Performance Max is in an indeed powerful campaign type. But if left unstructured, it can very easily go out of your control and become unpredictable.
There’s no doubt that pMax can scale fast, discover new converting audiences, and stitch together signals you didn’t even know existed.
But then, there are challenges, as well.
It works best when you feed it the right structure, something most advertisers overlook.
I’ve seen this across dozens of accounts – a single pMax campaign running the entire catalog with a broad ROAS target. Or multiple pMax campaigns based on product category segments.
What ends up happening in these strategies is that Google pushes the easiest-to-convert products, not necessarily the products which are most profitable for you or your clients. A good number of high-ROAS products get buried and the low-ROAS ones hog the spend.
That’s where the ROAS Segments Framework for Performance Max comes in.
Instead of focusing only on the products that easily sell, this strategy aligns your advertising budgets to your business’s bottom line.
It does not treat your catalog as one big block, rather it breaks things down based on ROAS expectations. Your budgets align better with your business priorities, not just with clicks or conversions probabilities.
If you’re already running pMax and want:
- Better control over what gets promoted,
- Cleaner data and faster learnings,
- And, a structure that scales profitably, not just quickly,
Then this framework will help.
I worked this out to solve real issues for one of my clients and then validated it with dozens of different accounts.
The first case study that I published about this ROAS Segment Framework for Performance Max can be read in Google Ad’s support section.
The Problem with Bundled Campaigns
The default approach to Performance Max is simple:
Put all your products into a single campaign. Let Google figure it out.
Or have separate campaigns for different categories and assign budgets based on your priorities.
And to be fair, this does work—if your entire catalog behaves the same way, has similar margins, and you’re okay with sacrificing visibility for convenience.
But in the real world, that’s not always the case.
Most businesses have a mixed bag:
- High-performing products with excellent margins,
- Low-ticket items that convert fast but contribute little,
- And new launches that need testing but struggle for attention.
When you put all of them into one campaign:
Google prioritizes what’s easiest to convert, not what’s best for your bottom line.
High-ROAS products often don’t get enough budget.
You lose clarity, because there’s no way to isolate performance by intent or value tier.
What looks like ‘automation at scale’ often turns into missed opportunities and inefficient spend.
This isn’t a pMax issue. It’s a structure issue.
Which is exactly what the ROAS Segments Framework is built to fix.
The Core Idea: Segment by ROAS, Not Product Type
Forget product categories. Forget brand vs. generic.
When it comes to structuring pMax, the segmentation that aligns with profitability is ROAS.
That’s the heart of this framework:
You split your campaigns based on how much return each product group can realistically generate, not what type of product it is.
Why does this matter?
Because in every catalog, some products consistently bring in 600–800% ROAS. Others hover around 200–300%. A few might just break even but help move inventory or acquire new customers.
When you group them all together, Google doesn’t care which is more profitable. It optimizes for ease of conversion.
But when you break them into separate campaigns:
- You control how much budget each ROAS tier gets.
- You set different bidding strategies and ROAS targets per campaign.
- And most importantly, you start scaling what actually deserves scaling.
Here’s a simple mental model:
- High ROAS Campaign → Profit maximization
- Mid ROAS Campaign → Stable performers
- Low ROAS Campaign → Test, learn, and control loss
This structure gives Google clear signals. But more importantly, it gives you clarity, control, and the ability to make smarter budget decisions.
Step-by-Step Implementation
Setting up the ROAS Segments Framework isn’t complex, but it does need upfront thinking.
The good news? Once done, it makes scaling and optimizing much easier.
Here’s how to build it from scratch:
Analyze Product-Level ROAS
Start with your historical data.
Export product-level performance from Google Ads or your backend analytics. I usually prefer creating a dashboard inside Google Ads itself, as when needed you can export CSV or Google Sheet to build your Asset Groups.
Whatever you choose, have the following:
- ROAS per SKU
- Volume of conversions
- Profit margins (if available)
Then, bucket them.
Something like:
- High ROAS: Above 600%
- Mid ROAS: 300–600%
- Low ROAS: Below 200%
You can customize these bands based on your business goals. The point is: every product should fall into one clear segment based on return, not gut feeling.
Tag Products in Your Feed
Once you have the buckets, apply them in your product feed.
Use Custom Labels or a supplemental feed to tag each product with its ROAS tier:
- custom_label_0 = high_roas
- custom_label_0 = mid_roas
- custom_label_0 = low_roas
If your inventory changes often, set up a Google Sheet or automation to update labels dynamically based on fresh performance data.
Create Separate pMax Campaigns
Now, set up one Performance Max campaign per ROAS tier:
- Campaign A → High ROAS
- Campaign B → Mid ROAS
- Campaign C → Low ROAS
Each campaign should only target products from its segment, using the custom labels you defined.
This gives you total control over budget distribution and bidding strategy.
Set Smart ROAS Targets
Apply different ROAS targets for each campaign.
Be realistic. Take the average ROAS of every segment, say in the last 90 days, and set tROAS about 20% below it.
Align Creative to Performance Tier
Even creative strategy should change by tier:
- High ROAS: Focus on proven angles, urgency, reviews.
- Mid ROAS: Mix of proven + test creative.
- Low ROAS: Heavier on storytelling, brand building, and hooks.
You can also start without adding any creative information with these campaigns. That way, you can start as feed-only shopping type campaigns. And, you can can add creative assets to them.
Remove Outliers
During analysis the products for the ROAS segments, you might see that certain products might have got a very small number of conversions, but the their revenue multiple will be much higher. If you can remove such outliers if you are operating with a large number of SKUs. I usually, filter out any which has got less than 5 conversions in the last 90 or 180 days.
So what happens to these outliers, when do you keep them?
You can keep them in a Catch All campaign, exclude them, or keep them in the brand-specific campaign that I am going to talk next.
Brand Exclusion
If your brand has a good volume of searches, you might see that your segments are going after brand searches. It’s because brand search usually have better conversion rates and higher ROAS. They are low hanging fruits. Note that it’s not specific to the ROAS Segments Framework for pMax, it can happen to any structure of segment you use.
The easy fix to this is that you create create a brand exclusion and add to all the pMax campaigns.
Now, create a new pMax campaign with all the products in your feed and give it a very high ROAS, usually about 10% higher than what your brand keywords are comfortably achieving. It will now fix the problem. Your ROAS fragments will be forced to look beyond brand searches, and the pMax that we created with the highest ROAS target will revolve around brand searches only.
You can have the control that you need.
Real-World Results
This framework wasn’t built in a spreadsheet. It was born in real accounts with real budgets, real products, and real problems.
Let me give you a sense of what happens when you switch from a bundled pMax setup to a segmented ROAS-based structure.
More Control Over Budget Allocation
In a D2C skincare brand account, the high-margin serums were getting about 15% of the budget, despite consistently delivering 600%+ ROAS. The reason was that the campaign was optimizing for Conversions, the other low-priced products converted well (however, with lower overall ROAS) and thus took most of the budget.
Post-segmentation, we gave those products their own campaign. Within 3 weeks, they scaled to 50% of spend and accounted for 80% of revenue. The client asked for the next two quarters projection and increased the allocated monthly budget to scale further.
Scalable Profitability
A fashion brand running 2,000+ SKUs moved to this framework.
We scaled their high-ROAS campaign by 2.7x in 6 weeks without compromising efficiency.
The difference isn’t just in the metrics, it’s in the decision-making.
You know where to push, where to hold, and where to fix. That’s what makes this framework stick.
Important Points To Understand
The ROAS Segments Framework works charm, but like anything that gives you more control, it asks for more responsibility. Here are a few things to be aware of before you roll it out:
- Requires Solid Product-Level Data
- This isn’t a “just try it and see” framework.
- You need at least 30–60 days of historical ROAS or contribution margin data to segment your products meaningfully. If that data’s missing or unreliable, your segmentation may misfire.
Ongoing Maintenance
Products don’t sit in one tier forever.
A mid-ROAS product can become high-performing post-launch or during a promo. You’ll need to review performance at intervals and shift SKUs between campaigns when needed.
It’s not set-and-forget.
Budget Distribution Needs Attention
If you’re not careful, your high-ROAS campaign might spend too little because Google sees “limited volume.”
You’ll need to manually nudge budget splits and monitor impression share, especially during the first few weeks.
Risk of Signal Fragmentation
Splitting campaigns means you split signals, audiences, conversions, etc.
So, if your site doesn’t get much volume, or you have a lean catalog, segmenting too early might work against you. In such cases, start with two segments, not three.
This isn’t to scare you off. It’s to set you up with realistic expectations.
When Not to Use This Framework
You Have a Small Catalog
If you’re selling fewer than 100 SKUs, you likely don’t have enough product variety to benefit from segmentation. One well-optimized pMax campaign with solid assets and a strong landing page is enough.
Or Category based segments will ideally be better for you.
You Don’t Have ROAS History
This framework is powered by data.
If you’ve just launched, or have too little sales volume to confidently bucket products, hold off. Instead, focus on building initial performance and trust signals first.
You Lack Feed Control
If your current eCommerce stack doesn’t let you customize product feeds (no access to custom labels, supplemental feeds, or dynamic updates), segmentation becomes messy.
Get your feed in order first.
Your Products Behave Uniformly
Some businesses genuinely have uniform margins and similar conversion behaviour across products. If splitting things out doesn’t give you materially different ROAS patterns, then keep it simple.
Bottom line: Use the ROAS Segments Framework only when there’s real performance asymmetry across your catalog and you’re ready to act on it.
Final Thoughts: Why ROAS Segments Framework for Performance Max Matters
Because, Performance Max is here to stay.
It’s smart, aggressive, and incredibly effective, if you set it up with intent.
But the moment you let automation run without structure, your budget starts making decisions you never signed off on. Products with low margins get priority. High-performing SKUs get buried. And optimization becomes guesswork.
The ROAS Segments Framework isn’t about outsmarting Google.
It’s about giving the algorithm better signals, and giving yourself the visibility and control you need to scale profitably.
Instead of treating your catalog like a monolith, you treat it like a portfolio, with clear expectations and budgets tied to outcomes.
That shift alone can change how you manage campaigns, allocate budget, and scale intelligently, without burning cash or chasing volume for the sake of it.
If you run pMax and want it to actually reflect business logic, not just conversion math, this framework will take you there.
As the next step of this Framework, you can start pushing your actual product level margin data in the feed and then fragment campaigns and set realistic ROAS. It will take you beyond ROAS and optimize for real profit margins. We are already testing it with a few accounts, and will be doing a post soon.
