eCPM implementation

The eCPM (effective CPM) is used to compare the performance of different type of campaigns (CPM, CPA, CPC, Sponsor). Everything is converted to CPM in order to compare campaigns and to order and visualize them according to their eCPM value: i.e.: eCPM is the way to express in terms of money, the relation among the actions carried out by a campaign (e.g.: between impressions and clicks) which allows to compare the performance of different type of campaigns (CPM, CPC, CPA, Sponsor) with the same weight.

For example:

Supposing we have a campaign that pays 0.10 dollars per click. After a week, it has registered 100,000 impressions and 2,000 clicks.

Therefore it has generated $200 (0.10 x 2,000).
After 100,000 impressions you get a eCPM of $200 /100,000 impressions = $200 / 100 CPM = $2.
(100 cpm results from diving 100,000 / 1,000 cpm = 100. Everything becomes CPM)

Consequently the cost of eCPM is of: $2.


  1. Click on "Global preferences" and select "Optimize by eCPM campaigns with priority 2 or 3 or 4, etc" according to your needs.

  2. Indicate the "Percentage of impressions used for research in eCPM optimization": The range of percentages is from 0% to 60%. The investigation percentage of 1% is default and can be modified at any time.

  3. The campaign can be created under, CPM, CPC, CPA, or Sponsor.
    NOTE: The type of campaign option (CPM / CPC / CPA / Sponsor) has no influence on the calculation of payments or eCPM.The only thing that influences is when the campaign ends.

  4. "Priority": Select the same priority that was selected in the "Global Preferences".

  5. Complete the information of the campaign price (mandatory)

The system will calculate the eCPM of each campaign and will display the campaign with the higher eCPM more frequently.

The eCPM is calculated by Campaign and Space and Campaign, Banner, and Space, for any campaign that has a specified price.

In version 4.3.1, there was a significant improvement by replacing the trial test period to a continuous test: All campaigns are now tested continuously rather than for an initial period. This adapts better to the changes in eCPMs relative to the campaign and achieve higher earnings.

The system now allows to divide all impressions between research and earnings. During the testing of new campaigns or old campaigns with the previous eCPM that were not as effective, but may be profitable later. This allows a substantial improvement in the functioning of the optimization.

Once you know the eCPM of all the campaigns uploaded, e-planning can easily sort them by first placing the highest to the least paying.

If there is a Space associated to several campaigns, sort by eCPM in a descending order with the most paid first. This will allow control of each campaign and will display the first that comply with all the filters and targeting variables. For example: if a Space has two campaign A and B with eCPMs of 0.89 and 0.88, and none have filters or targeting, and both have the same priority, it will always display campaign A, at least while eCPMs are maintained.

Calculating eCPM

The eCPM calculate each space that is associated to a campaign or ads. For example: A campaign may have an eCPM of $2, but in a given space can be $3 and in another $1. This is because the campaign may be more effective in one space than in another. When choosing which campaign and ad will display in a given space, the eCPM utilize will be for that particular space. No matter if the campaign has an eCPM of $100, if in a given space your eCPM is 1 cent, it is most likely that it will never display.

Once the campaign is selected to display, the ad must be chosen. The ads are also sorted by eCPM. However, e-planning does not display always the highest eCPM, but randomly chooses one with a higher eCPM ad preference. For example: if there are 2 ads with eCPMs of 2 and 1, for every two times 1 displays, it will display one time 2. In the case of ads, it also takes into account the weight. The weight does not matter if all have the same weight. But if an ad has more weight than another, it will have more chances to show.


   ©1998-2015 Caraytech S.A. All rights reserved.