The Poor Performer exclusion section examines how excluding underperforming segments from ads impacts ROAS. 📈 This can help you refine your exclusion strategy, deciding products to leave out for better results.
Key points to keep in mind
Excluding Poor Performers can improve ROAS—if the segment is set up and applied correctly.
Aim to exclude products responsible for around 15% of the worst-performing spend for best results.
Too little exclusion won’t make a difference, too much can hurt performance.
To see real impact, the segment needs to be excluded from enough high-spending product sets.
If excluding Poor Performers doesn’t improve ROAS, it’s often due to segment setup, product set logic, or overlapping filters. (see below)
How to analyze and exclude poor performers?
1. Select the segment 🫵
You may have multiple poor performers for various campaigns and product types - choose the primary segment for analysis.
📝 Note: After selecting a segment, you can filter the view by product sets. This is useful if you want to exclude data from product sets that aren’t relevant—like always-on or seasonal sets, or those used in traffic campaigns where Poor Performers are usually not excluded.
2. Look at how much of the total spend comes from this segment 💸
Ideal Range: Aim for around 15%.
Too Low (<10%): Segment may be empty and not working effectively.
Too High (>40%): Segment may be excluding too much and could be counterproductive.
3. See how much spend you are currently excluding with this segment 📊
Check if the segment was excluded from enough spending product sets to have a meaningful impact.
Tip: A minimum of 30% exclusion rate is recommended for significant overall effect. Everything above 70% is great, but not necessary.
Example: An 11% exclusion rate may only affect a small portion of spending and have limited overall impact.
Positive case:
Negative case:
4. Check how ROAS changes when poor performers are excluded vs included 💡
See the ROAS difference between products that exclude the poor performers segment and those that don't.
Example: If ROAS is better when certain products are left out, you might want to exclude low-performing products from more product sets to improve results.
5. Examine if poor performers are excluded from all key product sets 🔍
Check the spend from poor performers within each product set.
Identify where you're losing potential revenue due to not excluding the segment.
📝 Note: It is reasonable that Poor performers are not excluded from 100% of spend. You may have seasonal products or specific categories you want to sell regardless of performance or with different performance criteria altogether. For these, a different, more specific Poor Performer segment might be created or no exclusion is applied at all.
6. See how removing Poor performers affects the results 📈
When poor performers are set correctly and excluded from a significant amount of spend, the spend share of the poor performers segments (the red line) typically goes down while the overall ROAS (light blue line) usually goes up.
What do to if excluding Poor performers didn’t improve ROAS?
What if the segment that kept the poor performers still has a higher ROAS than the one that excluded them. 🤷♀️ Excluding poor performers should boost ROAS, so something must be off. Let’s figure out the problem.
The segment is too small
🤔 Problem: The segment is not big enough (less than 10% of total spend). It may be empty and not working effectively.
💡 Solution: One option is to set up a lower threshold for the minimum spend to help filter out more Poor performers. But be careful not to exclude too many products.
Excluding Poor performers along with Deadstock
🤔 Problem: If deadstock is defined too strictly, you might filter out too many products. (e.g. Deadstock = all products with no revenue, older than 30 days.)
💁♀️ Example: High percentage of products excluded as deadstock (e.g., 92%). This leaves little room for other products and can severely limit the campaign's potential.
💡 Solution: Consider adjusting the timeframe of the product age condition (e.g., to half a year) for a more meaningful exclusion.
Problem with AND/OR conditions
🤔 Problem: Using "OR" instead of "AND" between the conditions "IS NOT" will select everything and the exclusion is not applied.
💡 Solution: Ensure the correct logic is used in product set conditions. Break down complex exclusion groups and rebuild them with the correct "AND" logic.
📝 Note: Be aware that if your company operates in multiple markets with shared management, similar mistakes can occur across those markets. Examine the product set for these patterns to identify and rectify similar mistakes.
Not all product sets are comparable
🤔 Problem: You may be comparing data sets that might not be suitable for direct comparison such as those with different spend levels or product categories.
💁♀️ Example: The product set where the poor performers are not excluded has a better ROAS because it’s used in remarketing campaigns or because they carry products under heavy sale.
💡 Solution: Check out what product sets are actually making the difference and whether they are comparable. Use product set to filter out incomparable product sets, for example.
3 Key takeaways
The Poor performers exclusion page should clarify the majority of typical cases. It will help you reveal how the poor performance exclusion performs and provide signals and direction to understand possible issues.
Always consider the context of your data. Don't jump to conclusions based on high ROAS alone; investigate the underlying factors driving performance differences.
Go deep into the detail of the product set and explore it for potential issues. Be extremely careful with stock exclusions as they can significantly impact ROAS.