Exactly that happened when the “netzstrategen” switched a shop's Google Shopping setup to the Labelizer from Label Up. Within just a few weeks, a broad, difficult-to-control product portfolio transformed into a clearly structured performance system that Google immediately understood.

The result:

57% less cost.
4% increase in sales despite negative market trends.
And a setup that works almost ridiculously easily.

The key was not a complex AI, no massive overhaul, and no weeks of analysis. But a radically clear segmentation based on the only data points that really matter:

Clicks. Costs. Conversions.

With four labels, a clear prioritization, and a bold approach, netzstrategen created a setup that promotes the best products and mitigates the weakest. And Google responded immediately.

The solution: A radically clear Labelizer strategy

Instead of using categories or brands, Stephan Sperling, Google strategist at netzstrategen, opted for an uncompromising performance-based segmentation. These segments were created exactly as such in Label Up and played out clearly separated in Performance Max. The four label groups that made the breakthrough:

1. Super Champion (Conversions ≥ 1 & ROAS over 700 percent)

These are the absolute top products. They are purchased and fall within the clear profit area of the shop.
Google should always maximize exposure for these products.

Effect:
More visibility for the most profitable items that Google would not have clearly preferred in a mixed setup.

2. Champion (Conversions ≥ 1 & ROAS under 700 percent)

Products that clearly perform but are not yet in the “elite area.” Perfect for controlled growth.

Effect:
Budget is intentionally allocated to products that have conversion potential and should scale.

3. Waster (Over 50 clicks & 0 conversions)

The toughest and most important group.
Products that burn money but do not yet bring anything back.

Effect:
Google receives a clear signal: These products should receive less budget. Costs decrease immediately. This is the direct source of the 57% cost reduction.

4. Loser (Under 50 clicks & 0 conversions)

The broad “neutral field.”
Products that Google is still testing and that perform neither well nor poorly.

Effect:
PMax can test freely here, but without budget pressure. At the same time, transparency massively increases.

How Stephan used his structure with custom labels directly in PMax

Another crucial factor in the successful Labelizer strategy was the simple mirroring of the labels from Label Up to Google Ads.
Stephan assigned the four segments Super Champion, Champion, Waster, and Loser directly in the feed as custom_label. This allowed him to rebuild his entire PMax structure with just a few clicks. 

This is how the restructuring worked in practice

Once the labels arrive in the Merchant Center, they can be used in PMax as follows:

1. Filter products by custom_label in Google Ads
   In the Performance Max campaign, a product filter is set by Listing Group:
   »custom_label0 = SuperChampion«
   and so on.
   
   

2. Distribute budget and priority
   This allows the segments to be managed with different budgets, objectives, and assets.
   
   

3. Receive clearly separated performance data
   Each Listing Group shows transparently:
   • how much budget each segment consumes
   • how many clicks are generated
   • which segments deliver revenue
   • where costs escalate

With just a few filters, a performance-based campaign system is created that remains easily manageable in Google.

How the low performers were actively improved

After the restructuring, Stephan went one step further:
He deliberately set his Waster and Loser campaigns without target ROAS to the Google bidding strategy "maximize conversions".

This ensured that:

• Google could freely test whether products with weak history would still generate conversions
• individual items from these segments could rise to the Champions
• the low-tier segments did not burn unnecessary budget despite tests
• the overall system became dynamic: products rise or fall based on real performance

Stephan built a setup that is not only structured but self-optimizing.