際際滷

際際滷Share a Scribd company logo

The power of RapidMiner, showing the direct marketing demo

W
Wessel Luijben

Direct marketing allows businesses to target specific customer demographics like age, gender, and marital status. It has expanded beyond mail to include digital channels like text, email, and online ads. RapidMiner's direct marketing wizard helps businesses invest in the highest converting marketing actions and reduce costs through improved targeting. It provides a table of top customers to target, shows which customer properties most impact response rates, and evaluates the predictive model to determine if more customer data is needed.

1 of 27
Direct Marketing
Jeffrey van der Eijk Founding Partner, Responsible for data driven innovations Economics, University of Amsterdam More than 10 years of QlikView, Business Intelligence, business innovations experience Jeffrey@Xomnia.com
Wessel Luijben Lead Data Scientist Rapid Miner Expert Artificial Intelligence, University of Amsterdam More than 10 years of RapidMiner and Artificial Intelligence experience Wessel@Xomnia.com
Direct Marketing is easy to implement, by selecting columns such as age, sex or marital status in order to reach the target audience.
The power of RapidMiner, showing the direct marketing demo
21 3 4
The power of RapidMiner, showing the direct marketing demo
Direct Marketing What was once a postal delivery channel now expanded to include text message, emails, online ads and other digital advertising. Business must account for spending across these platforms and provide analytics to support ROI metrics. The RapidMiner Direct Marketing Wizard allows businesses to invest only in marketing actions with the highest conversion rates and helps reduce costs by improving targeting.
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
Age more than 2 standard deviations
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
The power of RapidMiner, showing the direct marketing demo
Customer Selection At the very top, we have the a table of customers that we recommend for addressing in your marketing campaign. These are the customers for which the generated model achieves the highest confidence values. Influence Factors The following two images help you understand your customers and how the previous table was derived. The bar chart shows which of the properties of your customers have the biggest impact on whether or not they respond to your campaign. The decision tree next to it displays how you can identify potential responders. Follow a path from the root to a leaf and observe the criteria identifying this customer group. The leaf itself indicates whether there are more responders or more non-responders in that group. Sample Evaluation The predictive model generated by this template utilizes the customer data you provided. It is often costly to obtain such data since evaluation campaigns have to be run, etc. The elements in the bottom row help you to decide whether or not you have collected enough data. The "Learning Curve" on the left visualizes how the prediction quality increases as more and more data samples are added while generating the model. If this line is still rising at the far right or rather unstable, this indicates that the model may still benefit from adding more data.

More Related Content

The power of RapidMiner, showing the direct marketing demo

  • 2. Jeffrey van der Eijk Founding Partner, Responsible for data driven innovations Economics, University of Amsterdam More than 10 years of QlikView, Business Intelligence, business innovations experience Jeffrey@Xomnia.com
  • 3. Wessel Luijben Lead Data Scientist Rapid Miner Expert Artificial Intelligence, University of Amsterdam More than 10 years of RapidMiner and Artificial Intelligence experience Wessel@Xomnia.com
  • 4. Direct Marketing is easy to implement, by selecting columns such as age, sex or marital status in order to reach the target audience.
  • 8. Direct Marketing What was once a postal delivery channel now expanded to include text message, emails, online ads and other digital advertising. Business must account for spending across these platforms and provide analytics to support ROI metrics. The RapidMiner Direct Marketing Wizard allows businesses to invest only in marketing actions with the highest conversion rates and helps reduce costs by improving targeting.
  • 23. Age more than 2 standard deviations
  • 27. Customer Selection At the very top, we have the a table of customers that we recommend for addressing in your marketing campaign. These are the customers for which the generated model achieves the highest confidence values. Influence Factors The following two images help you understand your customers and how the previous table was derived. The bar chart shows which of the properties of your customers have the biggest impact on whether or not they respond to your campaign. The decision tree next to it displays how you can identify potential responders. Follow a path from the root to a leaf and observe the criteria identifying this customer group. The leaf itself indicates whether there are more responders or more non-responders in that group. Sample Evaluation The predictive model generated by this template utilizes the customer data you provided. It is often costly to obtain such data since evaluation campaigns have to be run, etc. The elements in the bottom row help you to decide whether or not you have collected enough data. The "Learning Curve" on the left visualizes how the prediction quality increases as more and more data samples are added while generating the model. If this line is still rising at the far right or rather unstable, this indicates that the model may still benefit from adding more data.