How to Interpret Performance Max Asset Labels for Google Ads

Google assigns labels based on how specific assets contribute to your conversion goals compared to the rest of the assets in your account. A label of Best indicates that the asset is one of the top performers in your campaign. A Good label means the asset is performing well enough to stay in the mix. A Low label signifies that the asset is underperforming relative to others and likely requires immediate attention or replacement. These rankings are not based on a static score but fluctuate as the algorithm gathers more information about user interactions.

For example, in a typical account for an e-commerce brand selling home goods, you might have five different lifestyle images in one asset group. If three of those images are consistently labeled as Low, the system indicates that they are not driving the expected engagement or conversion rates when displayed. This does not mean these assets are inherently poor, but rather that they are failing to capture interest in the specific context of your campaign. By interpreting these labels accurately, you save time on manual testing and allow the system to favor higher-performing creative.

Assets labeled as Low are effectively consuming budget without providing the expected return. Because Performance Max campaigns rely on machine learning, they require high-quality inputs to function properly. When you leave underperforming assets in your campaign, you limit the potential of the ad combinations Google can generate. Replacing these assets with new, distinct variations provides the system with fresh data points, which can lead to better performance across your targeted audience segments over time.

In a typical engagement, we suggest a cycle of review. Once you identify Low assets, do not delete them immediately. Wait until the campaign has had enough time to accumulate significant data. A common pattern is to wait at least two weeks for a new asset to gather enough impressions for a reliable label. If the status remains Low after this period, you have sufficient evidence to swap the asset for something new. This keeps your creative library fresh and helps the system find the best possible combinations for your specific audience.

At Alien Road, we follow a structured approach to asset management. For every new Performance Max campaign, we ensure that we upload the maximum allowed number of assets across all categories. This gives the system more options to test from the start. We monitor the labels weekly, focusing on moving assets from Low to Good or Best. If an asset remains Low despite being visually distinct, we categorize it as a non-performer and remove it from the rotation to make room for new creative concepts.

In a typical client setup, we document which creative themes perform best. For instance, if video assets with a human narrator consistently earn a Best label, we prioritize creating more video content with that specific style. We treat the asset labels as a feedback loop rather than a final judgment. This process allows us to maintain a high standard of creative quality while ensuring the budget is always directed toward the most effective components of the campaign.

One major mistake is reacting to labels too quickly. If you change your assets every few days, you prevent the machine learning model from reaching a stable state. Another common mistake is ignoring the context of the asset. An asset might be labeled Low simply because it is only being shown in a specific placement where users are less likely to convert. Always ensure that your asset variety covers different angles, such as product-focused, lifestyle, and promotional messages, to avoid bias in your data.

While there is no single rule for the exact number of assets, Google recommends providing a wide variety to maximize the potential of your campaign. You should aim to fill as many slots as possible in your asset groups. A good practice is to keep at least five images, five headlines, and five descriptions, though more is generally better. The more inputs you provide, the more combinations the system can test. By maintaining this variety, you increase the likelihood of achieving Best labels for your most important messages.