The Model Insights Report highlights the signals used by our models to identify and target users likely to convert. These signals represent media consumption characteristics of users prior to conversion so that we can score our future targeting opportunities in real time.
Model Insights provide transparency into your converting audience’s behavior, and more specifically, how it changes. Our targeting capabilities are built on top of our real-time pulse of the Internet, and this report gives you visibility into how we adapt to live browsing signals so that the data can make the tactical decisions for you.
Once you have chosen the date ranges to compare, you can see the shifts in conversion affinity for specific domains, geos, and topics of content. Understand exactly what pieces of a customer journey, whether obvious or obscure, drive them to the outcome you want.
The term definitions below will help you interpret the data:
- Affinity: A user’s likelihood to convert, relative to the general Internet population. The higher the affinity number, the more likely the user will convert. For example, you’re looking at Geo features and see that San Francisco has an affinity of 10x, New York 0.5x, and Los Angeles 1x. This means a user browsing in San Francisco is 10x as likely to convert as an Internet user selected at random, whereas users in New York and Los Angeles are half as likely and equally likely to convert, respectively, as the general Internet population. Note: an affinity measurement is similar to an index, expressed differently. For example, an affinity of 1.5x is equal to an index of 150.
- Domains: Websites visited by users prior to conversion.
- Topics: Topics of websites visited by users prior to conversion.
- Geos: Locations from which users have browsed prior to conversion (based on the Internet service provider).
To export, select “Download” to the right of each category. The report will be downloaded to your desktop.