Target CPM provides an alternative way of setting a floor price that allows for increased fill rate and yield, while still maintaining an average minimum price for your inventory. Leveraging Google’s machine learning expertise, target CPM dynamically adjusts the floor price on matching inventory to maximize yield.
The target CPM feature is optional and available for unified and First Look pricing rules. You have complete control over whether to use target CPM. You have the flexibility to specify whether each individual rule is a floor or a target CPM.
How target CPM works
- Ad Manager dynamically sets the floor prices on individual bid requests; they can be either higher or lower than the target CPM you indicate.
- Ad Manager’s goal is to have the average CPM for your inventory be equal to or greater than the target CPM.
- There are a small number of cases when Ad Manager can’t match the target CPM, such as when:
- The rule is new and/or doesn’t have enough data yet for full optimization.
- There is significant fluctuation in the inventory traffic.
- For unified pricing rules with target CPM enabled:
- The dynamic allocation temporary CPM is still respected, ensuring publishers optimize overall yield—not just indirect yield at the expense of guaranteed demand sources.
- The floor is dynamically adjusted to allow all your non-guaranteed demand sources like Authorized Buyers, Exchanges via Open Bidding, and remnant line items to win more auctions, increasing your fill rate.
Query 1: Bid from Buyer A @$2.10, Bid from Buyer B@$2.05
Query 2: Bid from Buyer A @$1.90, Bid from Buyer B@$1.80
Scenario A (Floor @$2.0 for Buyer A and Buyer B)
- Only Query 1 (at transaction price $2.10) is filled
- eCPM: $2.10
- Revenue: $2.10
Scenario B (Target CPM @$2.0 for Buyer A and Buyer B)
- Query 1 (at transaction price $2.10) and Query 2 (at transaction price $1.90) are filled
- eCPM: $2.00 (-5%)
- Revenue: $4.0 (+95%)
You can report on the Ad Exchange revenue uplift generated by target CPMs with the “Optimization type (Beta)” dimension in historical reports. You’ll see a breakdown of data by target CPM.
You can run an experiment to measure the uplift from applying the target CPM to a rule.