Use Adobe Campaign for Predictive Modeling and Propensity Scoring

Today demands on marketers for audience acquisition go beyond segmentation based on solely on behavior and intent. They are searching for net new prospects who might act in accordance to the marketing program goals. Tight marketing budgets and faster delivery to market, make it difficult for marketing departments to hire and build a Quants Team, to deliver attribution models, and lead scoring algorithms.

Adobe Marketing Cloud products provide attribution models based on industry best practices. Adobe Analytics has Predictive Marketing module that helps analyst do everything from anomaly detection, clustering audience and assigning propensity scores within Data Workbench. An audience that is tailored within Workbench can then be targeted using other Adobe Solutions including Target. More information here.

If you want to use your own ML engine and data that is already in Campaign or if you don’t have or don’t want to connect Analytics to Adobe Campaign, then there is an option within Adobe Campaign the Marketing Analytics module. Adobe Campaign is an automation tool that helps you target, segment, deliver communications to customers across multi-channel journeys and allows marketers to work on predictive audience information within Adobe Campaign using a simple template approach.

Follow these steps

To create predictive marketing models, go to the Profiles and Targets > Jobs > Predictive models node of the Adobe Campaign console tree.

Predictive Model

Predictive Model creation

In the Modeling tab, you can insert data to enable the template to predict customer and prospect behavior.

recipients

The algorithm lets you build a model to define a target value based on a set of variables. The engine uses the following analysis techniques: Weka and Kxen engines. Like any other campaign template, the predictive template works on target dimensions. Use the Add button to select each variable you want to insert into the model. These are the fields of the table which the analysis will concern

expression-select

Once the model is designed, you can see the prediction calculated by the model.                                             

predictive-model-graph

 

Once the Model template is in place, you can to use the predictive intelligence in  campaign workflows with an out of the box component called Predictive Split. The Predictive Split activity lets you split a defined population using a predictive model that we designed earlier.
predictive-split

The predictive workflow should look something like this:predictive-workflow

And after the Predictive Split, you can add any delivery that you want. Real world customer journeys are complex but predictive split can be helpful and simple way of upgrading your campaign workflows to have predictive intelligence.

0 comments