Viewability Prediction Dashboard

Browsi’s machine learning algorithm powers the prediction of potential ad unit spots on the page that will be viewed by a single user in real time - all before demand for the slot is fetched.

Browsi provides publishers the option of using predicted viewability by connecting the data directly to the ad server.  This metric can then help decide within the ad server which demand, if any, should fill new highly viewable placements and at what price, keeping your impression waste to minimum. 

The Viewability Prediction dashboard provides all the data needed in order to better understand inventory breakdown into those quality tiers.

Browsi has three (3) viewability prediction tiers:

• 0-30% Low viewability prediction

• 31%-69% medium viewability prediction

• 70%-100% high viewability prediction
  
  

You can now measure what’s the share of each tier among your inventory and compare its performance in terms of CTR, Fill Rate, eCPM.

Each impression from the inventory falls into one of the three tiers, providing insights for inventory allocation - Direct campaigns / Programmatic, according to the importance of the viewability rate in this setting.

Filters

Data can be filtered per Ad units, Devices, Line item types and time period.

Rate and Distribution

In the following tile you will be presented with an overall viewability rate from GAM and an inventory distribution to the three viewability tiers.

Open

Performance per Viewabilty tier

In this tile you can select a metric and understand how it performs in each viewability prediction tier.   

The bar's color reflects the volume of impressions for that tier.

Open

The data can be viewed in a table or a bar chart.
In the table mode you can see all data for all viewability prediction tiers, while its coloring assists you with quickly spotting the number of performance tiers (usually 2-3) and their ranges.

This data would assist you with determining the optimal Direct/Programmatic pricing.

Instead of the traditional pricing setting in GAM, of using the same pricing globally for a specific ad unit, you now have the ability to understand its performance granularity and calculate its eCPM per each viewability prediction range.

 Line Item type viewability prediction share

This graph can assist you with spotting wrong allocation of high quality inventory to Line items which are less likely to pay for the high viewability.

Foe example, “House” Line item type is getting a large share of your high viewability inventory?

Consider targeting the campaigns under that type specifically to the lower viewability inventory.

Viewability prediction Share over time

This graph demonstrates the distribution of viewability prediction tier over your inventory and the shift between them over time. After utilizing Browsi features which have impact on viewability (Like refresh, lazy loading, setting higher viewability threshold, etc) you should expect to see the increase of the higher tier over the medium and lower ones.