Conversion Modeling 8211 Filling Data Gaps In The Customer Journey
Management Summary
Why do we need conversion modeling?
Recent changes in data collection and storage have impacted traditional methods of campaign analysis. Consent management, Privacy Shield, browser updates that prevent the use of third-party cookies tear holes in the measurement and analysis of online conversions. Cross device measurements are also made more difficult by the lack of cookies.
Conversion modeling now helps to fill these data gaps by reconstructing general campaign performance using machine learning technology.
For example, in DV360, unattributable clicks and impressions that likely caused conversions – but were not measurable – are attributed using machine learning. Currently observable signals such as: B. device, date and time as well as conversion type and model across all active campaigns are monitored and customer behavior is presented precisely and aggregated.
Conversion modeling requires a strong database
Conversion modeling depends on a strong base of conversion data. To maximize this data, advertisers should ensure the following actions as a first step:
- Site-wide integration of a tagging solution for first party data such as global site tag or Google Tag Manager
- Activate Enhanced Attribution to ensure maximum collection of click-through conversion data.
Activate Modeled Conversions in Campaign Manager
Conversion modeling can be found in the attribution category. For DV360 accounts connected to the Campaign Manager, the attribution models can be found under the Floodlight configuration in the Campaign Manager.
For DV360 accounts that are not connected to Campaign Manager, the attribution models can be found in the Floodlight Groups section of DV360.
Why you should start today!
The effectiveness of your marketing investments can only be effectively assessed if the customer journey is presented in detail and without gaps. Without conversion modeling, data gaps cannot be filled and essential questions cannot be answered based on data!
