Optimization for Ad Manager is designed to help you deliver the highest-valued inventory by determining the best ads to serve. Looking at ad request parameters, historical delivery patterns and user signals, Ad Manager uses a predictive model to estimate the click-through rate (CTR) of an ad and compare it to delivery schedule needs and forecasting. After accounting for delivery goals for all ads eligible to serve the impression, Ad Manager uses the estimated CTR to determine the best ad to serve.
If no historical delivery data is available, your ads serve normally. Over time, as more data becomes available, you'll begin to notice the effects of optimization and results should improve.
User rules help to define and categorize users. Both performance and revenue optimization employ user rules to help predict CTR. These rules are created dynamically, and incorporate new data within a matter of hours. This requires no extra work from you, and there is no waiting period before implementation.
For example, you might set up a user rule to categorize users browsing with Chrome on Saturday in New York, or users viewing Outdoor Sports content in a leaderboard ad associated with
User rules support multiple signals, including:
- Inventory unit
- User features (geography, OS, browser, domain)
- Time and day
- AdSense contextual signals to optimize your ads—we use Google's web crawling infrastructure to analyze the content of your pages along with the content of the click-through URL you specified in Ad Manager.
Measuring the impact of optimization
Optimization uses a control group and an optimized group to measure the lift that you can achieve by optimizing your ad delivery. Users who visit your site are automatically placed into one of these two groups at random, ensuring that the only difference between them is whether they receive optimized ads. Typically, 80% of users will be in the optimized group, and 20% in the control group.
How often data is updated
eCPM is calculated for each impression subject to optimization.
The model used for predictions in optimization is updated every hour.
New data (i.e., new ad units) are incorporated into the model once per day.
Performance vs. Revenue optimization
These two optimization modules both utilize the same underlying ad server logic to achieve their ends, but each module optimizes a different type of ad according to different criteria, and each module produces a different set of reporting metrics.
The performance module works with absolute goal ads (CPM and CPC) to improve your clickthrough rates (CTR) or conversions for your premium advertisers by predicting which ads are most relevant to users.
Performance optimization is measured in terms of click lift percentage:
Click lift = (optimized CTR / control CTR - 1) * 100
Performance optimization applies to the following ad and cost types:
- Standard CPM and CPC
- Bulk CPM and CPC
The revenue module works with any type of CPC inventory (remnant or reserved) to improve your eCPM (effective CPM). Revenue optimization helps you deliver the same number of clicks from fewer impressions by finding users who click through more frequently, thus freeing up impressions to serve elsewhere in your network.
Revenue optimization is measured in terms of freed up impressions and eCPM lift percentage:
eCPM lift = (optimized eCPM / control eCPM - 1) * 100
Revenue optimization applies to the following ad and cost types:
- Standard CPC
- Bulk CPC
- Price priority CPC
How line item priority affects optimization
Revenue and performance optimization are applied within the context of of an ad’s priority. Typically, priority is treated in the same manner by our ad servers, whether or not optimization is turned on for a network. However, on rare occasions, optimization and other factors such as Ad Exchange revenue and dynamic allocation can outweigh the value of priority when determining which line items Ad Manager selects. Learn more about priority and the ad selection process.