I recently got an update from Pegasystems focused on their recent release 7 of the Pegasystems Decision Management product. Last year (2013) Pegasystems had a good year, breaking $500M in revenue while showing a 55% increase in license and backlog. They are forecasting $576M in 2014 and report particular interest in predictive analytics, analytic marketing and customer next best action. From a Decision Management product perspective 2013 included a very large scale rollout for a US telco (35,000 call center agents, thousands of stores, IVR and soon online) as well as some large scale batch decisioning deployments. They are also seeing more of a focus on using the Decision Management product outside marketing especially in credit risk management, predictive maintenance and healthcare delivery.
Pegasystems Decision Management 7 includes a number of new capabilities
- New interaction tracking capabilities
The customer journey and interaction history is now managed and extensible so users can define exactly what they want to track for each interaction. All of this data can be used in the strategy designer by business users so they can build decisioning strategies to control every decision in the customer journey. This information is also available for KPI and results tracking. In particular the Decision Canvas now has a customer interaction component that allows any subset of these interactions (all those interactions in a particular channel or relating to a particular product for instance) to be brought into the decision strategy without any coding. These interactions can then be fed into other decisioning elements. KPIs that are based on this interaction data can be easily added to Visual Business Director to show how interactions are contributing to business performance. Decision strategies are also defined to determine what the customer response is and how it should be attributed. This allows complex responses to be defined using the same decisioning approach and the results of these response decisions are then stored back to the interaction history as well.
- New Decision Strategy Canvas
The whole UI for Pega 7 has been updated and this has flowed through to the decision strategy canvas where data can e brought in, business rules executed, models executed/built, champion-challenger defined etc. using a drag and drop palette. As in previous releases the decision flow is kept separate from the process flow: A process can have a node representing a decision that is then defined in a separate decision flow. Testing support has been updated and decision strategies are now tested every time they are modified or saved so business users can get immediate feedback. These tests can be defined to use synthetic data or customer data retrieved from a specified data source.
- Adaptive Analytic Models
Adaptive analytic models have been key to the Pega decisioning component for a while . In Pega 7 the analysis and reporting on these models has been extended to include things such as which possible predictors are being used or not used, the performance of individual predictors, bins and more. This exposes the details of how the model is working and adapting to the user and all this is live, reflecting the model as it constantly adapts. PDFs and static reports can be generated and the data on the performance of the models can be saved out for detailed analysis.
In the near term they see two main threads – Big Data and Complex Event Processing. They are integrating event-based data with the customer interaction data and general structured data to drive decisioning. In addition they plan to allow users to define what measures customer value for their organization and then use adaptive models to see what drives this value. These models can then be embedded in decisioning strategies.
Core to this is the extension of existing work around the automatic development of a customer analytic record. This record is based on snapshot data, wherever it may have come from. A sequence of such snapshots, which Pega call a Customer Movie, creates a view of the customer journey that includes traditional data, interaction data, “big data” etc. These movies are kept compact by only storing changes.
The decisioning technology is also being extended to handle events and event correlation. The same canvas-based approach will allow events to be correlated using one model, decisions to be defined in another and processes in a third. These can be connected to provide a decision that responds to a set of events and triggers the relevant process.
Pegasystems is one of the vendors in our Decision Management Systems Platform Technologies Report and you can get more information on Pegasystems Decision Management here.