Performance Max Campaigns: Why 45% of All Conversions Now Go Through PMax

These numbers can't be ignored. And yet we still meet e-shop owners who have either never tried PMax, or switched it on back in 2023, got dismal results, and never came back to it. And honestly — we don't blame them. PMax in 2023 was a completely different tool from what it is today.

In this article we'll show you what has changed over the past year and a half, who PMax makes sense for (and who it doesn't), which mistakes are killing e-shops' campaign performance, and why managing PMax by hand no longer makes economic sense in 2026.

What Performance Max Is (and Why You Need to Care)

Performance Max (PMax for short) is a Google Ads campaign type that automatically shows your ads across all of Google's channels at once — in Search, on YouTube, in Gmail, on Google Maps, across the Display network and in Discover. Instead of building a separate campaign for each channel, PMax handles it for you.

It sounds simple. Under the hood, though, it's one of the most sophisticated advertising systems Google has ever built. Based on your data (product feed, creative, audiences), the algorithm decides who sees your ad, when and where. And it does so in real time, thousands of times a second.

Why do we mention it? Because Google is actively pushing the entire platform toward automation. PMax is part of the so-called "Power Pack" — the trio of PMax + Demand Gen + AI Max for Search, designed to cover the whole purchase funnel from first contact to conversion. Anyone not using PMax today is voluntarily ceding ground to competitors who are.

And the competition is. 82% of advertisers run PMax alongside other campaign types. The question isn't whether to try PMax — it's whether you've set it up correctly.

What Changed in PMax in 2025–2026 (and Why Old Articles Lie)

If the last thing you read about PMax was in 2023, forget it. Over the past year and a half Google has made fundamental changes that fix most of the old problems. Here are the most important ones.

Negative keywords — finally

This was PPC specialists' most common complaint from day one. "I can't exclude nonsense search queries." Since January 2025, you can. And the limit? 10,000 negative keywords per campaign. That's a huge step forward.

But there's a catch few people mention. Negative keywords only work on Search and Shopping inventory. Display, YouTube, Gmail and Discovery placements stay unfiltered. So yes, it's a great improvement, but not a cure-all.

Channel reporting — finally you know where the money goes

Another thing that frustrated PPC specialists: PMax was a "black box." You threw money in and had no idea how much went to YouTube, how much to Search and how much to Display. That's history. The new channel reporting shows a full performance breakdown by channel — Search, Display, YouTube, Gmail, Discover, even Waze Ads and Search Partners.

In practice, this means you can finally see whether your money is working where it should. A healthy e-shop should spend 60–80% of its budget on Shopping placements. If you see most of your budget flowing into Display, something is wrong.

Brand exclusions — separate out your branded traffic

This matters mainly for e-shops with a strong brand. PMax tended to "steal" branded traffic from your own Search campaigns, because branded conversions are cheap and the algorithm naturally chases the lowest CPA. According to the Optmyzr study, 91% of accounts have keyword overlap between Search and PMax.

The new brand exclusions solve this. For e-shops there's also a clever split — you can exclude branded search from text ads while keeping Shopping ads on your brand active. Which makes sense — in Shopping you always want to be visible.

What else changed?

• Search themes expanded from 20 to 50 — they help the algorithm find the right search queries faster.
• Demographic exclusions (beta) — you can exclude age groups or devices.
• High Value Mode — Google AI predicts which new customers will maximize lifetime value and automatically raises bids for those users.
• AI creative — Google Imagen 4 and Veo generate images and videos directly inside Google Ads.

To sum it up at a glance:

If you tried PMax a year ago and it didn't work — today it's a completely different game.

Who PMax Works For (and Who It Doesn't) — Honestly

We're not trying to sell you PMax at any cost. It works brilliantly for some types of business, but for others it can be an expensive learning phase on your dime. Here are the facts.

PMax is for you if:

You have an e-shop with at least 30 conversions a month. That's the absolute minimum for the algorithm to have enough data to optimize. Ideally 60 or more conversions a month — that's where PMax really starts to shine.

Your daily budget is at least €30 (ideally €50–100). Below that threshold the algorithm has no room to learn and results will be unstable.

You have a quality product feed. This is the be-all and end-all. Product titles have to match how people actually search, not your internal naming. Strong, distinctive images decide the click. A weak feed = weak results, no matter how well the campaign is set up.

According to DACH benchmarks (the German, Austrian and Swiss market, which is closest to us in account size and CPC), a properly configured PMax lowers average CPA by 31%. Those are far from negligible numbers.

Recommended structure by budget

The 2026 trend is clear — a hybrid PMax + Standard Shopping approach. PMax and Standard Shopping now compete at the Ad Rank level, so you get fair conditions for both campaign types.

7 Mistakes That Kill Your PMax Campaigns

We come across them with practically every new client. And honestly — we made mistake number 2 ourselves at the start.

1. You let Google generate the video

If you don't upload your own video to PMax, Google automatically builds one from your images and text. And the result? Let's put it diplomatically: it looks unprofessional. In the PPC community there's near-unanimous agreement on this — auto-generated video damages your brand image. Upload your own video; 15–30 seconds is enough. E-shops that added their own video report 25–40% better campaign performance.

2. One asset group for everything

Do you sell handbags, shoes and jewelry in your e-shop? And all in one asset group? That's like throwing winter coats and swimsuits into the same shop window. Every asset group needs a clear theme and intent. Otherwise the algorithm doesn't know who to show what, and the results get diluted.

3. You ignore brand cannibalization

91% of accounts have keyword overlap between Search and PMax campaigns. If you don't have brand exclusions set up, PMax is stealing your branded traffic — conversions you'd get more cheaply through classic Search. Set up brand exclusions. Now. Today.

4. Too little data, too small a budget

Below 30 conversions a month, results swing ±100 to 400% off target. One month a CPA of 150 CZK, the next month 600 CZK. That's not optimization, that's roulette. If you don't have enough conversions, start with classic Search campaigns and add PMax once you've gathered data.

5. You change settings during the learning phase

PMax needs 4–6 weeks to learn. Any major change — to budget, bidding strategy or assets — means a restart. Changed your ROAS target after two weeks because the results looked weak? You just threw away those two weeks of learning. Our advice: make major changes at most once every two weeks, and never change anything by more than 20% at a time.

6. "I'll switch it on and it'll just run"

48% of PPC professionals cite lack of control as their main frustration with PMax. And they're right about one thing — PMax really isn't autopilot. It requires weekly monitoring, a monthly creative refresh, and constant cleanup of search queries. Anyone who "sets PMax and forgets it" is paying for conversions they could have had more cheaply.

7. PMax as a replacement for Search campaigns

This is a fundamental misunderstanding. PMax complements Search — it never replaces it. Start with classic Search campaigns, build up a data foundation, and once you have 50+ conversions a month, add PMax as an extension. Not as a replacement.

How to Set Up PMax Correctly

Now for the practical part. Here's the process we recommend to the e-shops we work with.

Prepare your assets

The algorithm is only as good as the materials you feed it. The old saying "garbage in, garbage out" applies. Minimum vs. recommended number of assets:

Yes, you read that right. 15 headlines and 20 images. It's not a typo. The more quality assets you provide, the better the algorithm optimizes — it has more to choose from and more combinations to test.

Understand audience signals

This causes a lot of confusion. Audience signals in PMax are NOT targeting. They're hints for the algorithm about where to start looking for customers. The algorithm can — and will — ignore them if it finds a better audience elsewhere.

What should you set? Your own customer data (Customer Match lists) is the strongest signal. Next, custom segments based on keywords and competitor URLs. And interest and demographic groups as a supplement.

Set your bidding strategy

Start with "Maximize Conversion Value" WITHOUT a ROAS target. Let the algorithm gather data for 4–6 weeks. Once performance stabilizes, add a ROAS target — but set it realistically, based on what the campaign is actually achieving. Don't set a dream ROAS; set your current one + 10–15%.

Watch your search terms and negative keywords

Even with PMax, you need to review search queries regularly (ideally weekly) and add negative keywords. Remember — negative keywords only work on Search and Shopping, so Display and YouTube stay unfiltered.

TIP: Focus on n-grams. Find the two- and three-word combinations that keep appearing in low-quality queries and exclude them all at once. It saves you hours compared to weeding out individual queries.

PMax + AI: Why Managing Campaigns by Hand Makes No Sense in 2026

Here we come to a topic almost nobody writes about yet. PMax itself is powered by AI — that's its whole premise. But managing a PMax campaign? At most agencies and specialists in 2026, that's still done by hand. And that's a problem.

What eats up most of a specialist's time?

Account audits. Reviewing search queries. Checking assets. Generating new headlines and descriptions. Budget reallocation. Weekly reporting. This work is important, but it's routine and repetitive. And that's exactly the kind of work AI tools now handle.

What AI actually handles

Take a concrete example. A complete Google Ads account audit — checking for disapproved ads, missing extensions, campaigns with no conversions, keyword quality, search terms with high spend and zero conversions — takes an experienced specialist 4–6 hours. An AI tool like Claude Code does it in 15 minutes.

And it's not just audits. AI can:

• Automatically clean up search terms and suggest negative keywords sorted by priority.
• Generate 15 headlines and 5 descriptions perfectly matched to your brand tone — and secure an "Excellent" Ad Strength.
• Analyze your budget and suggest where to move money from high-CPA campaigns into ones that perform better.
• Write custom scripts for automatic alerts — for instance, when daily spend exceeds 150% of the average, you get an email before it can do any damage.
• Create a weekly report with comparisons against the previous week and month, anomaly detection, and concrete action items.

The cost argument

Paid optimization tools like Optmyzr, Opteo or Adalysis cost $500 to $2,000 a month. And they offer off-the-shelf features that don't always fit your account. AI solutions built on tools like Claude Code cost a fraction of that and are tailored precisely to your needs.

That doesn't mean a PPC specialist is redundant — quite the opposite. AI frees them from the routine so they can focus on strategy, creative, and decisions that require human judgment. But hand on heart — if your specialist spends hours cleaning up search terms and copying data into reports, you're paying for work a machine does faster and more accurately.

At LK Media, we moved to an AI-first approach back in 2025. Managing campaigns that used to take a whole day now takes hours. And we devote the rest of the time to what AI can't do — thinking strategically about the client's business.

When Not to Use PMax

We won't tell you PMax is for everyone. It isn't.

Service businesses with a handful of leads a month. If you run a law firm generating 5 inquiries a month, PMax doesn't have enough data to learn. Stick with Search campaigns.

A budget under 20,000 CZK (≈ €800) a month. With a low budget and few conversions, PMax is an expensive lottery. Better to invest in well-configured Search campaigns with precise targeting.

B2B with a long sales cycle. If six months pass between the first click and signing the contract, PMax optimizes toward the wrong signal. You need offline conversions linked to your CRM, which is a technically demanding setup.

A new e-shop with no data. No history, no conversions, no Customer Match list. Start with Search and Shopping, gather data, and add PMax as the next step.

What instead of PMax? Classic Search campaigns with precise keywords, Standard Shopping campaigns with a quality feed, and possibly Demand Gen for building awareness. This trio reliably keeps you in front of customers, even without PMax.

Want help setting up or optimizing your PMax campaigns? Get in touch at info@lkmedia.cz — we'll gladly take a look at your account and tell you honestly whether PMax makes sense in your situation.