Response attribution, while challenging, provides tremendous benefits. Not only does it provide full visibility into channel performance, it may be used as an input to optimize long term media spend as well.
The key benefit for cross channel attribution is to provide visibility into which marketing channels are driving sales well and which are performing poorly. This will allow you to adjust your marketing tactics accordingly and track performance over time.
In many environments several challenges exist when implementing a cross-channel response attribution model. They include:
- The process of attributing a ‘sale’ to one or more channels is challenging from both a data and analytic perspective. It is best to start with bringing your data house in order before implementing the attribution model(s).
- The modeling techniques can be complex and difficult to communicate. All teams (business, database, and analytics) need to be a part of the process of choosing the proper analytic techniques. Analytics should drive this process but gain feedback and buy-in from all teams at every step of development.
- Implementation of the response attribution process typically requires a database, a data transformation layer, and an analytic engine. Be sure to have a vision on how the implementation should occur before starting development. That vision will help identify complexities during the implementation that may be minimized during development.
- Once you have your response attribution house in order, then it may be used to help optimize media spend over time, through the use of media mix models. When mapping out the response attribution process spend time to understand how it will be used within media mix models as that will enable a more seamless process from response attribution to media mix.
Cross-channel response attribution can be difficult, but it is worth the effort. Think agile, start simple, gain buy-in and momentum, and enhance over time. It is human nature to take an all-in approach and build out the ‘perfect’ solution right away. I would argue that approach, in many environments, is fraught with pitfalls given the challenges that exist in cross-channel response attribution. Build out a roadmap with milestones and quick wins and a solution built out in phases so you can learn and enhance over time while providing value quickly.
