≡ Menu

Pre-configured Digital Decisioning

Organizations that invest in a Digital Decisioning approach usually do so based on the need to solve a specific business problem. This solution focus has a number of benefits. Organizations understand the business problem to be solved and can quantify it. They are also able to put measurements in place to validate that the problem has been addressed. This moves the focus away from the technology and makes it easier for organizations to justify the investment in Digital Decisioning.

Pre-configured Digital Decisioning systems are targeted at this audience and provide an “out-of-the-box” implementation of a Digital Decisioning solution. This often includes the ability to deliver the solution through Software-as-a-Service (SaaS) or as an add-on to existing applications (example: campaign management applications). The solutions are usually made up of a combination of business rules, predictive analytics, and/or optimization that has been wrapped up in a single application, allowing the Digital Decisioning user to work in a single, simplified, user interface.

Here are some examples of the applications that can be addressed through pre-configured Digital Decisioning :

  • Targeted direct marketing
  • Next best offer
  • Inbound and outbound marketing
  • Retail assortment planning / merchandising
  • Optimizing customer service
  • Credit risk
  • Insurance fraud
  • Healthcare fraud
  • Human capital management
  • Debt collection
  • Price optimization

There are both pros and cons to using a pre-configured Decision Service versus creating a solution based on the separate pillars of a BRMS, Predictive Analytics Workbench, and Optimization System.

Pros of a pre-configured solution

  • Simplicity. The focus of the application is targeted around solving a single business problem. The business rules are usually limited in scope, complexity, and number. The number of predictive analytic models is typically small, and there is limited ability for controlling how the predictive analytic models are built. The optimization method is usually fixed and has a user interface (not code). The application provides the user with a simple focused solution.
  • One integrated user interface. A single interface exists with which the user can interact to create, configure, deploy, and maintain the Decision Service.
  • Lower IT requirements. If the solution is being deployed in-house, then it is only one application that is being installed and managed.
    SaaS option. Providing data can be easily transferred to/from the cloud based application. The management overhead of these solutions is low when purchased as SaaS, and these implementations have even less IT overhead.
  • Implementation focus. The pre-configured solution addresses a single business problem that the organization (business, IT, and analysts) can rally around.
  • Easier to deploy updates. Reduced complexity of the solution usually makes it easier for the user to update the deployed application with new rules, predictive analytic models, or new constraints for an optimization model.
  • Content. Solutions can be delivered with pre-built reports, pre-defined rules, and predictive analytic models, reducing the time to deployment.

Cons of a pre-configured solution

  • Less flexibility. As soon as you try applying the pre-configured solution to an application for which it was not intended, problems usually start to occur. They make simplifying assumptions that may not be reasonable in new situations as they are built to solve one problem and solve it well.
  • Few options outside marketing or fraud. For marketing and fraud problems, there is a good chance that a pre-configured application exists. For other applications, there is more limited availability.
  • Poor integration with existing machine learning/business rules initiatives. Most pre-configured solutions are closed and cannot easily be integrated with existing business rules or predictive analytic models. Decisions will now be defined in multiple systems, increasing the complexity of managing your decision inventory.
  • Limited cross-silo learning. A pre-configured solution may solve one problem very well, but it is unlikely to give your organization the sense of the potential for Digital Decisioning more generally. Even more than with a custom solution there is a danger that the approach will be pigeonholed as a marketing or fraud solution with no validity in other areas of the business.