Download edition

Example

To explain how the weights of banners affect banner selection, let’s assume the following sample scenario:

• There is a campaign with two banners of identical dimensions, both are 728 x 90 pixels. Initially, both banners have a weight of 1.
• There is a zone with dimensions 728 x 90 pixels.
• The sample campaign with the two banners is linked to the zone, and it is the only campaign linked to the zone.

Zone probability

In order to understand how this plays out, we’ll have to look at the “Probability” tab of the zone. Opening this tab will display a table that shows how the ad selection process will handle all the banners currently linked to the zone.

To open the Probability tab for a zone, navigate to the Inventory tab in your Revive Adserver account, and next click the Websites link in the left column. Finally, click the Zones link for the website you’re interested in.

In the Zones overview for the selected website, click the “Probability” link.

The Probability tab

The Probability tab of any zone shows a table of all the banners that are currently linked to the zone and that are currently active. There can be separate sections for “Override” campaigns, for “Contract” campaigns, and for “Remnant” campaigns. In this article’s example, there is only a single remnant campaign available. This is what Probability for the sample zone looks like:

Understanding Probability

Under this sample scenario, each of the banners has a probability of 50%. When a visitor opens a web page with this zone on it, this triggers a request to this particular zone, and the ad selection process in Revive Adserver will use the probability percentages to decide which banner to select. Let’s assume that during a period of time, a total of 1000 ad requests gets processed. The outcome will be that banner A will be selected roughly 500 times, and banner B will be selected roughly 500 times as well.

Having 2 banners each with 50% probability does NOT mean that a single visitor will see the two banners alternating perfectly as they refresh the webpage a few times. All it means is that each banner has a 50% probability of being selected for delivery.

Adjusting banner weight affects probability

Now let’s assume that there is a need to display one of the two banners more frequently than the other. This can be done by changing the weight of that banner. For this next example, the weight of banner A has been left at 1, and the weight of banner B has been changed from 1 to 2.

The effect is that banner B now has a probability of 66,67%, and banner A has a probability of 33,33%. This means that for every 1000 requests to the zone, banner B will be selected roughly 667 times, and banner B will be selected roughly 333 times.

The probability percentages are calculated by summing the weights of the two banners: 1 + 2 = 3.

Banner A gets 1 / 3 of that total, thus 33,33% and banner B gets 2 / 3 of the total, thus 66,67%.

Complexity

A real ad server account will – of course – be much more complex than this very simple example. There will almost certainly be many more campaigns, each with many more banners in them of all sorts of sizes, and there will also be many more zones. To add to the complexity, campaigns also have weights.

Understanding the reason and functionality of banner weights is simply the first level in a multi-level calculation model that ultimately determines the probability percentage of each banner in each zone.

Maximum banner weight

The maximum weight for a banner is 127. If you attempt to enter a higher number (for example 200), and save the changes, it will simply be adjusted down to the maximum allowed value of 127.