Decision Manager Help. Version 7.1.7

Transcription

1 Version 7.1.7

2 This document describes products and services of Pegasystems Inc. It may contain trade secrets and proprietary information. The document and product are protected by copyright and distributed under licenses restricting their use, copying distribution, or transmittal in any form without prior written authorization of Pegasystems Inc. This document is current as of the date of publication only. Changes in the document may be made from time to time at the discretion of Pegasystems. This document remains the property of Pegasystems and must be returned to it upon request. This document does not imply any commitment to offer or deliver the products or services described. This document may include references to Pegasystems product features that have not been licensed by your company. If you have questions about whether a particular capability is included in your installation, please consult your Pegasystems service consultant. For Pegasystems trademarks and registered trademarks, all rights reserved. Other brand or product names are trademarks of their respective holders. Although Pegasystems Inc. strives for accuracy in its publications, any publication may contain inaccuracies or typographical errors. This document or Help System could contain technical inaccuracies or typographical errors. Changes are periodically added to the information herein. Pegasystems Inc. may make improvements and/or changes in the information described herein at any time. This document is the property of: Pegasystems Inc. 1 Rogers Street Cambridge, MA Phone: (617) Fax: (617) Publishing date: October 29, 2014 Contact Product Support Visit the Pega Discovery Network

3 Understanding Decision Manager Decision Manager User Activity Boundaries Business Sandbox Cycle Revision Activation Cycle Production Cycle Technical Overview Application Overlays Create Revision Management Application Overlay Create Direct Deployment Application Overlay Update Application Overlay Delete Application Overlay Proposition Group Authorization Revisions & Change Requests Assign Change Request Send Back Change Request Submit Change Request Reject Change Request Withdraw Change Request Withdraw Revision Submit Revision Activate Revision Roll-back Revision Revision Management Import Revision Discard Revision Activate Revision Roll-back Revision Decision Manager Roles & Privileges Default Roles Privileges Decision Manager Portal Dashboard Revisions Change Requests Work on Change Request Send Back Change Request Manage Revisions Managing Revisions Create New Revision Revision Details & Business Objectives

4 Change Requests Change Revision Update Change Request Withdraw Revision Approve Change Request Reject Change Request Submit Revision Proposition Data Proposition Data Overview Propositions & Proposition Attributes Manage Adaptive Models Adaptive Models Actions Clear Model Delete Model View Model Parameters Upload Responses Visual Business Director Key Performance Indicators Data Sources Views View Business Rules Monitor Adaptive Models Models Overview Behavior Report Performance Overview Predictors Overview Monitor Interaction History VBD Planner Dimensions Dimension Filter X/Y Axis Data Modes Timeline Multiple Grids Run Simulations Runs Preview Data Browse Reports Adaptive Decisioning Interaction History Simulation Reports Plan Business Rules Business Rule Types

5 Decision Tables About Decision Tables Define Table Configure Results Decision Tables in Strategies Decision Trees About Decision Trees Configure Decision Define Decision Decision Trees in Strategies Map Values About Map Values Configure Map Value Define Matrix Map Values in Strategies Scorecards About Scorecards Define Score Calculation Define Results Predictive Models About Predictive Models Define Predictive Model Results PAD Model Graphical Representation Map Segments PMML Model Statistics Predictive Models in Strategies Decision Parameters About Decision Parameters Decision Parameters in Strategies Change Decision Parameters Strategies About Strategies Strategy Design Patterns Next Best Action Capture Results Segmentation Reusable Logic Strategy Execution Strategy Design Toolbar & Context Menu Strategy Components General Settings Name & Description Source Components Properties Mapping

6 Pages & Alternative Pages Sub Strategy Components Import Category Data Import Components Interaction History Components Proposition Data Components Business Rules & Decision Analytics Categories Decision Analytics Category Predictive Model Components Scorecard Components Adaptive Model Components Business Rules Category Decision Table Components Decision Tree Components Map Value Components Split Components Enrichment Category Set Property Components Data Join Components Decision Parameters Components Aggregation Category Group By Components Iteration Components Financial Calculation Components Arbitration Category Filter Components Prioritize Components Segment Filter Components Contact Policy Components Geofence Filter Components Selection Category Champion Challenger Components Switch Components Exclusion Components Component Connections Connecting Components Expressions Segment Filtering Expressions in Strategies Understanding the Expression Context Using Component Properties in Expressions External Input Strategy Results Auto-Run Results Glossary

7 Understanding Decision Manager The Decision Manager portal provides access to the different aspects of changing Decision Management enabled applications: Revisions and change requests Applications and simulations Monitoring, visualization and reporting Proposition data management Adaptive models management The administration activities that support the work of business users in the Decision Manager portal are done through Designer Studio: Application overlays Access level to the Decision Manager portal Access level to proposition groups Revisions Decision Manager Decision Manager allows business users to make controlled changes in the business sandbox and perform monitoring in production within the boundaries defined by IT. The access group configuration determines what users can do in the Decision Manager portal. Business users work in the Decision Manager portal and system architects in Designer Studio. System architects import revisions, manage the state of revisions, define user access to propositions in a proposition group and configure application overlays. The state of a revision determines which versions of the application's resources and artifacts are active in the system. Although Decision Manager functionality can be available in different systems, the Decision Manager portal itself is typically used in the business sandbox and in production. Typically, each system has a different view of the data stores that support adaptive learning, interactions and business monitoring (online data stores in the production cycle, offline data stores in the remaining cycles). For this reason, unless data is replicated, the actual monitoring activities take place in the production system, and offline analysis in the business sandbox. Related Topics User Activity Boundaries (page 5) Business Sandbox Cycle (page 6) Revision Activation Cycle (page 7) Production Cycle (page 8) User Activity Boundaries In Designer Studio, system architects define the boundaries for the business user activity in the business sandbox and in production through the Decision Manager portal. They are responsible for creating application overlays, defining the exposure of propositions to Decision Manager portal users and maintaining access group user membership. 5

8 Business Sandbox Cycle In the Decision Manager portal, revision managers and strategy designers work in the application overlay to achieve the revision objectives. Revision managers address production issues reflected in the business sandbox, and define the revisions and change requests to resolve these issues. Change requests take the form of assignments strategy designers need to complete so that the revision can be processed by revision managers. Strategy designers perform the work necessary to achieve the objectives of the revision by addressing the goals of the change request assigned to them. They make the necessary changes, test applications, run simulations and analyze the impact of the changes, a cycle that is repeated until the objectives are met. They make the changes visible to revision managers by submitting the change request. If the strategy designer rejects the change request, the changes are rolled back to the version before the definition of the revision. 6

9 Revision managers process the changes submitted in the context of the change request. They can run applications and simulations to validate the change request, and then accept or reject the changes. Once all change requests are completed (accepted or rejected), revision managers submit the revision to package the changes in the form of a JAR file. Revision managers hand over the packaged changes to system architects for deployment in another system. Revision Activation Cycle System architects import the revisions handed over to them by revision managers. The revision import process allows for reflecting in other systems the changes made and tested in the business sandbox. In this process, a step for validation is introduced so that test users can validate the impact of the changes in the end user application. After validation, system architects manage the state of the revision by making the revision available to all users, or by discarding the changes. 7

10 Production Cycle Once system architects activate the revision for all users in the production environment, production supervisors monitor the performance and accuracy of interactions through the Decision Manager portal. Production administrators can perform the same monitoring activities as supervisors but they go beyond a supervisor's observer activities. They can monitor adaptive models, update the data used in adaptive models monitoring and reporting, perform on adaptive models, update proposition data and run applications. 8

11 Technical Overview The following topics allow you to understand the system operations underpinning the actions performed by system architects and portal users: Roles Actions System architects in Designer Studio Working with application overlays (page 9) Configuring proposition authorization (page 11) Strategy designers and revision managers in the Decision Manager portal Working with revisions and change sets (page 11) System architects in Designer Studio Managing revisions (page 13) Application Overlays System operations underpinning the actions performed by system architects working with application overlays: Create revision management application overlay (page 11) Create direct deployment application overlay (page 10) Update application overlay (page 10) Delete application overlay (page 11) Application overlays have an important relationship with the access groups that define what users can do in the Decision Manager portal. Refer to the Decision Manager roles and privileges (page 14) section for an overview of the technical details concerning the implementation of these roles and privileges. In the process of creating application overlays, the wizard does not associate operators with the access groups. System architects have to configure the operator records accordingly. 9

12 Create Revision Management Application Overlay The system architect runs the new overlay wizard, resulting in the following operations: Create an initial revision ruleset version with revision ruleset dependencies. Create a branch ruleset for the revision ruleset. Create the overlay application with revision ruleset version in the application's ruleset list. Create the access groups with the selected role configuration. These access groups are configured to include pydecisionportal in each access group's portal settings. Create a copy of the pxdecisioningclass field value (location of the SR class) in the revision ruleset. Update the current application to contain the revision ruleset in the first row of the application's ruleset list. Associate the branch ruleset with the application overlay. Create a data instance of Data-Decision-RevisionRecords to contain the list of the selected revision rules. Copy the instances selected for the application overlay into the revision ruleset. Create Direct Deployment Application Overlay Application overlays for the purposes of direct deployment are automatically generated for each application in non-production systems. Generating these overlays is triggered by system architects accessing the application overlays landing page for the first time, resulting in the following operations: Check if there is an existing application overlay for direct deployment. If this application overlay does not exist, proceed with creating it. Create the overlay application with revision ruleset version in the application's ruleset list. The name of the application overlay follows this pattern: RTC-<first ten characters of application name>. Create an initial revision ruleset version with revision ruleset dependencies. This ruleset uses application validation, not ruleset validation. The name of the ruleset is based on the following pattern. Create a branch ruleset for the revision ruleset. Associate the branch ruleset with the application overlay. Create the <application_name>:directdeploy access group. Create a copy of the pxdecisioningclass field value (location of the SR class) in the revision ruleset. Create a data instance of Data-Decision-RevisionRecords to contain the list of the selected revision rules. Initially, this list is empty. By default, the name of a direct deployment overlay is constructed by prefixing RTC- to the first ten characters of the application's name. You can override the default prefix by configuring the DirectDeploymentOverlayName dynamic system setting (owning ruleset: Pega-DecisionManager). If any of the automatically generated names used in the process of creating a direct deployment application overlay exist, the names are appended with a count (for example, BranchRuleset1). Update Application Overlay The system architect updates a revision management application overlay, resulting in the following operations: Nothing happens if the update results in no changes. If the update results in removing rules, update the data instance of Data-Decision-RevisionRecords to remove these instances. If the list of rules changes: Create a new minor revision ruleset version (for example, ). Copy the new instances to this new minor revision ruleset version, and lock the ruleset version. Update the branch ruleset with the new minor revision ruleset version. 10

13 Update the current application's revision ruleset version with the new minor revision ruleset version. Update the application overlay's revision ruleset version with the new minor revision ruleset version. The system architect updates the list of rules (decision parameters) a direct deployment application overlay, resulting in the following operations: Every update consists of updating the revision records but the ruleset version remains the same (that is, version is not incremented). If the update results in removing rules, update the data instance of Data-Decision-RevisionRecords to remove these instances. If the update results in adding rules, update the data instance of Data-Decision-RevisionRecords to add these instances. Delete Application Overlay The system architect deletes an application overlay, resulting in the following operations: Where used by operators, remove the access groups from the operator's access group list. Delete the access groups associated with the application overlay. Delete the overlay application. Delete the data instance of Data-Decision-RevisionRecords containing the list of the selected revision rules. Delete the branch ruleset. Deleting an application overlay can only be done for revision management application overlays, as direct deployment application overlays are system-defined, not user-defined. Proposition Group Authorization System architects configure proposition authorization. Data instances of Data-DecisionPropositionGroupsAuthorization determine the propositions production administrators and strategy designers can manage through the Decision Manager portal. System architects add proposition groups to an access group to define the propositions groups Decision Manager portal users can work with. This action results in creating a data instance of Data-Decision-PropositionGroupsAuthorization per proposition group. Each data instance is deleted when system architects remove a proposition group from the access group's proposition groups authorization. Revisions & Change Requests System operations underpinning the actions performed by Decision Manager portal users working with revisions and change sets: Assign change request (page 11) Send back change request (page 12) Submit change request for approval (page 12) Reject change request (page 12) Withdraw change request (page 12) Withdraw revision (page 12) Submit revision (page 12) Activate revision (page 13) Roll-back revision (page 13) Assign Change Request The revision manager defines and assigns a change request, resulting in the following operations: Copy the instances included in the change request to the branch rule set when then strategy designer edits them. 11

14 Copy and check out the instances to the strategy designer's personal ruleset. Change status to in progress. Send Back Change Request The strategy designer sends back the change request, resulting in the following operations: Verify checked out instances. Remove instances from the strategy designer's personal ruleset. Change status to pending-update. Submit Change Request The strategy designer submits a change request for approval by the revision manager, resulting in the following operations: Check in instances into the branch ruleset. Remove instances from the strategy designer's personal ruleset. Change status to pending-approval. Reject Change Request The revision manager rejects the change request requested for approval, resulting in the following operations: Remove from the branch the instances that are not included in an already approved change request. If the instances were included in another change request within the same revision, and that change set was already approved, rejecting the change request does not remove the instances from the branch, but instead replaces them with the versions of the approved change request. Change status to resolved-rejected. Withdraw Change Request The revision manager withdraws the change request assigned to the strategy designer, resulting in the following operations: Remove instances from the strategy designer's personal ruleset. Change status to resolved-withdrawn. Withdraw Revision The revision manager withdraws revision, resulting in the following operations: Verify checked out instances. Remove instances from the strategy designer's personal ruleset. Remove instances from the branch. Change status to resolved-withdrawn. Submit Revision The revision manager submits the revision, resulting in the following operations: Copy modified instances from the branch ruleset to the revision ruleset. The revision ruleset version is the next available version. After copying, lock the revision ruleset. Remove instances from the branch. Generate the revision package containing the last created and locked revision ruleset. If it is a generic revision (rule revision management purposes, not direct deployment), change status to resolved-completed. If it is a direct deployment revision, and there are changes in decision parameters rules, change status to pending-activated. 12

15 Activate Revision The revision manager activates changes included in a direct deployment revision, resulting in the following operations: List access groups of the current version of the application the overlay is built on. For every access group, update the version of the production rulesets to match the new revision ruleset version. Change status to resolved-activated. Roll-back Revision The revision manager rolls back a resolved-activated revision, resulting tin the following operations: List access groups of the current version of the application the overlay is built on. For every access group, update the version of the production rulesets with the previous available resolved-activated revision ruleset version. In the absence of such a previous version that can be used, remove the revision ruleset from each access group's production rulesets. Change status to resolved-rolledback. Revision Management System operations underpinning the actions performed by system architects managing the packaged revisions generated by revision managers in the Decision Manager portal: Import revision (page 13) Discard revision (page 13) Activate revision (page 13) Roll-back revision (page 14) Import Revision The system architect imports a revision, resulting in the following operations: Increment the application patch version. Create the roll-out access group based on the current access group. Update the roll-out access group to point to the new application version. Append the roll-out access group to the list of access groups for the selected test users. Change status to testing. Discard Revision The system architect discards a revision under testing, resulting in the following operations: Remove the incremented application patch version. Remove the roll-out access group. Clear the roll-out access group from any any operator selected to test the revision. Change status to discarded. Activate Revision The system architect activates revision under testing, resulting in the following operations: Update the current access group with the roll-out access group's configuration. Update the built on version of the application overlay to point to the new version (that is, the version of the application that was under testing). If this revision overrides decision parameters coming from direct deployments in non-production systems: Remove the versions of decision parameters instances from the direct deployment overlay. Remove the production ruleset versions pointing to the direct deployment revision ruleset. Update all revisions to resolve the roll-back of direct deployment decision parameters. 13

16 Change the resolved-activated direct deployment revision in the Decision Manager portal to resolved-rolledback. If this revision does not override decision parameters coming from direct deployments in nonproduction systems: Changes to decision parameters that are defined in both this revision and the direct deployment revision are not taken into account. The resolved-activated direct deployment revision in the Decision Manager portal remains in resolved-activated state. Change the status of the currently active revision to inactive. Change the status of this new revision to active. Roll-back Revision The system architect rolls-back a revision in production, resulting in the following operations: Create a new application version. Copy the contents of the previously active revision to this new application version. Update the current access group with the new application version. Update the built on version of the application overlay to point to the new application version. Change the status of the currently active revision to rolled-back. Change the status of this new revision to active. Decision Manager Roles & Privileges Access to the Decision Manager portal is governed by access roles, and each access role is associated with privileges. The access to the Decision Strategy Manager portal requires the applicable access groups to list pydecisionportal in their portal settings. The Decision Strategy Manager portal also provides access to the. Default roles (page 14) Privileges (page 15) Default Roles Decision Manager defines default access roles that can be used to configure what users can do in the Decision Manager portal. Typically, roles apply to the business sandbox or the production environment. Business sandbox roles: Access Role Description DecisionManager:StrategyDesigner Strategy designers work on change requests, change proposition data, test applications, run simulations, analyze results through reports and Visual Business Director, manage the assets used in the Visual Business Director planner and submit the change request for approval by revision managers. They also have access to the calendar. DecisionManager:RevisionManager Revision managers create revisions and change requests, assign change requests to workbaskets or strategy designers, manage the assets used in the VBD planner, run applications and simulations to validate changes, approve changes and package revisions. Revision managers can also act as strategy designers when change requests are assigned to them. DecisionManager:DecisionArchitect This role is reserved to Decision Management experts because it combines the privileges given to strategy designers, revision managers, production supervisors and production administrators. Although typically applicable to the business sandbox, it can also apply to the production environment. Production environment roles: 14

17 Access Role Description DecisionManager:Supervisor Production supervisors keep interactions under observation through reports and Visual Business Director. They can also update adaptive reporting data. The direct deployment access role provides users with the ability to directly deploy revisions originated from direct deployment overlays. User can manage revisions, work on change requests run applications and view reports. Production administrators have the same access as supervisors. Additionally, they can update proposition data, manage the assets used in the Visual Business Director planner, manage adaptive models and run applications. DecisionManager:DirectDeploy DecisionManager:Administrator Privileges Decision Manager defines the privileges used to set up access roles according to the level of access to the Decision Manager portal. These privileges are designed to be granular so that you can set up custom roles, and use them instead of the default Decision Manager roles. Access Users with this privilege can... Privilege Adaptive Models Monitoring Monitor adaptive models, monitor predictors and access ADM reports. Update adaptive reporting data. Manage adaptive models through the actions that allow for controlling the adaptive data store (clear, delete and train models). Add, edit and delete propositions. pymonitoradmmodels Access reports in the Interaction History category. Access the report browser. Run applications. pymonitorinteractionhistory Adaptive Models Management Proposition Data Reports Test Applications Visual Business Director Use the VBD planner to monitor actuals. Use the VBD planner to monitor simulation data sources. Manage KPIs, data sources and views. Simulations Run simulations. View business rules, revisions and work Revisions list. Access the revisions facilities. Change Requests View business rules, work on change requests, open rules, change rules and submit changes. Plan Access the Plan facilities. pyupdateadmreportingdata pymanageadmmodels pymanagepropositiondata pymonitorreports pytestflows pymonitorvbdactuals pymonitorvbdsimulation pymanagevbd pyrunsimulations pymanagerevisions pyworkonchangesets pyviewcalendar 15

18 Decision Manager Portal The Decision Manager portal allows business users to make changes in an enterprise application. Your assigned role determines the type of actions you can perform. By default, an application overlay in the business sandbox defines the decision architect, strategy designer and revision manager roles. In production, the default roles are supervisors and administrators. The portal is available in the multiple environments set up to support enterprise applications, allowing for: Improving the effectiveness of Decision Management enabled applications by streamlining the change process. Providing a controlled change environment following a process based on case management principles. The navigation and home page are filtered for each role. The portal's header allows for switching applications and running the processes designed to test the changes made to the application. Dashboard The Dashboard page allows strategy designers and revision managers to work on a revision's business objectives. It displays the list of open work items, and allows for viewing the list of work items assigned to the team members. Use the Refresh button to update the list. The type of work items displayed in the dashboard is specific to your role: Revision managers work on revisions (page 16) Strategy designers work on change requests (page 16) Revisions As a revision manager, you create new revisions, and approve/reject changes submitted by strategy designers (page 19). Change Requests As a strategy designer, you work on change requests assigned to you or to a workbasket. 16

19 Working on change requests (page 17) Sending back change requests (page 18) Work on Change Request 1. Go to the dashboard, and open the link to the change request. 2. The change request provides the change request details and includes the links to business rules. 3. In the process of working on the change request, you typically: Change business rules (page 38). Run flows to test and validate your changes. Repeat this process until you achieve the change request's objectives. 4. Once your work in the change request is complete, submit the change request to the revision manager. Click Submit. Review the list of changes. If the list contains more instances than necessary (for example, if you changed an instance but that change is not relevant), click Cancel, go back to the change request to discard the corresponding instances, and submit the change request again. Provide a comment, and click Submit. 17

20 5. The change request goes to pending-approval state, and is now assigned for approval by the revision manager. Send Back Change Request Use the change request's Send back action to reassign the change request to the revision manager. 1. In the change request, click Send back. 2. Select the reason. Use the entries in the drop-down list to request: Changes to the due date Changes to the list of rules included in the change request Reset of rules back to the original version Reassignment of the work item Other actions 3. Provide a comment. 4. Click Submit. Changes you have made when working on the change request are discarded when you submit. Manage Depending on your role, these facilities allow you to perform the actions relevant to the way you need to influence changes in the business sandbox or manage assets in production. 18

21 Roles Actions Revision managers Create and manage revisions (page 19) Configure and manage data sources, Key Performance Indicators and views used in Interaction History monitoring (page 27) Strategy designers Manage proposition data in the business sandbox (page 24) Configure and manage data sources, Key Performance Indicators and views used in Interaction History monitoring (page 27) Production administrators Manage proposition data in production (page 24) Manage adaptive models in production (page 25) Configure and manage data sources, Key Performance Indicators and views used in Interaction History monitoring (page 27) Revisions As a revision manager, you can create and manage revisions. Each revision has a defined set of business objectives that describe the change. It also contains the change requests that define the issues strategy managers need to address for the revision to be approved. Only one revision is allowed to be in progress at any given point in time, a restriction that is in place to avoid revision conflicts. Once the revision is either approved or rejected, revision managers can create a new revision. A special type of revisions, direct deployment revisions, follow a different flow and work differently from revisions created for the purposes of revision management: The ability to create these revisions depends on how the system is set up: the system must include a direct deployment overlay. This direct deployment overlay can only include decision parameter instances. When processing this type of revisions, you activate the revision directly and no revision package is generated in the process of approving the revision. The status of such a revision changes to resolved-activated. The Revisions tab displays the list of existing revisions and corresponding details. Name: name of the revision as assigned by the revision manager. ID: system assigned ID. Revision: revision number. Objectives: the business objectives. Due By: time elapsed since the target completion date, or time remaining to reach this date. Completed: time elapsed since the revision was processed. Assigned To: user or workbasket the work item is assigned to. Status: status of revision. Managing Revisions Click the New button to define a new revision (page 20). Once the revision is created, you can continue adding as many change requests (page 22) as necessary to accomplish the objectives. If strategy designers requires updates, update the change request (page 22) before reassigning it. You can withdraw the revision (page 23), in which case you discard the revision's change requests. Change requests submitted for your can be rejected (page 23) or approved (page 23). 19

22 Finally, when all change requests have been processed, you can submit the revision (page 23). You can access revision history details in the revision's Audit tab. Create New Revision A revision is defined by the business objectives and the change requests required to successfully meet these objectives. Define revision details and business objectives (page 20) Define change requests (page 20) Revision Details & Business Objectives Define business objectives and revision details: 1. In the Revision creation options, select the type of revision: Create new revision: use this type when address issues found in production (default). Base this revision on a previous one: used this type when rolling back a revision that has been submitted and approved. 2. Enter the name of the revision. 3. Define one or more objectives. 4. Set the urgency and the target completion date. 5. Click Complete Set-up & Create Change Request to add change requests to this revision. Change Requests Add as many change requests as necessary, and assign them to the appropriate user (typically, a strategy designer): 1. In the Define Change Request Details section: Assign the change request using the appropriate option (resource, workbasket or assign it to yourself). 20

23 Enter the name of the change request. Associate the change request with one of the revision's objectives. Enter a description about the work required for this change request. Set the urgency and the target completion date. 2. A new change request does not contain any instances. To add instances that can be edited in the context of this change request go to the Select Business Rules section: Select the business rules that are required to work on this change request. You can add all business rules in the revision to the change request or select a sub set of rules. Use the Type selection to filter by rule type. Use the Name selection to include all instances of the rule type or just a single instance. Click Include in change request. The selected instances are now part of the rules included in the change request, and the list of instances that can be added to the change request is updated to exclude them. If you select all business rules, none can be added to other change requests until the current change request is completed. 21

24 3. Submit the change request: Click Submit & Create new to save and assign the change request, and proceed with adding more change requests. You can use this action as long as there are business rules available to include in change requests. Click Submit to save and assign change request. Change Revision Provided there are business rules still available for new change requests, you can add more change requests to the current revision. 1. Go to Manage > Revisions, or use the link in your dashboard. 2. Click the link in the Name column. 3. Use the Create Change Request button to define more change requests (page 20) to include in this revision. Update Change Request Update change requests in pending-update state. 1. Go to Manage > Revisions, and expand the current revision to view the pending update change requests. These work items are also listed in your dashboard: 2. Open the change request. 22

25 3. The update request details are displayed at the top of the change request. 4. Click Update. 5. Reassign the work item to the appropriate user or workbasket. 6. Make the necessary changes. 7. Click Submit. Withdraw Revision Any revision in progress can be cancelled by withdrawing it as a whole. 1. Go to Manage > Revisions, or use the link in your dashboard. 2. Click the link in the Name column. 3. Use the Withdraw button to cancel the current revision. Approve Change Request Approve change requests in pending-approval state. 1. Go to Manage > Revisions, and expand the current revision to view the pending approval change requests. These work items are also listed in your dashboard. 2. Open the change request. 3. Click Approve. 4. Provide a comment to justify approving the change request. 5. Click Submit. Changes included in the submitted change request are ready to be packaged when approving the revision, and the change request's status changes to resolved-completed. Reject Change Request Reject change requests in pending-approval state. 1. Go to Manage > Revisions, and expand the current revision to view the pending approval change requests. These work items are also listed in your dashboard. 2. Open the change request. 3. Click Reject. 4. Provide a comment to justify rejecting the change request. 5. Click Submit. Changes included in the submitted change request are discarded when you submit, and the change request's status changes to resolved-rejected. Submit Revision When all change requests have been processed, you can submit the revision. 1. Go to Manage > Revisions, and click the name of the current in progress revision. 2. Click Submit. 3. Provide a comment to justify submitting the revision. 4. Click Submit. 23

26 5. The revision's status changes to resolved-completed, and changes in resolved-completed changes requests are packaged in the archive you can download in the Overview tab. The archive contains the changes that can be imported in a different system for testing and, ultimately, activation. Proposition Data As a strategy designer in the business sandbox or as a production administrator, you can manage proposition data. Production administrators typically manage proposition data in production environments, strategy designers do this in the business sandbox. The proposition group authorization associated with your access group determines the propositions that you can access. Proposition Data Overview Use these facilities to view the propositions in your application, manage propositions and define values for proposition properties. 24

27 New: add a new proposition using the New Proposition dialog. Bulk edit: select the group for which you want to create, edit or delete propositions using the Edit <Group> Data Table instances dialog or Excel. Delete: delete the data instance representing the selected proposition(s). All propositions with a check mark are deleted when you confirm this action. Search: enter a value in the text box and click Search to narrow the amount of information displayed in the Propositions tab. The value is matched against any value in the columns displayed in the grid. Propositions & Proposition Attributes The New Proposition dialog allows you to define new propositions. The definition of a new proposition consists of providing the following values and settings: Field Description Proposition name Enter the name of the proposition. Description Enter an optional description. Business issue Select the business issue. Group Select select the group. Properties Proposition properties are displayed depending on the scope of this proposition. Define values for the applicable properties. After defining the details, click Submit. You can also use the Submit & add new button to continue adding more propositions. The expand control provided to the left of a proposition's check box displays the area to define proposition attributes, create a new proposition based on this proposition and delete the proposition. Edit: edit the values of properties that apply to the proposition. Duplicate: open the Duplicate Proposition dialog. This dialog works similarly to the New Proposition dialog and allows you to create a new proposition based on the details of an existing one. Delete: delete the proposition. Manage Adaptive Models As a production administrator, you can manage the adaptive models in the ADM system. Adaptive Models The overview of adaptive models shows the information pertaining to every adaptive model. The information displayed in the overview depends on how adaptive models rules are defined and how each adaptive model component is defined in the strategy. 25

28 Business issue: the issue in the decision hierarchy. Group: the group in the decision hierarchy. Proposition: the name of the proposition the adaptive model is modeling. Direction: the direction. Channel: the channel. Rule: the name of the adaptive model rule that configures the adaptive model. Applies to: the class the adaptive model instance belongs to. Responses: the number of responses. Performance: the model performance. Action: actions to manage the adaptive statistics for a given model. Show Last Responses allows you to analyze response model and predictor information for the last five records sent to the ADM database. You can use this information to troubleshoot the learning process: Check if response information contains the data required by the predictors. Check response information that has not been factored in the ADM system. For example, responses that have not been taken into account because the process of running data analysis was not triggered. Actions Use the actions menu to manage the adaptive statistics for a given model, view model settings and train models. Clear Model This action allows you to remove the adaptive statistics (page 90) associated with the adaptive model. In this process, everything is cleared, including numeric predictors boundaries. 1. Locate the model. 2. In the Actions menu, select Clear Model to open the Clear Adaptive Model dialog. 3. Press the Clear button to remove the adaptive statistics from the ADM system. Delete Model This action allows you to delete the current model. If used by a strategy, the model is recreated again when the strategy is executed. Deleting models implies the loss of the adaptive statistics (page 90) associated with that model. 1. Locate the model. 2. In the Actions menu, select Delete Model to open the Delete Adaptive Model dialog. 3. Press the Delete button to remove of the model from the ADM system. View Model Parameters Models are created based on the values and behavior defined in the adaptive model's settings. You can view these settings through this action or by opening the adaptive model. 1. Locate the model. 2. In the Actions menu, select View Model Parameters to open the Adaptive Model Parameters dialog. 3. Inspect the settings used for creating models. Upload Responses The Upload Responses wizard allows you to train the model by uploading existing customer data representing previous behavior or sample data. The use of previous results allows for Adaptive Decision Manager to create models that are able to predict behavior. ADM only considers positive and negative cases that correspond to the possible outcomes taken into account by the settings defined in the adaptive model rule. Typically, this action is not performed in production, only in systems that connect to the offline analytics data store. The CSV file should contain the input data for each case (page 90) and a set of interaction results (page 91). Follow the steps described below to use a CSV file to train models. 26

29 1. Locate the model you want to train. 2. In the Action drop-down of the model, select Upload Responses. 3. Press Start. 4. In the Select File step, use the Browse button to select the CSV file containing historical data and click Next. 5. In the Select Outcome step, select the column that provides the historical outcome for each case and click Next. 6. In the Map Outcome step, define how the outcome in the historical data should be mapped to the outcome defined in the adaptive model. 7. Click Finalize to make the historical data available to the adaptive analytics engine (page 90). Visual Business Director As a strategy designer in the business sandbox or production administrator, you can manage the settings and assets used to configure business monitoring and visualization in the VBD planner. Key Performance Indicators (page 27) Data sources (page 29) Planner views (page 29) Key Performance Indicators The Key Performance Indicators landing page allows you to manage the available Key Performance Indicators. Click New to add a new KPI, click the link in the Name column to change an existing KPI. Once defined, KPIs are calculated every time the interaction rule writes results to Interaction History, Visual Business Director and database tables. 27

30 Name: name of the KPI. Expression: expression used to calculate the KPI. Updated and Updated by: the time stamp and the user name corresponding to the last change. The Edit KPI dialog allows you to add or change KPIs. Name: the name of the KPI. Select the type of operation that defines the KPI formula. 28

31 Operator Description Sum Sum of the values provided by a measurement (Interaction History properties that can be used as KPIs) when the possible outcome filtering is applied. Count Count (volume) of the results is the selected possible outcome values. Average Average of the values provided by a measurement when the possible outcome filtering is applied. Ratio The ratio between two volumes defined by the first and second KPIs. Difference The difference between two volumes defined by the first and second KPIs. Available outcomes (count, average and sum): list of possible values in the outcome dimension. Selected outcomes (count, average and sum): list of available values in the outcome dimension selected for this KPI. Display data in Visual Business Director (sum and count) Cumulative: VBD displays values accrued over time. Non-Cumulative: VBD does not display values accrued over time. Compare data sources in Visual Business Director Higher values are better: values below a reference value are displayed in red and green when above. Higher values are worse: values below a reference value are displayed in green and red when above. Data Sources The Data Sources tab displays the data sources. The list of available data sources consists of VBD's view of the Interaction History (Actuals), or any other data source created through running an interaction rule that is configured to write simulation results to VBD. Name: the name of the data source as defined in the interaction rule. Start date: the start date for monitoring data purposes. End date: the end date for monitoring data purposes. Number of records: the total number of records in the data source. You can perform the following actions to manage the selected data source. Based on the information provided in the Visual Business Director grid node topology, you can use the Delete and Clear actions to free up memory resources. Remove data sources by using the Delete button. This action results in permanently losing the data and information associated with the data source. Delete the records in a data source by using the Clear button. This action results in permanently removing all records associated with the data source but it does not delete the data source. Define the start date for monitoring data purposes by expanding the data source, defining the date in the Start date field and pressing the Save button. Views The Views tab provides an overview of saves planned views. Views are saved in the VBD planner through the Save View As action in the planner's context menu. Use the Delete button to delete the selected views. Click the view in the Name column to open the VBD planner with the settings defined in the view. Name: the name of the view as defined by the users when saving the view. Updated and Updated by: the time stamp and the user name corresponding to the last change. View Business Rules As a strategy designer or revision manager, you can use View Business Rules to get an overview and general details about the business rules instances (page 38) contained in the application overlay. 29

32 Name: the name of the instance. Type: the type of rule (decision parameters, decision tables, decision trees, map values, predictive models, scorecards or strategies). Updated: time elapsed since the instance was last updated. Updated By: user information corresponding to the last update. Monitor Adaptive Models As a production administrator, you can monitor the performance and statistics of adaptive models in the ADM system. Reporting on adaptive models is based on the last data snapshot, which means time and date may be not the same as when you request the report. Use the Update Reporting Data link to take a new snapshot of the models and predictors in the ADM system. Models overview (page 30): inspect active and inactive predictors. Predictors overview (page 31): inspect the usage of predictors for all adaptive models in the ADM system. Models Overview Models Overview displays all models in the ADM system in a bubble chart, and allows you to compare models in terms of success rate and performance. You can apply filtering to focus on models for a particular direction, channel or a combination of the two. The data used to visualize the models is displayed below the chart (model in the action context, response count, performance and acceptance rate). Use the reporting actions to: Obtain the behavior profile (page 30) Obtain the performance overview (page 31) Behavior Report The behavior report contains the behavior profile. The behavior profile allows you to observe the treatment of predictors (page 93) and the difference between positive and negative cases. The grouping is automatically determined by ADM. The behavior profile is centered around predictor grouping (page 92) and predictive performance (page 91). 1. Locate the model in the data grid. 2. Click the Select Report action. 3. Select Behavior Report. Model information provides information about the modeled proposition. Column Description Predictors Performance Bounds/Symbols Name of the properties used as predictors. Bins Positives Total predictive performance. Bounds for numeric predictors or number of symbols for symbolic predictors. Number of bins (the number of bins is influenced by the adaptive model rule's grouping settings). Number of positive cases. 30

33 Negatives Number of negative cases. Type Predictor's data type: numeric or symbolic. List of active predictors: predictors actively used for modeling, also known as filtered behavior. List of inactive predictors: predictors that are not used for modeling. Predictors are inactive if their performance is below the performance threshold defined in the adaptive model rule. The Classifier section displays the model's predictive performance (unfiltered behavior). For every predictor, use the expand control to view the break down of statistics per interval/category. Performance Overview Performance Overview allows you to monitor predictive performance (page 91) over time. 1. Locate the model in the data grid. 2. In the Select Report action, select Performance Overview to display the graph showing the model's performance over time. Predictors Overview Predictors Overview displays predictor information in a bar chart. This overview allows you to analyze predictor use across models in the ADM system. Use the drop-down to select the criteria for reporting on predictors: Average Performance: overview of predictors based on the average predictive performance. Maximum Performance: overview of predictors based on the highest predictive performance. Minimum Performance: overview of predictors based on the lowest predictive performance. Number of Models Where Predictor is Active: overview of active predictors based on number of models. Number of Models Where Predictor is Inactive: overview of inactive predictors based on number of models. The data used to visualize the predictors is displayed below the chart (predictor name, number of models where the predictor is active, number of models where the predictor is inactive, minimum performance, maximum performance and average performance). Monitor Interaction History As a strategy designer, production administrator or production supervisor, you can use the VBD planner to monitor Interaction History and analyze simulation results. VBD planner overview (page 31) Filtering on dimensions (page 32) Using data modes (page 33) Using the timeline (page 34) Multiple grids in the VBD planner (page 35) VBD Planner The VBD planner provides the visualization based on different views of the business, and allows you to assess the success of your business strategy through Key Performance Indicators (KPIs). You can use the VBD planner provided that one or more KPIs have been defined for your application. Interaction History data (dimensions, properties and KPIs) forms the basis for visualizing decision results with a 3D graphical view of the different dimensions and KPIs, such as accept rate, conversion rate, average price, volume, number accepted or number of processed responses. 31

34 The button on the top left corner of the VBD planner allows for refreshing the data. On the top right corner, buttons are available for zooming in/out, changing the position of the VBD planner and setting it to default view. The context menu allows you to capture the current state of the VBD planner to clipboard or file, save a view and load a predefined view. Bar charts on the VBD planner's walls show the calculation for a single dimension and target values. Each line chart shows a different KPI over the current time period selected in the timeline. Clicking a chart makes the grid use the KPI and double clicking allows you to change which KPI is displayed by the chart. Bars show the performance for a combination of two dimensions. Color saturation is applied based on quantity below/above target and each bar provides a view of the statistics. The VBD planner supports changing the dimensions displayed in the X (right) and Y (left) axis, using different data modes, displaying multiple grids and filtering on different dimensions. Related Topics Dimensions (page 32) Data Modes (page 33) Timeline (page 34) Multiple Grids (page 35) Dimensions Together with KPIs, dimensions are used in the VBD planner to construct the business view. The X/Y axis dialog determines the visualization of dimensions. Dimension filter (page 32) X/Y axis (page 33) Dimension Filter The dimension filter allows for dynamically defining customer interactions using multiple dimensions. The filter works as a slider displaying the hierarchy of the different dimensions. The tree view shows the different levels up to the selected item. Use single (left) click selection to select/deselect levels. If the item is already part of the defined filter settings, it will be deselected. Each dimension is displayed in a different color. The color saturation shows the proportion of selected items in the dimension's hierarchy. 32

35 X/Y Axis You can configure how dimensions are displayed in the X and Y axis by double clicking the level displayed in the axis and using the set axis dialog to select a different dimension or level within the dimension. The VBD planner reflects the changes when you confirm the new dimension/level selection in the set axis dialog. For example, you can perform outcome time based analysis by setting the Y axis to outcome, and setting the X axis to the time period for showing customer behavior. Y Axis 1. Double click the axis label on the left. 2. Use the Set Y Axis dialog to change it to show All Outcomes. X Axis 1. Double click the axis label on the right. 2. Use the Set X Axis dialog to change it to show Day. Data Modes The top part of the panel on the far right of the VBD planner's back wall provides the facilities for setting different data modes. The VBD planner can operate in regular, reference data or delta mode. 33

36 The regular mode shows the source versus reference data sources. Switch to this data mode by clicking the icon. The reference mode shows the interaction history records corresponding to the time period selected in the Reference drop-down. Switch to this data mode by clicking the icon. The delta mode provides the comparison between the Reference and Source data sources, allowing you to analyze their relative effects. Switch to this data mode by clicking the icon. Timeline The timeline is the interface allowing for browsing recorded historical performance and predicted future performance. This interface consists of timeline console and timeline display. The and buttons allow you to hide/show the display and the console. The console allows you to set the date settings for displaying data in the VBD planner and select the data sources. Selection Description From Selects the start date for showing the data. The earliest date to be displayed is determined by the data source's configured start date. Selects the end date for showing the data. To Source Displays the duration between start and end date. Duration is displayed in days, hours, minutes and seconds. Drop-down for selecting which data source to use as source. Reference Drop-down for selecting which data source to use as reference. Duration Source and Reference display the data sources listed below. If present, you can also select data sources created in the process of executing interaction rules that write results to VBD. 34

37 Actuals - MONTH: compare to last month aligned on days. For example, compare September 15 with August 15. Actuals - QUARTER: compare to last quarter aligned on days. For example, compare September 15 with June 15. Actuals - YEAR: compare to last year aligned on days. For example, compare September 15, 2012 with September 15, The selected data sources are listed on the left side. Each data view has the corresponding line showing when data is available. Two time sliders allow you to change the start and end date. The recommended method for changing the time span is by using the From/To selection in the timeline console (page 34), but you can also change it by dragging the vertical sliders or using the scroll buttons. When doing this, the timeline console reflects the changes. Multiple Grids The VBD planner supports adding multiple grids, each visualizing different aspects of the business strategy. Enabling multiple grids is done by using the + icon at the far right of the X axis. After adding multiple grids, you can define filters for each individual grid using the facilities provided in the back wall. You can remove a grid by clicking the close button on the top right corner of the back wall. The timeline selection remains common to all grids. Run Simulations As a strategy designer or revision manager in the business sandbox, you can use simulations to test and analyze changes to the enterprise application in the business sandbox. Run simulations (page 35) Preview data (page 36) Runs Use these facilities to define and manage simulation definitions. Use the button at the end of the corresponding row to delete a simulation definition. The overview shows the definitions available in your application. ID: the simulation identifier. Interaction: the name of the interaction rule. Strategy: the name of the (first) strategy rule selected by the interaction rule. A simulation can only run one strategy; if the interaction rule has multiple strategies, the simulation takes the first strategy. Input Definition: the name of the input definition. Output Definition: the name of the output definition. Records Processed: the number of data records processed. Updated and Updated By: the period of time elapsed since the last update to the simulation definition (creating, updating or running the simulation) and the operator name. Status: the work item status. Use the icon to delete the simulation definition. Use the New button to open the Simulation Setup form. This form allows you to create a new simulation definition. Simulation Definition Description Input Definition Select the input definition. After selecting the input definition, you can use the corresponding links to: 35

38 Save View the input definition instance. Preview the data retrieved by the input definition (page 36). Select or enter the name of the interaction rule. The Simulation Setup form displays information about the interaction rule configuration. Save the simulation. Save and Execute Save and run the simulation. Interaction The progress of running the configuration is displayed after clicking Save and Execute. Click Refresh to update the information displayed under Simulation Progress. Click Stop to stop the process of running the simulation. Click Delete to complete the work item and delete its results. Click Re-Execute to run the simulation again. When the status is Resolved-Completed, go to the Report tab to view proposition reports. If the interaction rule is configured to write results to VBD, you can perform analysis by clicking the link that opens the VBD planner (page 31). Preview Data The data preview options allow you to define the criteria used when previewing the data retrieved by the input definition: Sample: select this option to preview a sample of the data for a number of customers. One Customer: select this option to preview the data for a specific customer. Preview Size: size of the data sample in number of records when the option is Sample. Partition Key: displays the field used for partitioning when the option is Sample. Customer ID: provide the ID of the customer when the option is One Customer. Fetch for Strategy: only retrieve the inputs used by the strategy. Browse Reports As a strategy designer in the business sandbox, or as a supervisor or administrator in production, you can view reports to monitor the integrity of the application and analyze results. The default reports allow you to analyze the information captured for adaptive models, propositions and interactions. Adaptive Decisioning Active predictors: displays the list of active predictors. All models: displays the list of adaptive models. All models performance: displays model performance statistics. List of all models based on a particular proposition: displays the list of adaptive models for a given proposition. Proposition analysis based on success rate: displays the proposition success rate corresponding to models for the last 30 days. Interaction History Acceptance rate: percentage based report displaying the acceptance per proposition group. Channel distribution: displays the distribution of channels selected by the strategy, regardless of rank. Interaction History: displays the interaction records (page 91) for the last 30 days. Proposition distribution: displays the distribution of propositions selected by the strategy, regardless of rank. Simulation Reports Every successful simulation run generates standard proposition, segment and channel centric reports. The reports are available through the Report tab of the Simulation Execution Progress form, or through 36

39 the Browse Reports link in the Decision Manager portal navigation. You can also define your own reports by using the New button to create a new report definition in the class created by the output definition. Distribution of the top ranked propositions selected by the strategy. Average priority of propositions selected by the strategy. Recurrence of propositions considered by the strategy and corresponding rank. Average priority in the proposition rank report. Distribution of segments selected by the strategy. Distribution of all propositions selected by the strategy, regardless of rank. Distribution of channels considered by the strategy, regardless of rank. Distribution of the second highest ranked proposition selected by the strategy. Distribution of the third highest ranked proposition selected by the strategy. Plan Strategy designers have access to the calendar that provides a view of scheduled tasks assigned to them. 37

40 Business Rules Changes to one or more business rules (page 38) are required when working on a change request. The change request defines the instances the strategy designer can modify in order to be able to complete the work required for it to be submitted to the revision manager. Business Rule Types Decision tables (page 38) Decision trees (page 41) Map values (page 45) Scorecards (page 49) Predictive models (page 52) Decision parameters (page 55) Strategies (page 58) When opening a business rule included in the context of a change request, the header allows you to: Save: save changes. Discard: discard changes. Actions Run: in the majority of business rule types, you can use the run action to test your changes by providing inputs and analyzing the result. For example, you change the segments of a decision table that uses customer age, and you can enter age values to test if you logic places the customer in the correct segment. Refresh: refresh the rule. Any changes you have not saved are lost when you use this action. Get Help: read the documentation provided for the rule type. Close: close the rule. Decision Tables About decision tables (page 39) Defining the decision (page 39) 38

41 Defining results (page 39) Using decision tables in strategies (page 40) About Decision Tables Decision tables are used in Decision Management to bring classification to the strategy in the form of segments that define a result for conditions applied on characteristics of the data you use to determine the logic. In strategies, you use the decision table through the decision table component (page 73). You can then leverage the possible results defined in the decision table to combine with components that filter or split the decision path. Define Table It is in the Table tab that the decision logic is defined. The toolbar provide the controls to add the necessary columns and rows, analyze conflicts and completeness. The header of each column is mapped to a property. In the column cells, you apply values (or expressions) in the cells to define the criteria for the return results defined in the Actions column. If you want to provide upfront validation for the allowed results that can be used in the Actions column, you have to configure first the Results tab. The following example describes how you can define the decision table provided in the introduction to decision tables (page 39). This decision table works with two properties (Active and ValidStartDate), and two results (Eligible and Not Eligible). 1. The first property to the added to the first Conditions column is the Active property that inputs two possible values: true, or false. The first cell for the value of this property is true. 2. With the cursor in one of the cells of the first Conditions column, click the add column icon to add the next column. 3. Map this column to the ValidStartDate property, a property of date time type. In the label, you can override the default value (in this case, the label is the property name containing spaces). The reason for adding this property is to provide the decision with the information on whether the proposition's is still valid. To achieve this, compare the current date and time to the start date of the proposition by using expression. 4. Return Eligible if both conditions are true, otherwise Not Eligible. Configure Results It is in the Results tab that validation and processing criteria can be defined. The use of this tab is not required to design valid decision tables, but it provides powerful mechanisms to enforce consistency, define additional processing instructions and restrict decision table functionality. Allowed Results: this section allows you to validate the possible results through a property. Leave this setting blank if this is not the case for your decision table. Additional Allowed Results: the main function of this section is to provide a number of possible results that validate the return values defined in the Table tab. Another function of the settings defined in this tab is the ability to conditionally set property values depending on a given result. The following example illustrates both uses, although the main function in this example is the ability to set a property (the IsEligible property) at decision time. Add two additional allowed results rows, one for Eligible, and the other for Not Eligible. 39

42 For each allowed result, set the IsEligible property to true (Eligible) or false (Not Eligible). Preset Properties: this section allows you to define a list of property updates to be applied before evaluating the logic defined in the Table tab. The order of the rows is important as the list is executed in sequence. Decision Tables in Strategies To illustrate the use of decision tables in strategies, we provide an example where a decision table is used to evaluate the eligibility criteria in a product offers strategy. This strategy defines eligible propositions and prioritize these offers by marketing weight and initial base propensity. The Eligibility decision table component references the decision table designed to set the eligibility criteria, which is read by the Eligible Propositions filter component to exclude any product offer that is not valid in the context of this recommendation. 40

43 Decision Trees About decision trees (page 41) Configuring decision (page 41) Defining decision (page 43) Using decision trees in strategies (page 44) About Decision Trees Decision trees allow you to perform tests based on if conditions and define the instruction to apply when the criteria for given condition is met. In Decision Management, you can use decision trees to bring classification to the strategy in the form of segments that define a result for conditions applied on characteristics of the data you use to determine the logic. In strategies, you use the decision tree through the decision tree component (page 73). You can then leverage the possible results defined in the decision tree to combine with components that filter or split the decision path. Configure Decision In the Configuration tab, define the validation and processing criteria. The use of this tab is not required to design valid decision trees, but it provides powerful mechanisms to enforce consistency and restrict decision tree functionality. Options: these settings influence the capability of your decision tree and restrict how it can be used. Allow changes to function list: this instance allows changes to functions in the Decision tab, and not only constant vales. Without specifying the list of functions available in the design of the decision, all functions are allowed. Allow adding of nodes to the decision tree: this instance allows adding entries in the Decision tab. Allow selection of 'evaluate property' option: this instance is configured to evaluate properties. This option adds the evaluate instruction to the list of default instructions (return, continue and otherwise). If you enable this option, define which property can be evaluated, and the label to display in the Decision tab. Allow selection of 'call decision' option: this instance can call map values, decision tables and other decision trees. This option adds the call Map Value, call Decision Table and call Decision Tree instructions to the list of default instructions. 41

44 Allow selection of additional return actions: this instance allows operators to take additional actions. This option adds the take action instruction to the list of default instructions. When using the take action instruction in the Decision tab, you can define the list of actions for each entry by pressing Actions and defining the list of instructions (for example, set property) that apply when the logic applied in the entry applies. Result: provide result validation and processing instruction to enforce the classification in the decision tree value logic. Results defined by property: the possible results can be defined through a property. Leave this setting blank if the possible results are defined through the decision tree. Additional Allowed Results: provide allowed results to enforce the classification in the decision tree logic. If the results are also defined by a property, these results define additional possible results. For each result, you can set properties by referencing the property and defining the value to set when the decision tree logic returns the specified allowed result. Preset Values: this section allows you to define a list of property updates to be applied before evaluating the logic defined in the Decision tab. The order of the rows is important as the list is executed in sequence. In the example used to illustrate the decision tree functionality, the instance is configured to allow adding entries and setting properties. The Decision tab is limited to a predefined value (High, Medium, or Low). For each possible result, add three entries in the Additional Allowed Results for each possible result. In this example, the properties are set per entry in the Decision tab, and not through the Target property setting for each allowed result. 42

45 The example you see above relies on a combination of two specific decision tree options. If you need to restrict decision tree operations, you could configure the Options settings as follows: The most basic decision tree configuration only allows you to select a property, compare it to a value and return a result. This is the way you can work with decision trees with Allowed To Add Nodes as the only enabled setting under Options. To build the example provided to illustrate the use of decision trees in Decision Management, set property values by using the Target property and Value fields in the Additional Allowed Results section. Define Decision It is in the Decision tab that the instance's logic is defined. Each entry is a condition that returns a result and, at the end, the otherwise condition catches the remaining cases. To complete the example that illustrates the decision tree functionality, the RepayPattern property set for each condition has three possible values: Same day, Same week or Same month. For each value, the RiskFactor property is set accordingly. 1. Add one entry to define if the repay pattern in "Same week". Return Medium and press Take Actions to set the risk factor to 2. 43

46 2. Add another row and repeat the same steps for the "Same day" value. Return Low and press Take Actions to set the risk factor to In the otherwise condition, return High and press Take Actions to set the risk factor to 3. Setting properties is not an action you could perform in a basic decision tree. In that case, you can only set the properties in the Configuration tab, as described in alternative in restricted decision tree operations (page 43). Decision Trees in Strategies To illustrate the use of decision trees to return a result that can be used for characteristic based classification in Decision Management, we provide an example of a decision tree that decides on the result based on the possible values of the RepayPattern property, and sets the value of another depending on the result (RiskFactor). The map value is used in a centralized classification strategy that can be reused in other strategies as a reusable piece of logic since it is enabled with the external input setting (page 88). Set property components define the initial budget for the medium, high and low risk segments. 44

47 The decision tree is imported through the Risk Classification decision tree component: Map Values About map values (page 45) Configuring map value (page 46) Defining the matrix (page 47) Using map values in strategies (page 48) About Map Values Map values allow you to create matrices that define segmentation based on two vectors. The vectors (row and column) contain the criteria that applies to the input properties and, based on business logic, you apply possible results. In strategies, you use the decision tree through the map value component (page 74). You can then leverage the possible results defined in the map value to combine with components that filter or split the decision path. 45

48 Configure Map Value In the Configuration tab, define processing instructions, validation logic and the two vectors (row and column) used for segmentation. Security: these settings influence delegation restrictions. Allow updating of the matrix configuration in delegated rules: delegation allows operators to update the settings in this tab. Allow use of the expression builder on the matrix view: the expression builder can be used in the Matrix tab (page 47). Input Rows and Input Columns: define the two vectors against which conditions are applied to. Label: define how to display the row/column name. Source: define if the source is a property or a custom value of the type selected in the Data Type drop-down (none, string, integer, double, boolean, date, date time, or decimal). Configure rows and Configure columns: configure the matrix by adding as many rows and columns as necessary to define the criteria. Results: provide result validation and processing instruction to enforce the classification in the map value logic. Results defined by property: the possible results can be defined through a property. Leave this setting blank if the possible results are defined through the map value. Additional Allowed Results: provide allowed results to enforce the classification in the map value logic. If the results are also defined by a property, these results define additional possible results. For each result, you can set properties by referencing the property and defining the value to set when the map value logic returns the specified allowed result. Preset Values: this section allows you to define a list of property updates to be applied before evaluating the logic defined in the Matrix tab. The order of the rows is important as the list is executed in sequence. In the example used to illustrate the map values, the matrix is based on age and salary conditions: Input Row: Property:.Age Label: Age Rows: <= 25 <= 35 <= 45 <= 55 Input Column: Property:.Salary Label: Salary Columns: <= <= <= > Five possible results (High, Low, Medium High, Medium Low or Medium) apply, setting the RiskFactor property accordingly (5, 3 or 1 ). 46

49 Define Matrix The Matrix tab provides the mechanism to define segmentation. The two dimensions of the segmentation (row and column) depend on the row and column properties defined in the map value configuration (page 46). Add as many rows and columns as necessary to set the segment for cases that meet the defined criteria. To complete the example used to illustrate map values, the matrix selects from the allowed results (page 46) for a combination of age and salary characteristics. The Default column and row values apply in cases that do not meet the criteria, or when there is missing data. 47

50 High risk conditions: Regardless of age, if yearly salary is less than or equal to If age is between 46 and 55, and yearly salary is less than or equal to Medium high risk conditions: If age is between 26 and 45, and yearly salary is higher than but less than or equal to Default if yearly salary is higher than but less than or equal to 50000, and age does not meet the defined criteria. Default if age is less than or equal to 25, and salary does not meet the defined criteria. Default if age and salary do not meet the defined criteria. Medium risk conditions: If age is less than or equal to 25, and yearly salary is higher than but less than or equal to If age is between 36 and 45, and yearly salary is higher than but less than or equal to Default if yearly salary is higher than but less than or equal to 50000, and age does not meet the defined criteria. Default if age is between 36 and 45, and salary does not meet the defined criteria. Medium low risk conditions: If age is between 26 and 35, and yearly salary is higher than but less than or equal to If age is between 46 and 55, and yearly salary is higher than Default if yearly salary is higher than but less than or equal to 70000, and age does not meet the defined criteria. Default if age is between 46 and 55, and salary does not meet the defined criteria. Default if age is between 26 and 35, and salary does not meet the defined criteria. Low risk conditions: If age is less than or equal to 45, and yearly salary is higher than If age is less than or equal to 25, and yearly salary is higher than but less than or equal to Default is yearly salary is higher than 70000, and age does not meet the defined criteria. Map Values in Strategies To illustrate the use of map values to return a result that can be used for characteristic based classification in Decision Management, we provide an example of a map value that decides on the 48

51 result based on risk when combining age and salary criteria. The map value is used in a centralized classification strategy that can be reused in other strategies as a reusable piece of logic since it is enabled with the external input setting (page 88). Set property components define the initial budget for the medium, high and low risk segments. The map value is imported through the Risk Classification map value component: Scorecards About scorecards (page 49) Defining score calculation (page 50) Defining results (page 51) About Scorecards Scorecards create segmentation based on one or more conditions, and a combining method. The score based segmentation can be mapped to results by defining cutoff values used to map a given score range to a result. For example, your application can use a scorecard to calculate customer segmentation based on age and income, and then map particular score ranges to defined results. The output of a scorecard is a score and segments defined by the results. Scorecards are used in strategies through scorecard 49

52 components (page 70). In flows, scorecards are referenced through the decision shape by selecting the scorecard model type. Define Score Calculation Define the predictors by adding properties, determining how to calculate the score and assigning a weight. 1. In the Scorecard tab, use the Combiner Function drop-down to select the method for combining the score. Combiner Function Description SUM AVERAGE Combines by total sum of score values between predictors. Combines by total sum of score values between predictors divided by the number of predictors. MIN Combines the score for each predictor and takes the lowest value. MAX Combines the score for each predictor and takes the highest value. 2. The grid under the Combiner Function drop-down allows you to define the properties, conditions, score and weight attributed to cases matching the conditions. In the Predictor Expression column, use the down arrow to select existing single value properties. Predictors can also be determined through expressions (for == "Employed",.Salary,.CreditAmount). In the Condition column, define the criteria for matching the predictor value to a score. Add as many conditions as necessary. In the Score column, enter the score for cases falling in the defined condition. You can define the score explicitly (for example, 20), through a property (for example,.score) or a computation expressed by an expression (for example,.score*.penaltymargin Optionally, define a fallback score for any case that does not match the previous conditions in the Otherwise row. In the Weight column, define the coefficient of the predictor. By default, every predictor is assigned the same weight (1). Changing the default value results in calculating the final score as weight multiplied by score (for example, 0.5*30). Maintaining the default value implies that, effectively, only score is considered because the coefficient is 1 (for example, 1*30). 50

53 Add as many properties as necessary to segment your customer base, and repeat the process specified in the previous steps. 3. Map scores to results (page 51). Define Results After defining the way scores should be calculated (page 50) and the conditions for cases falling in a given score, map score ranges to results by defining the cutoff value. 1. Go to the Results tab. 2. Information is provided about the score range. The maximum and minimum scores depend on the combiner function selected in the Scorecard tab. If you use expressions to calculate the score, minimum and maximum scores display as 'unknown' because they can not be calculated. 3. Check the Audit Notes option if you want scorecard execution details captured in the work item's history. 4. In the Result column, enter the action for the score range specified in the Cutoff Value column. 5. In the Cutoff Value column, define the score range according to the minimum and maximum score the scorecard calculates. 6. As you add more actions, score ranges are defined top-down and automatically calculated based on the cutoff value defined in the previous result. 51

54 Predictive Models About predictive models (page 52) Defining predictive model results (page 52) Viewing statistics (page 54) Using predictive models in strategies (page 55) About Predictive Models Predictive models predict behavior for one or more segments (classes) based on customer data. Predictive models instances use a PAD model or a third party model in PMML (page 91) format (3.0, 3.1, 3.2, 4.0 or 4.1). Predictive models are used in strategies through predictive model components (page 70). In flows, predictive models are referenced through the decision shape by selecting the predictive model type. Define Predictive Model Results The segmentation provided by the predictive model needs to be assigned to actions applying to a given class or class range. Defining predictive models depends on the type of model: PAD model (page 52) PMML model (page 54) PAD Model PAD models are constructed to generate the largest possible number of classes (segments) that exhibit predicted behavior, steadily increasing as the class number increases. However, business strategies translate to the two or three alternatives typically associated with the probability of predicted behavior (high, medium and low). Remapping the classification defined in the predictive model to the smaller number of business strategies allows you to increase the quality of business. For example, if a lower propensity (page 92) class is reassigned to the medium propensity class where fewer customers are presented with a product offer but a greater proportion responds, although the volume of business decreases, the quality increases. Graphical representation (page 52) Mapping segments (page 53) Graphical Representation Graphical representation is provided for predictive model results. Two graphs side by side shows the score distribution (page 92). 52

55 View Parameter: drop-down that allows you to selected the model output used to plot data. Use the data points to inspect the values. X-axis: visually represents the Classification. If you combine classes, they are also aggregated in the chart. Y-axis: visually represents the Percentage. The graph on the left allows is based on the non-aggregated classification; this is also the graph displayed in the Statistics (page 54) tab. As you aggregate classes, the graph on the right displays aggregated statistics according to the number of results. Map Segments Map the classes output by the model to decision results. To understand the effect of combining the different classes to create predictive based segmentation, examine the grouped statistics in the graph displayed on the right. If the original PAD model does not contain aggregation/grouping statistics information, N/A is displayed. 1. Go to the Results tab of the predictive model instance. 2. By default, this tab displays the total number of classes corresponding to the segmentation created by the predictive model. 3. You aggregate classes by clicking by placing your cursor between the bars corresponding to the class, and clicking so that classes up to that one can be binned under the same result. A red bar signals the class aggregation. Click Apply to reflect the changes, and the segments in the Result column. The graph on the right reflects the recalculated class distribution. Use Reset to set the class aggregation back to default so that you can restart the process of aggregating classes. 53

56 PMML Model The results of a PMML model consist of model output fields. In this tab, you select the output field that provides the classification result. Typically, this is the model's default outcome field. Statistics The Statistics tab displays statistical information corresponding to the classification in the PAD model. This tab is not present if the predictive model is constructed based on a third party model. Similar to the Results tab, it shows a graphical representation (page 52). The main difference between the two graphical views is that the one in the Statistics tab does not reflect the class aggregation. This tab also displays the number of classes, and PAD model attributes. 54

57 Predictive Models in Strategies Predictive models are used in strategies through the predictive model component (page 71). For example, a Next Best Action strategy importing the sales and retention strategies, and selecting the decision path based on the likelihood of customer attrition predicted by the predictive model rule. The switch component selects Retention if the result of the predictive model is High, and Sales in any other case. Decision Parameters About decision parameters (page 55) Using decision parameters in strategies (page 56) Changing decision parameters (page 57) About Decision Parameters Decision parameters offer a flexible mechanism for the type of input values that require frequent changes without having to adjust the strategy. Changes to the values of decision parameters become directly available when you update the instance. Decision parameters can assume the form of a simple list of 55

58 values (typically, this is the case with global decision parameters), or a set of values that are available in a specific context (for example, proposition parameters and channel centric parameters). Decision parameters are used in strategies through the decision parameters component (page 77). The values of decision parameters are typically defined by business users through the Decision Manager portal, but this functionality is not tied to the facilities in the portal and can be used in Designer Studio as well. The content of decision parameters rules is defined by the extension points that system architects use to configure the data model and user interface supporting decision parameters. Decision Parameters in Strategies Decision parameters are used in strategies through the decision parameters component (page 70). The use of decision parameters allow for values that influence the decision to be independently adjusted according to business priorities. In revision based on direct deployment overlays, this means that revision managers can directly change the values of decision parameters without the need for system administrators to deploy a new revision through Designer Studio. Decision parameters can influence the decision at the application (global) level. For example, a Next Best Action strategy that selects the best offer based on the call reason, using global parameters that set maximum, minimum and target budget ratio, plus a global flag to determine if agents can offer sales propositions. Decision can also be scoped, and influence the decision by providing a flexible mechanism to set specific proposition parameters. For example, a product offers strategy that defines eligible propositions and prioritizes offers by marketing weight and initial base propensity. Marketing weight and initial base propensity are proposition parameters (in this example, propositions in the Phones and Tablets groups). 56

59 Change Decision Parameters You can change the values of decision parameters through the Decision Parameters tab. You validate the changes by saving the decision parameters rule and running the strategy, or running the flow a flow designed to test the changes. Checking in the changes makes the changes available to all users but, typically, the changes to decision parameters are made available when system architects activate the revision that contains the corresponding instances, or when revision managers activate a direct deployment revision. Decision parameters can define a list that applies to the entire application. Decision parameters can also define a set of values that are applicable in a specific context. This is the case for parameters that apply to propositions in a specific business issue or group, or channel centric parameters. 57

60 Strategies About strategies (page 58) Designing strategies (page 62) Auto-run results (page 88) About Strategies Typically, you design a strategy (page 93) to deliver a personalized recommendation for a single decision (page 90). For example, a strategy to recommend the most important issue to be dealt with for a particular customer, via a channel or system, and at a given point in time. Combined with the current objectives and priorities of the company, predicted risks and interests are part of the strategy. A recommendation can be part of a sequence. After determining the most important issue to address, the decision chain may need to address which credit strategy to use, which retention strategy or which product to offer first. Every decision employs a combination of strategy components (page 63) that define the underlying logic required to deliver a recommendation. Components allow you to create personalized customer interactions consistently across contact channels. The advantage of building decision strategies from these smaller components is that each one can be readily understood, developed, edited and tracked on an ongoing basis. You can use components to model sophisticated customer behavior and there are some common design patterns that you end up reusing frequently. In the context of using strategies in combination with propositions, you create a strategy to deliver the decision for one business issue or group. The scope in the proposition hierarchy corresponds to the issue or group level of the action dimension. The level at which you create the strategy determines the properties it can access, and these are the properties that define the output structure of components in the strategy. You can develop decision strategies as a self-contained single strategy, or multiple strategies combined using sub strategy components. Combining strategies allows for concurrent development of large scale strategies by creating smaller strategies that can be developed in a relatively independent manner. The other use case of multiple strategies is reusing a logical pattern. 58

61 The result of a strategy is a page (clipboard or virtual list) containing at least the results of the components that make up its output definition. Related Topics Strategy Design Patterns (page 59) Strategy Execution (page 61) Strategy Design Patterns Next Best Action (page 59) Capture results (page 60) Segmentation (page 60) Reusable logic (page 61) Next Best Action The standard approach for finding the Next Best Action (page 91) for each customer consists of segmenting customers, assessing the propensity, selecting the action for each customer segment and, finally, selecting the best decision path. The following list describes a sequence that can be used as a starting point when planning your strategy. 1. Plan the final decision and work backwards. Starting point that allows you to define the strategy plan(s), such as the most important issue to address, what drives the decision, the most appropriate proposition (and how to determine it), the probability factors, characteristics and preferences to take into account in the decision. a. What do you want to deliver? b. What action to take in order to achieve this? c. What data is required? 2. Build from customer, product, environment and other required information to deliver the decision. a. Define propositions. b. Import propositions. c. Prioritize between propositions. d. Balance issues. e. Finalize decision. The visual orientation of the strategy is a logical translation of the output orientation working backwards from the Next Best Action end point. Structurally, this can be explained by using a top-down tree model. For example, assume that you need to build a strategy that addresses the following: A number of segmentation components are available that classify customers based on product and risk of customer attrition. Different issues need to be addressed, such as sales, recruitment and retention. Arbitration between the different propositions is done with a Next Best Offer (page 91) prioritization: In the sales context, the offer that has the highest cross sell score. In the risk of customer attrition context, the offer that addresses cases falling in segments with the highest customer attrition risk. Depending on the issue to be addressed, a final recommendation needs to be issued. The diagram below visualizes the concepts used when planning the strategy that provides the decision. A strategy implementing the logical structure abstraction is the final result. The fundamental NBA pattern starts from the final decision point and has a right-to-left orientation, but the the flow of the arrows starts with import components (page 67), then segmentation components (page 70) for which possible actions are defined, next the data enrichment components (page 75), proceeding with arbitration components (page 81) and, finally, the end selection component (page 84) that delivers the best action in the interaction. 59

62 Capture Results Starting with PRPC 7.1, the information necessary to capture the interaction results is delivered through the strategy, not the interaction rule. The minimal design of a strategy for capturing interaction results consists of a data import component that provides the page containing the results of running the strategy that delivers the decision and provides the necessary information for writing the interaction results to the Interaction History tables. Segmentation Split components provide strategies with a mechanism to derive segments from results created by decision analytics and business rules components. Combined with the exclusion component, you can create powerful segmentation trees that assign results based on a percentage of cases, and then conditionally exclude results based on relevant conditions. 60

63 Reusable Logic Through the external input strategy setting (page 88), you design strategies as centralized pieces of logic that can be used by more strategies. The strategy referred to by the sub strategy component has the external input option switched on. This external input connects to the starting components that define the reusable chain of components. In another strategy, the sub strategy component refers to the reusable strategy and it is driven by other components. When you run this strategy, the sub strategy component effectively results in replacing the component with the chain of components that are propagated by the sub strategy. Strategy Execution Strategy execution is performed in the opposite direction of the dependency chain represented by the gray arrows, taking the last component, recursively executing the dependent components and calling out the components whose configuration is tied to other decision instances, data references reading data records and named pages or properties from a page depending on the Applies To class of the strategy. In general, components that reference a rule or a page are subject to auto-mapping, which means that properties with the same name in the referenced rule/page and in the data class defined for the strategy are automatically mapped even if not explicitly mapped through components. The data class can be the 61

64 strategy result class defined for the strategy or the class corresponding to the scope of the strategy in the proposition hierarchy. Typically, the last component is a selection component that, through the Results component, is exposed to other rules using the strategy. Components whose configuration is tied to other rules are components in the prediction/segmentation category and data import components. Each component creates its own page list from which the embedded pages are of the class the strategy properties belong to. This mechanism allows you to acquire and enrich data. The result of executing a strategy can be a single result or a list. List processing can be implemented by importing a set of propositions by group or by combining data. The only components that do not combine data are champion challenger and switch components. Strategy Design A strategy is defined by the relationships of the components that are used in the interaction (page 91) that delivers the decision (page 90). The Strategy tab provides the facilities to design the logic delivered by the strategy. Toolbar and context menu (page 62) Defining components (page 63) Connecting components (page 86) Defining expressions (page 87) Enabling external inputs (page 88) Defining the results of a strategy (page 88) Toolbar & Context Menu The strategy toolbar in the Strategy tab displays buttons that correspond to the same functionality as provided through the editing of flows with Modeler. The Auto-Run on/off button allows you to turn on and off the strategy's auto-run mode (page 88). You can access the context menu by right clicking the working area without selecting any component. The context menu allows you to add strategy components (page 63), select all components, enable and disable external inputs (page 88), annotate your strategy in the same way as you would do in a process flow and use the zoom options. 62

65 The right click mouse action after selecting a component allows you to open the component's Properties dialog, use the Open <RuleType> action to open the referenced rule (adaptive model, decision tree, decision table, map value, predictive model, scorecard, strategy, decision parameters, segment, geofence or contact policy), use the Delete action to remove the component, and use the Copy, Cut and Paste actions in the same strategy or across strategies. A paste action results in clearing the source components, but it retains expressions contained in the original component's definition. Strategy Components Editing components is done via the properties dialog of the selected component. This dialog is displayed when you double click the component, or when you right click the component and select Properties from 63

66 the context menu. A component's properties dialog consists of elements common to all components and tabs specific to the type of component. General component settings (page 64) Sub strategy components (page 66) Import components (page 67) Business rules and decision analytics components (page 70) Enrichment components (page 75) Aggregation components (page 78) Arbitration components (page 81) Selection components (page 84) General Settings Name and description (page 64) Source components (page 64) Properties mapping (page 65) Pages and alternative pages (page 65) Name & Description Every component is assigned a default generated name. The Name field allows you to change the generated name to a meaningful name in the context of your strategy. This field defines the component identifier and supports defining names containing space characters. Below the Name field, Component ID displays the actual component name in the clipboard. The actual component name is the user-defined name excluding space characters. This is also the name used to refer to components in an expression. The Description options allow you to define how to handle the description of the component. Use generated: displays information based on the component's configuration. Use custom: allows you to define a description that will be displayed in the component's summary. Source Components The Source Components tab applies to most components. This tab displays the components that connect to the current component. The order can be changed by dragging the row up or down. 64

67 Properties Mapping Some components allow you to map the properties brought to the strategy by the component to properties available to the strategy. This is done through the Properties tab (group by and financial calculation), Output Mapping tab (adaptive model and predictive model), Score Mapping tab (scorecard), Properties Mapping tab (data join, decision parameters and data import) or Interaction History tab (interaction history and proposition data). Pages & Alternative Pages You can supply data to components that reference an external rule instance (predictive model, scorecard, adaptive model, decision tree, decision table, map value, sub strategy, and decision parameters). This is particularly useful if you want to drive the component results using customer data. This capability requires a specific set up for the referenced instances and the strategy referencing them: In the referenced rule instance, the data is included in the rule instance's pages and classes. In this example, a predictive model defines a page named Purchase, which is mapped to the Purchase data class: Pages from the reference rule instance's pages and classes are listed under Available pages & classes in the strategy component. If the Supply with data check box is enabled, data passed by the page is used to evaluate and execute the component. 65

68 It is also possible to provide an alternative page. If the alternative page data is not available, it falls back to the originally set page (in the example above, Purchase). Sub Strategy Components A strategy (page 93) can use other strategies through sub strategy components. Including strategies allows for using specialized group or issue level strategies that address a specific business case, and combining them in a more generic strategy that is typically at the top level class in the proposition hierarchy. The strategy design pattern used when including sub strategies can be seen as always including more specialized cases to address all issues in an NBA strategy. Sub strategy components reference other strategies. They define the way two strategies are related to each other, access the components in the strategy they refer to, and determine how to run the strategy if it is in another class. Basically, a sub strategy component defines which strategy to import and, explicitly defined, the decision component (page 88). This is accomplished in the Source tab through configuring the strategy and, if applicable, the component. Additionally, you also define how to run the imported strategy. A sub strategy component can represent a reusable piece of logic provided the strategy it refers to has the external input (page 88) option enabled, and the sub strategy component itself is driven by other components. Current Page: the imported strategy runs on the class the strategy belong to. 66

69 Another Page: the imported strategy runs on a defined page. Field Description Page Define the page. When this field points to a page group or list, the decision making process is iterated over as many times as defined in the page group or list. For example, if a strategy runs through a sub strategy component over a list containing two customers, and assuming the strategy outputs three offers, the sub strategy component results in a list containing six offers. After defining the page, the Class field displays the corresponding class. Select the strategy and, if applicable, the strategy component. Class Strategy Import Category Components in this category bring data into the current strategy. Data import (page 67) Interaction history (page 68) Proposition data (page 69) Data Import Components Data import components import data from pages available to the strategy. In the Source tab, use the Smart Prompt to select the page. Data import components that refer to named or embedded pages can map the page's single value properties to strategy properties through the Properties Mapping tab. 67

70 Data import components defined in releases previous to PRPC 7.1 were subject to automapping. That is still the case, but the mapping by matching name between target and source is implicitly done when the strategy is executed. You only have to explicitly map properties if exact name matching can not be applied or you want to override the implicit target/source mapping. If using named pages, these pages have to be present in the strategy's Pages & Classes. Interaction History Components Interaction history components import the results stored in Interaction History for a subject ID. In the Interaction History tab, use the filter settings to add time criteria, conditions based on Interaction History properties and specify the properties that should be retrieved. If you do not define any conditions or properties, the component retrieves all results for the subject ID. Defining criteria reduces the amount of information brought to the strategy by this component. Some properties are always retrieved by the interaction history component (for example, subject ID, fact ID and proposition identifier). 68

71 Database limitations related to data type changes apply if you are filtering on days. This setting is not suitable if you are working with dates earlier than January 1, Proposition Data Components Proposition data components import propositions (page 92) defined in the proposition hierarchy. In the Proposition Data tab, use the proposition hierarchy to define which propositions to import. Use the Business issue drop-down to select the issue. In the Group/Proposition drop-down lists, you can either use the Import All option or specify a group/proposition. The configuration in this tab is directly related to the level of the strategy in terms of the proposition hierarchy (business issue and group). In the Interaction History tab, check the Enable Interaction History option to bring results stored in Interaction History to the strategy as specified in the conditions and properties settings. The settings 69

72 defined in this tab are similar to the interaction history component (page 68) but, unlike the interaction history component, the component only retrieves results for the subject ID if you define which properties to use. Business Rules & Decision Analytics Categories Components in the business rules and decision analytics categories typically use customer data to segment cases based on characteristics and predicted behavior and place each case in a segment (page 93) or score (page 92). Decision analytics (page 71) Business rules (page 73) Some common configuration applies to these components. Select if the component should be defined on the Applies To class or the Strategy Result class for predictive model, scorecard, decision tree, decision table and map value components. Applies To: the component is evaluated one time on the primary page of the current strategy. Strategy Results: the component is evaluated on every incoming step page. Predictive model and adaptive model components map the output of the corresponding instance to strategy properties through the Output Mapping tab. In the case of scorecard components, this is done through the Score Mapping tab. Select the instance in the name field. Depending on the type of component, select a predictive model, scorecard, adaptive model, decision parameters, decision table, decision tree or map value instance. Use the SmartPrompt to select an existing instance. Adaptive models, decision tables, decision trees and map values allow for defining parameters. When the instances are on the Applies To class, the Properties dialog displays the parameter values in the Define Parameters section. Through segment filtering connections (page 86), you can create segmentation trees. For example, you start by defining a strategy path for cases falling in the accept segment and another one for cases falling in the reject segment. 70

73 Decision Analytics Category Predictive model (page 71) Scorecard (page 71) Adaptive model (page 72) Predictive Model Components Predictive model components reference predictive model instances (page 52). Scorecard Components Scorecard components reference scorecard instances (page 49). 71

74 Adaptive Model Components Adaptive model components provide segmentation based on adaptive models in ADM. These components reference adaptive model instances. The Adaptive Model tab is displayed for this type of segmentation components. Select the adaptive model instance. Since the scope in the proposition hierarchy is propagated through proposition data components, if proposition data components connect to the adaptive model component, the Adaptive Model field is the only setting available in the Adaptive Model tab. If other components (typically, proposition data) do not connect to the adaptive model, set the remaining fields according to what the scoring model created in ADM is going to model. Use the Business Issue, Group and Proposition fields to select the hierarchy and proposition defined when managing propositions (page 24). Depending on the scope the strategy was added to, business issue and group can be predefined. 72

75 Use the Direction and Channel fields to select values defined in the action dimension. Business Rules Category Decision table (page 73) Decision tree (page 73) Map value (page 74) Split (page 74) Decision Table Components Decision table components reference decision tables used to implement characteristic based segmentation by referencing a decision table that takes into account customer data to segment on a given trait (for example, salary, age and mortgage). Decision Tree Components Decision tree components reference decision trees. Decision trees can often be used for the same purpose as decision tables. 73

76 Map Value Components Map value components reference map values that use a multidimensional table to derive a result. For example, a matrix that allocates customers to a segment based on age and salary. Split Components Split components branch the decision results according to the percentage of cases the result should cover. These components are typically used to build traditional segmentation trees in strategies, allowing you to derive segments based on the standard segments defined by the results of other components 74

77 in the business rules and decision analytics category. You define the result (pxsegment) and the percentage of cases to assign to that result. Enrichment Category Components in this category add information and value to strategies. Set property (page 75) Data join (page 76) Decision parameters (page 77) Set Property Components Set property components enrich data by adding information to other components, allowing you to define personalized data to be delivered when issuing a decision. Personalized data often depends on segmentation components (page 70) and includes definitions used in the process of creating and controlling a personalized interaction, such as: Instructions for the channel system or product/service propositions (page 92) to be offered including customized scripts, incentives, bonus, channel, revenue and cost information. Probabilities of subsequent behavior or other variable element. These components enrich data through the Target tab. Use this tab to add comments and set strategy properties for which you want to define default values. Comments can be defined through adding rows, setting the Action drop-down to Comment and entering the appropriate comment. Properties can be set through adding rows, setting the Action drop-down to Set and mapping the properties in Target and Source. 75

78 Set property components created with releases previous to PRPC 6.2 SP2 DSM Edition supported overriding property values in the strategy through the Overrides tab, but this functionality has been removed since then. Strategy components defined in this way will show the configuration in the Overrides tab if there was a previous configuration, but no changes can be made. Data Join Components Data join components import data in an embedded page, named page or strategy component and map strategy properties to properties from the page or component. Data join components enrich data through the Join and the Properties Mapping tabs. These type of component can be used to join lists of values; for example, a data join component that has one or more components as source and uses the results of another strategy component to define the join conditions. In the Join tab of this component: Use the Type drop-down to select the type of data: Pages or Component. In the Name field, select the page or component. The Class field displays the class context. The criteria to match data is defined as one or more conditions in the Join When all Conditions are Met section. This is a value pair between properties in the data join component and, depending on what you selected in the Type field, properties in the page or strategy component. You can determine the amount of data to include in the exclude source component results check box, an option that, if enabled, results in performing an inner join operation. 76

79 In the Properties Mapping tab, map the properties you want to use in the component's decision path. Data join components defined in releases previous to PRPC 7.1 were subject to auto-mapping. This is no longer the case: you have to explicitly map the properties required for strategy execution purposes. Decision Parameters Components Decision parameters components import the data defined in decision parameters instances into the strategy. In the Decision Parameters tab, select the decision parameters instance. The when conditions allow you to match properties brought by the decision parameters instance and properties defined by the decision parameters component. The conditions can be provided by a property or an expression. 77

80 In the Properties Mapping tab, configure the mapping settings. The Define Mapping check box turns on/ off implicit mapping by name. Automatically mapped properties lists the properties that are subject to implicit mapping. For decision parameters properties that do not have an implicit counterpart among the strategy results (that is, name matching does not apply), you can explicitly map them by using the Enable additional mapping option. Aggregation Category Two components fall in this category. Group by (page 78) Iteration (page 79) Financial (page 80) Group By Components Group by components set strategy properties using an aggregation method applied to properties from the source components. The Properties tab of this component allows you to define the aggregation operations. So that you can use the results of a list of elements, the Group output rows by setting is available in this component. The properties that can be used to group the results are strategy properties; that is, properties of Data-pxStrategyResult and properties available to the strategy depending on its 78

81 applicability in the context of the proposition hierarchy. For example, selecting grouping by.pyname allows you to obtain the list of results for each proposition name. In the Aggregators section, select strategy properties in the Property column, the method for setting the property value based on an expression (SUM, COUNT, FIRST, MIN, MAX, AVERAGE, TRUE IF ANY, TRUE IF NONE, TRUE IF ALL or STDEV) and type the expression in the Source column. The properties that can be used in the Property column are strategy properties. The properties that can be used in the Source fields are properties of Strategy Result (SR) default properties, and properties available to the strategy depending on its applicability in the context of the proposition hierarchy. Properties that are not mapped in the component are automatically copied. In the For remaining properties setting, define how to handle the remaining properties by selecting one of the options from the drop-down. When using the options that copy with the highest/lowest value, specify which property in the SR class corresponding to the level of the strategy in the proposition hierarchy provides the value. First: copy with first value. None: empty. With highest: copy with highest value. With Lowest: copy with lowest value. Iteration Components Iteration components perform cumulative calculations based on the settings defined in the Parameters tab. Iteration components operate in two modes: 79

82 Without source components, you can define the properties, number of iterations and early stop conditions. The order of the properties is taken into account when performing the calculation. Depending on the setting used to control how to return the results, the component returns only the final calculation, or final calculation and intermediate results. With source components, the number of iterations equals the number of results in the source component. The result of running the iteration component contains the final calculation and no intermediate results. If the value of the arguments is set through source components, the order of the components in the Source tab is important because it is directly related to the order of arguments considered to perform the calculation. The settings you can use to define the iteration calculation consist of iteration settings, early stop conditions and results options: Iteration Settings Select the property for the set value action. Define the initial value for the set value action. Define the progression value for the set action. Define the maximum number of iterations in terms of results. Early Stop Conditions allow you to define conditions that apply before the maximum number of iterations. The conditions are expressed by the value of a property, the difference between the current and the previous value, or a combination of the two. In the Return settings, select if the component returns the last final calculation, or final and intermediate calculations. Financial Calculation Components Financial calculation components perform financial calculations using the following functions: 80

83 Internal rate of return calculates the internal rate of return for a series of cash flows. Modified internal rate of return calculates the modified internal rate of return for a series of periodic cash flows. Net present value calculates the net present value of an investment. The Properties tab of this component allows you to define the calculation and select properties that provide the arguments for each financial function. The arguments that you can select in the Target and Payments drop-down lists are strategy properties of type decimal, double or integer. If the value of the arguments is set through source components, the order of the source components is important because it is directly related to the order of arguments considered by the function to perform the financial calculation. Typically, the Payments argument should be a list of values and not a single value. So that you can use a list of values to provide the Payments argument, use a data import component to set properties that can be used by this component. Arbitration Category Components in this category filter, rank or sort the information from the source components. Enriched data representing equivalent alternatives is typically selected by prioritization components. Filter (page 82) Prioritize (page 82) Segment filter (page 83) Contact policy (page 83) Geofence filter (page 83) Segment filter, contact policy and geofence filter components are only available in a NextBest-Action Marketing (NBAM) application. 81

84 Filter Components Filter components apply a filter condition to the outputs of the source components. Filter components express the arbitration through the Filter Condition tab. The Filter condition field allows you to enter the expression used when filtering the results of the source components. Prioritize Components Prioritization components rank the components that connect to it based on the value of a strategy property or an expression. These components can be used to determine the service/product offer predicted to have the highest level of interest or profit. Prioritization components express the arbitration through the Prioritize tab. Two modes can be used to order the results: by priority or alphabetically. Each mode toggles its own specific settings. If Prioritize values is selected, Order by settings are displayed. If Sort alphabetical is selected, sort settings are displayed instead. 82

85 The Expression field is used to define properties providing prioritization criteria through an expression. The Output settings (Top and All) define how many results should be considered in the arbitration. The Top option considers the first results as specified in the field next to it and All considers all results. Segment Filter Components Segment filter components reference a segment rule (page 93), allowing for determining if a case falls in a given segment. The arbitration itself is expressed through the referenced rule. The segment rule is executed on customer data (the primary page of the strategy) and returns true if the case is part of the segment it represents. Segment filter components set the pxsegment property to the name of the referenced segment rule and also the pxrank property. If other components do not connect to it, the segment filter returns a list with a single row (the case is part of the segment) or an empty list (the case is not part of the segment). If there are components that connect to it, the segment filter returns all or no strategy results. Contact Policy Components Contact policy components reference a contact policy rule (page 90), allowing for determining if the customer should be contacted. As with the segment filter component, the arbitration itself is expressed through the referenced rule. Contact policy components typically have source components and return a subset of strategy results matching the criteria defined in the contact policy rule. The output options allow for refining the amount of results returned by the component. In case the order of the results is relevant, you need to prioritize them and provide that ordered input to the contact policy component. Geofence Filter Components Geofence filter components reference one or more geofence rules (page 91), allowing for determining if a customer has triggered a given geofence. Geofence filters typically have source components and return a subset of strategy results if a customer has triggered a given geofence based on the current 83

86 customer location. The customer location can be provided through properties representing the latitude and longitude or real-time events. Selection Category Strategies are balanced to determine the most important issue when interacting with a customer. The first step in applying this pattern is adding prioritization components (page 81) to filter the possible alternatives (for example, determining the most interesting proposition for a given customer). The second step is to balance company objectives by defining the conditions when one strategy should take precedence over another. This optimization can be accomplished by a champion challenger or a switch component that selects the decision path. Champion challenger (page 84) Switch (page 84) Exclusion (page 85) Champion Challenger Components Champion challenger components randomly allocate customers between two or more alternative components, thus allowing for testing the effectiveness of various alternatives. For example, you can specify that 70% of customers are offered product X and 30% are offered product Y. Champion challenger components express component selection through the Champion Challenger tab. Add as many rows as alternative paths for the decision and define the percentage of cases for each decision path. All alternative decision paths need to add up to 100%. Switch Components Switch components apply conditions to select between components. These components are typically used to select different issues (such as, interest or risk) or they select a component based on customer characteristics or the current situation. Switch components express component selection through the Switch tab. Add as many rows as alternative paths for the decision, use the Select drop-down to select the component and enter the expression that defines the selection criteria in the If field. The component selected through the Otherwise drop-down is always selected when conditions are not met. 84

87 Exclusion Components Exclusion components conditionally stop the propagation of results by restricting the selection to results that do not meet the exclusion criteria. These components are typically used to build traditional segmentation trees in strategies. Exclusion components express the selection of results through the Exclusion tab. Use the Type drop-down to select the type of data: Pages or Component. In the Name drop-down, select the page or component. The Class field displays the class context. The criteria to exclude results is defined as one or more conditions in the Exclude when all conditions below are met section. This is a value pair between properties in the exclude component and, depending on what you selected in the Type field, properties in the page or strategy component. If you do not define any condition, this component stops the propagation of the results of its source components. 85

88 Component Connections Connecting components (page 86) Expressions (page 86) Segment filtering (page 86) Connecting Components Connections between components are established through selecting a component and dragging the arrow to another component. Expressions Another type of connection represented by blue arrows is displayed when a component is used in another through an expression. Segment Filtering Segment filtering can be applied if segments are brought to the strategy through segmentation or segment filtering components. 86

89 Double click the connector. In the Connector Properties dialog, select an existing segment. You can also provide a value that is not available to the strategy when designing it, but present in its execution chain (for example, a segment in a strategy included through the sub strategy component). Click OK to apply the specified segment filtering. Expressions in Strategies Working with strategies means working with the strategy result data classes and the class the strategy belongs to. These classes can be combined in expressions or by introducing segmentation components (page 70) that work on the strategy result data class, and not the class the strategy belongs to. Understanding the expression context (page 87) Using component properties in expressions (page 87) Understanding the Expression Context Using the dot notation in the SmartPrompt accesses the context of an expression, which is always the strategy result class (for example,.pypropensity). To use properties of the Applies To context, declare the primary page (for example, Primary.Price). If the properties used in the expressions are page properties, you can omit the Primary keyword (for example, instead of Primary.SelectedPropositon.pyName, use SelectedPropositon.pyName). Pages of the Applies To context and pages added through Pages & Classes are displayed together in the smart prompt. When using page properties without declaring the Primary keyword, there is no disambiguation mechanism to differentiate between referencing the embedded page in the Applies To class (for example, a Customer.Name embedded page) and the output of a component (for example, Customer.Name, where Name is the output of a component named Customer). Using Component Properties in Expressions To use properties of one strategy component in another, declare the name of the component. For example, RetentionBudget.RetentionBudget * (GlobalControlParameters.MaximumBudgetRatio/100) < 80. If the component used in the expression outputs a list (multiple results), only the first element in the result list is considered when computing the expression. 87