Solved: Performance Max Campaign Setup for E-commerce

Hi there,

Thanks for reaching out!

I've had a good look through your situation with your PMax campaigns. It's a really common problem, especially for e-commerce stores new to the platform. People assume that because you can segment things, you *should*, but that's often the quickest way to burn through your budget with poor results. It sounds like you're falling into that exact trap.

I'm happy to give you some initial thoughts and a bit of guidance. The short answer is you're probably focusing on the wrong thing. Campaign structure is only a tiny part of the puzzle; the real issues likely lie in how you're feeding the algorithm and what's happening after someone actually clicks your ad. Let's get into it.

We'll need to look at how PMax really works...

Alright, so the first and most important myth to bust is this idea that granular control is the key to Performance Max. It isn't. In fact, it's the opposite. Traditional search campaigns rewarded advertisers for meticulous segmentation of keywords and ad groups. PMax is a completely different beast. It's a black-box algorithm designed to automate bidding and targeting across all of Google's channels (Search, Display, YouTube, Discover, Gmail, Maps). Its fuel is data, and lots of it.

When you take your budget of, say, €2,500 a month (around €80 a day) and split it across multiple campaigns, and then further split *that* across multiple asset groups, you're starving each part of the system. Each asset group is trying to learn independently, but it's only getting a tiny trickle of data. Imagine trying to train five different athletes for a marathon, but you only have enough food for one. None of them are going to perform well. That's what you're doing to your PMax campaigns.

You mentioned the more detailed setup "performed slightly better" but ran for longer and spent more. This is a classic cognitive trap. It's highly likely that the "better performance" was just statistical noise, or the algorithm finally scraped together enough data over that longer period to get a couple of conversions. It wasn't because the structure was superior; it was in spite of it. The increased ad spend just gave a flawed setup more runway to eventually get lucky. You're effectively paying more for the same, or worse, results. You're trying to force an old logic onto a new system, and the system is pushing back with high CPAs.

The core principle you need to embrace is data consolidation. PMax needs a large, unified pool of data about who is converting for *any* of your products. It uses this to build a powerful composite picture of your ideal customer. When you segment by product category—"hair products" in one asset group, "liquid soap" in another—you're telling the algorithm "don't let what you learn about soap buyers inform who you show hair product ads to". But what if your best customer is someone who buys both? Or someone who starts with soap and comes back a month later for a hair mask? You've just broken that learning path. You have to trust the machine to find these patterns, and for that, it needs all the data in one place.

I'd say you need to consolidate, not separate...

So, the immediate, actionable advice is to pause all your current campaigns and rebuild with a consolidation mindset. For a store with your budget and product count, here's what I'd recomend:

One Campaign, One or Two Asset Groups.

Yes, just one. This will be your "All Products" campaign. Pool your entire budget here. This gives the algorithm the maximum possible daily spend and data velocity to exit the learning phase quickly and start optimising effectively. You're not telling it where to spend the money; you're giving it the resources to figure that out for itself, faster and more efficiently than you ever could.

The only reason to consider a second asset group within this single campaign would be if you have a collection of products with a vastly different price point or margin that requires a different creative message. For example, your high-value "packages" could potentially sit in a seperate asset group if you want to test messaging around "value bundles" versus individual product messaging. But even then, I'd start with one asset group first, let it run for a month, and only segment if the data clearly shows your bundle products aren't getting any traction at all. Dont overcomplicate from the start.

Think about the data flow. Right now, your setup is inefficient, creating data silos that prevent the algorithm from building a holistic understanding of your customer. A consolidated structure creates a powerful feedback loop.

You probably should stop worrying about structure and start worrying about the funnel...

Now for the brutally honest part. Even with the perfect campaign structure, you will fail if your customer journey is broken. You said your ad strength is "very good" and you have "very good images and videos". This is a classic trap for marketers. Google's "Ad Strength" metric is a vanity metric; it primarily measures whether you've ticked all of Google's boxes by providing a variety of assets. It has almost zero correlation with actual conversion performance. An "Excellent" ad can have a terrible ROAS, while a "Poor" one can be your most profitable. Ignore it.

Your problem isn't the ad, it's what happens *after* the click. Your poor CPA and ROAS are symptoms of a leaky bucket. You're pouring expensive traffic from Google into a website that isn't converting it effectively. You need to become an expert diagnostician of your own sales funnel. Let's break it down:

-> Clicks but low Product Page Views? If people are clicking your ad but not bothering to look at individual products, it means you're attracting the wrong kind of traffic. Your ad might be too generic, or your audience signals are off (we'll get to that). Or, your homepage/category page is confusing, cluttered, or slow to load, and they're leaving in frustration before they even see a product.

-> Lots of Product Page Views but no 'Adds to Cart'? This is the most common and most critical failure point. If people are getting to the product page but not taking the next step, something is seriously wrong *on that page*. This is where your "good" images and videos are failing you.

• Product Photos: Are they professional? Do they show the product in use? Do they convey the "natural" and "skincare" value proposition? Or are they just sterile product-on-white-background shots? For a brand like yours, lifestyle imagery is everything.
• Product Descriptions: Is there any? Do they sell a benefit or just list ingredients? You're not selling soap; you're selling smoother skin, a moment of self-care, confidence. The copy needs to evoke that feeling.
• Pricing & Shipping: Is your pricing competitive? Is shipping cost a surprise at checkout? High, unexpected shipping costs are the number one cause of cart abandonment. Be transparent upfront.
• Trust: Your store is small. Why should anyone trust you with their credit card details? You need social proof. Customer reviews (with photos!), trust badges (secure payment logos), a clear returns policy, and an 'About Us' story are non-negotiable. Without them, you look untrustworthy, and people will not buy.

-> Lots of 'Adds to Cart' but few Purchases? This usually points to friction in the checkout process. Is it too long? Do you force account creation? Are there unexpected fees? Streamline this process relentlessly. Make it as easy as possible for them to give you money.

This is a pattern I see constantly. I remember one e-commerce client who had this exact problem. They were getting plenty of traffic from their ads, but the ROAS was terrible. We audited their site and found the issue wasn't the ads at all—it was the product pages. The product photos were low quality and they had almost no customer reviews. After they invested in a professional photoshoot and implemented a system to gather reviews, their on-site conversion rate doubled. This made their ad campaigns profitable almost overnight, without us needing to make major changes to the ad account itself. This goes to show that often, a poor ROAS isn't an ads problem; it's a business problem.

Use this calculator to diagnose your own funnel. Be honest with the numbers. The stage with the biggest percentage drop is where you should be focusing all your attention.

You'll need to feed the algorithm better signals...

Once you've fixed your leaky funnel and consolidated your campaign structure, the final piece of the puzzle is the quality of the data you're feeding the algorithm. PMax relies heavily on 'Audience Signals' to guide its initial targeting. This is your chance to tell Google, "Hey, don't just go after anyone interested in 'skincare'. Go after people who look like *this*."

Too many advertisers just throw in a few broad in-market or affinity audiences and hope for the best. This is lazy and ineffective. You need to provide the highest quality signals possible, and there is a clear hierarchy of what works best.

1. Your Own Data (The Gold Standard): This is non-negotiable. You must upload your customer lists to Google Ads.

• -> All Past Purchasers: This is your most valuable asset. Google will analyze these users to find thousands of similar people.
• -> High LTV Customers: If you can, segment your customer list and upload a list of your best, most frequent buyers. This tells Google to find more whales, not just minnows.
• -> Remarketing Lists: Create audiences for 'All Website Visitors', 'Product Viewers', and 'Cart Abandoners'. These signals tell PMax to go back and target warm leads who have already shown interest.

2. Custom Segments (High Intent): This is your second-best tool. Instead of using Google's pre-made audiences, build your own. Create a custom segment based on people who have recently searched for specific, high-intent keywords on Google. Don't use broad terms like "natural soap". Use buying-intent terms like:

• -> "buy organic shea butter soap"
• -> "best natural shampoo for dry hair"
• -> Your competitors' brand names
• -> Specific product names or types you sell

3. In-Market & Affinity Audiences (The Last Resort): These are the broadest and least effective signals. Use them only to supplement the higher-quality signals above. They can help the algorithm explore, but they should never be your primary targeting input.

Your asset group's signals should be layered like a pyramid, with the most valuable data forming the foundation.

And let's talk about realistic numbers...

Finally, we need to manage expectations about your CPA and ROAS. You say you're not satisfied, but what are you benchmarking against? E-commerce is a tough game, and what seems like a "high" CPA might actually be perfectly acceptable, or even good, depending on your product margins and customer lifetime value (LTV).

Based on my experience with many e-commerce clients, particularly in developed countries like those in Europe, here are some realistic ballpark figures. For a typical online store, a click might cost anywhere from €0.50 to €1.50. A healthy website conversion rate is between 2-5%. Let's do the maths:

Worst Case Scenario: High CPC (€1.50) / Low Conversion Rate (2%) = €75 Cost Per Acquisition (CPA)

Best Case Scenario: Low CPC (€0.50) / High Conversion Rate (5%) = €10 Cost Per Acquisition (CPA)

Your actual CPA will likely fall somewhere in that €10-€75 range. If you're currently in the upper end of that, it confirms that the problem is your website conversion rate. If you're already in the lower end and are still not satisfied, then your expectations might be unrealistic for your market.

The metric that matters more than CPA is Return On Ad Spend (ROAS). A common target for e-commerce is a 3x to 4x ROAS, meaning for every €1 you spend on ads, you get €3-€4 back in revenue. This target, however, depends entirely on your profit margins. If you have a 70% margin, a 2x ROAS might be hugely profitable. If you have a 25% margin, you'd need a 4x ROAS just to break even.

Use this calculator to work backwards. Set your average order value and your target ROAS to see what your maximum affordable CPA is. This number should become your new benchmark for success.

I've detailed my main recommendations for you below in a table to give you a clear action plan. This is the main advice I have for you:

As you can probably tell, getting Performance Max to work properly isn't just a case of setting up a campaign and hoping for the best. It's a complex interplay between campaign structure, audience signals, creative assets, and—most importantly—your website's own ability to convert visitors. Misinterpreting the data, like thinking a more expensive, segmented campaign is "better," is an easy and costly mistake to make.

This is where expert help can make a huge difference. An experienced eye can quickly diagnose these kinds of funnel issues and interpret the data correctly to build a strategy that actually works.

If you'd like to go through your specific account and website together, we offer a completely free, no-obligation strategy session where we can do just that. It might be helpful to have a second opinion before you spend another euro on a strategy that's not delivering.

Hope this helps!

Regards,

Team @ Lukas Holschuh