/description combinations that no human would approve.
When you need granular reporting. PMax is a black box compared to standard campaigns. You can’t see which search queries triggered your ads (Google shows “insights” but not complete query data), you can’t see performance by placement, and you can’t easily attribute conversions to specific assets. If your reporting requirements are detailed, this is a real limitation.
What Assets Does Performance Max Need?
PMax builds ads from asset groups. Each asset group contains the creative elements Google combines into ads. The quality and quantity of these assets directly affect campaign performance because Google needs options to test.
| Asset Type | Minimum | Recommended | Specifications |
|---|---|---|---|
| Headlines | 3 | 5-15 | Up to 30 characters each. Vary messaging angles. |
| Long headlines | 1 | 5 | Up to 90 characters. Used on Display and Discover. |
| Descriptions | 2 | 4-5 | Up to 90 characters. Short description (60 char) for smaller formats. |
| Images | 1 | 5-20 | market (1200×628), square (1200×1200), portrait (960×1200). |
| Logos | 1 | 2 | Square (1200×1200) and market (1200×300). |
| Videos | 0 | 1-5 | YouTube videos, 10+ seconds. If you don’t provide one, Google auto-generates it. |
| Call to action | 1 | 1 | Automated or select from preset options. |
| Final URL | 1 | Multiple via URL expansion | Landing page for each asset group. |
One thing that trips people up: Google will auto-generate video ads from your images if you don’t provide video assets. These auto-generated videos are almost always terrible; they’re slideshow-style with generic music and look like they were made in 2012. Always upload at least one proper video, even a simple one, to prevent Google from creating its own.
How Do You Control Where Performance Max Spends Your Budget?
This is the biggest frustration advertisers have with PMax. You set a budget and a goal, and Google distributes spend across networks as it sees fit. You can’t say “spend 70% on Search and 30% on YouTube.” Google decides.
In practice, what we observe across our managed accounts: PMax tends to allocate 50-70% of budget to Shopping (if you have a product feed), 15-30% to Search, and the remainder across Display, YouTube, and Discover. For non-e-commerce accounts without a Shopping feed, the split skews heavily toward Display and Discover, which often means cheaper but lower-quality traffic.
The controls you do have:
- Audience signals. You can tell Google which audiences to prioritize (your customer lists, in-market segments, custom intent audiences). These aren’t restrictions, they’re hints. Google can and will go beyond them, but strong audience signals help the algorithm find the right users faster.
- Asset group structure. Create separate asset groups for different products, services, or themes. Each asset group can have its own landing page, images, and copy. This gives the algorithm better-matched creative for different contexts.
- Brand exclusions. As of 2024, you can exclude your own brand terms from PMax to prevent it from cannibalizing branded traffic that would convert organically or through a cheaper branded Search campaign. Use this. Always.
- URL exclusions. You can exclude specific URLs from Final URL Expansion, which stops Google from sending traffic to pages you don’t want used as landing pages (blog posts, about pages, careers pages).
- Negative keywords at account level. Google added the ability to apply negative keywords to PMax at the account level in 2024. Use this to block irrelevant queries. It’s not as granular as campaign-level negatives in Search, but it’s better than nothing.
How Should Performance Max and Search Campaigns Work Together?
Google’s official guidance says PMax “complements” Search campaigns. In reality, there’s overlap, and managing that overlap is critical to avoiding cannibalization.
Here’s how the priority works: if a search query matches an exact match keyword in your Search campaign, the Search campaign takes priority over PMax. If it doesn’t match an exact keyword, PMax competes in the auction alongside your phrase and broad match keywords, and whichever has the higher Ad Rank wins.
What this means practically: PMax will absorb most of your broad match and some of your phrase match traffic. If you want Search campaigns to retain control over specific queries, use exact match keywords in your Search campaigns.
Our recommended structure for accounts running both:
- Branded Search campaign with exact match brand keywords. This ensures brand queries don’t get absorbed by PMax.
- Core non-branded Search campaign with exact and phrase match keywords for your most important terms. PMax won’t override these.
- Performance Max campaign with brand exclusions enabled and strong audience signals. Let PMax handle the discovery, the long-tail queries, and the cross-network distribution.
Run this structure for 30 days and compare: what queries does PMax find that your Search campaigns missed? Are PMax conversions incremental, or is it just taking credit for traffic you were already getting? The data will tell you how to adjust.
How Do You Measure Performance Max Effectively?
PMax reporting is its weakest point. Google shows you top-line metrics (cost, conversions, ROAS) but hides much of the detail you need. Here’s how to work around the limitations.
Use the Insights tab. PMax’s Insights tab shows search themes (grouped queries), audience segments driving conversions, and asset performance ratings. It’s not a search terms report, but it’s the closest thing you’ll get. Check it weekly.
Monitor asset performance. Google rates each asset as “Low,” “Good,” or “Best.” Replace “Low” assets with new ones every 2-3 weeks. The algorithm performs better with fresh creative, and removing underperformers stops Google from showing weak combinations.
Track incrementality. This is the big question: are PMax conversions genuinely new, or would they have happened anyway through your other campaigns? We measure this by running hold-out tests. Pause PMax for a week, see if total account conversions drop proportionally. If they don’t, PMax was cannibalizing.
Use scripts for deeper reporting. Google Ads scripts can extract placement-level data from PMax campaigns that isn’t visible in the interface. Mike Rhodes’ PMax script (free, widely used) pulls placement URLs so you can see where your Display and YouTube spend is actually going. We run this monthly for every PMax campaign we manage.
What Are the Most Common Performance Max Mistakes?
After managing PMax campaigns across 20+ accounts since its launch, these are the errors we see most often:
1. Not excluding brand terms. PMax loves brand traffic because it converts cheaply and makes the campaign’s CPA look great. But you were going to get those brand conversions anyway, probably at lower cost through a branded Search campaign or organically. Always exclude your brand from PMax and measure non-branded performance separately.
2. Using one asset group for everything. If you sell 5 different product categories, one asset group with generic messaging will underperform. Create separate asset groups per category with tailored headlines, images, descriptions, and landing pages. More asset groups means better creative-audience matching.
3. Skipping video. Google auto-generates a video from your images if you don’t provide one, and these auto-generated videos are embarrassingly bad. Upload at least one real video per asset group, even if it’s just a 15-second product demonstrate.
4. Setting unrealistic CPA targets. If your historical CPA is Rs 3,000 and you set a Target CPA of Rs 1,000 on PMax, the algorithm will restrict delivery so aggressively that the campaign barely spends. Start with a target 10-20% above your actual CPA and gradually reduce it as the campaign learns.
5. Judging too early. PMax needs 4-6 weeks of learning before you can trust its performance data. Making changes during the first two weeks resets the learning period. Set it up correctly from the start, then leave it alone for at least 30 days.
Should You Use Performance Max for Lead Generation?
You can, but with significant caveats. PMax for lead gen is a very different animal than PMax for e-commerce.
The core challenge: PMax optimizes for whatever conversion action you tell it to. For e-commerce, that’s purchases, with clear revenue values. For lead gen, it’s typically form submissions, which have no immediate revenue signal. PMax doesn’t know the difference between a qualified lead worth Rs 50,000 and a spam submission worth nothing. It just sees “conversion” and optimizes for more of them.
If you’re going to use PMax for lead gen, here’s what we recommend:
- Set up offline conversion tracking. Feed actual sales outcomes back to Google from your CRM. This tells the algorithm which leads became customers, so it can optimize for quality, not just quantity. Without this, PMax for lead gen is a gamble.
- Use lead form qualification. If you use Google’s Lead Form extension, add qualifying questions to filter out low-intent submissions.
- Monitor placement data. Use scripts to check where your Display impressions are going. Low-quality apps and sites can drive cheap but worthless form fills.
- Start with a smaller budget. Don’t give PMax your entire lead gen budget. Start at 20-30% and scale based on lead quality, not lead volume.
One client of ours, a B2B software company spending Rs 12 lakh monthly, moved 30% of their budget to PMax in Q3 2024. Leads increased 45%, but qualified leads dropped 20%. After implementing offline conversion tracking and excluding Display placements with high spam rates, qualified leads recovered within 6 weeks and eventually exceeded the pre-PMax baseline by 12%.
Performance Max is a powerful tool when matched to the right business model and set up with proper guardrails. It’s not a magic button that replaces campaign management. It’s a different kind of campaign management, one that requires you to think about inputs (assets, signals, data quality) rather than outputs (bids, placements, keywords). If that shift suits your business and team, PMax can deliver real results. If you need control and transparency, stick with Search and Shopping campaigns and add PMax as a supplementary channel. Get in touch if you want us to evaluate whether PMax fits your account.
Related Service