The Optimization Report provides a high-level view of how effectively the optimizer is improving campaign performance over time. Rather than showing granular breakdowns (e.g., by specific publisher or audience), this report focuses on how well the AI-powered model is working overall.
Performance Index
The Performance Index metric is a simple way to understand whether the optimizer is generating results more efficiently than standard campaign delivery. It looks at what percentage of the impressions served thanks to the optimizer resulted in conversions.
A score above 1.0 indicates stronger-than-expected performance, while a score below 1.0 indicates performance below expectations. For example, a Performance Index of 1.22x means that 22% of optimized impressions resulted in conversions.
Optimized Daily Win %
The report also shows the percentage of days where the optimizer outperformed standard campaign performance. This indicates how often the model is contributing positively to campaign results.
Optimized Conversions
This explains how many conversions were attributed to the optimizer.
Impressions Split Over Time
The report visualizes how impressions are distributed between optimizer-influenced campaigns and standard campaigns. This is shown as a time-series view (how allocation changes over time) and a cumulative view (overall split across the campaign).
You will notice that over time more and more optimizations will take place because the optimizer will have more data and get more confident, leading to better outcomes for your campaign.
Note: If your campaign doesn't generate many events, for example only a few website visits, the optimizer will not be able to optimize too much during your campaign.
Optimization Areas
The report highlights where the optimizer is focusing its efforts, including:
Geography (ZIP code)
Supply and deals
Day of week
Media and time of day
Each area is shown as a percentage, indicating where the model is making the most adjustments.
Note: Optimizations are generally around the media purchased; the optimizer will not make changes to audience targeting strategies or creative.
How to interpret:
Higher percentages indicate areas where the model sees stronger opportunities to improve performance
Lower percentages suggest less variation or fewer optimization opportunities
Daily Performance Index
This metric explains how the Performance Index trends daily.
How to interpret:
Optimized Wins — these are impressions won based on optimization. A higher share of optimizer-influenced impressions indicates a more aggressive optimization approach, while a lower share suggests a more conservative approach.
Baseline Wins — these are impressions won with baseline campaign delivery (not using optimization).
This helps you understand how actively the model is shaping your campaign results with a day-to-day view.
Key takeaways
The Optimization Report is designed to answer: Is the optimizer improving performance, and how actively is it doing so? It provides a clear, aggregated view of model effectiveness without requiring manual analysis of individual dimensions.
What you'll need to enable effective optimization
Campaign minimums
Spend: at least $500 per line item (lower spend may make optimization difficult)
Length: minimum 14 days (the longer the campaign duration the better the optimization results)
The optimizer uses a rolling learning window, and sufficient data volume is required to generate meaningful insights.
Considerations
Attribution pixel must be installed, tested and working before launch
Optimization impact may be limited early in the campaign while the model is learning
Highly restrictive set-ups (i.e. narrow targeting, limited networks) can reduce effectiveness
Learnings are specific to each line item and do not carry over
The optimizer only takes into account the last two weeks of the campaign