12 Portfolio Optimization

This chapter describe the Portfolio Optimization (PO) module, where a user can access and manage configurations for optimizing distribution over product areas to align with financial planning for upcoming periods. Depending on the role, a user will see the following link. Typically, the administrative user has access to this link.

Figure 12-1 Portfolio Optimization Link

Configuration

To configure Portfolio Optimization, complete the following steps:

1. Use the Manage System Configurations screen configurations in RSE_CONFIG. For further details, seeConfiguration . The relevant configurations in RSE_CONFIG that must be reviewed and edited as required are listed in the following table.

   Table 12-1 Configuration

   APPL CODE	PARAM NAME	PARAM VALUE	DESCR
   RSE	PORTFOLIO_LOWEST_PROD_REC_LVL	Default value is 6.	Lowest portfolio product recommendation level

2. Use the Portfolio Optimization screen to set up portfolio scenarios, as follows.
   1. In Portfolio Optimization, start by setting up a portfolio optimization type. Click the + icon in the area at the top of the screen or click the gear icon to the far right of the screen title bar. You must create a portfolio optimization type for each sales metric/scope intersection/recommendation intersection combination that is required.
   2. Once you are done setting up the portfolio optimization types, return to the main Portfolio Optimization screen and review the tiles that are now displayed in the type area at the top of the screen. Click Start Data Aggregation in any of the types that have not yet started data aggregation. When data aggregation is complete, the link will change to Data Aggregation Complete. You may see a new link appear below Start Portfolio Aggregation. Click this link if it is present. Once portfolio aggregation completes, the status will change to Portfolio Aggregation Complete. At this point, scenarios can be created and executed to generate recommendations.
   3. To create a scenario, first click on a configuration type tile at the top of the screen, then expand the Create Scenario section to show the Create form. In the form, you can set the configurations parameters related to the scenario you want to run. After modifying and reviewing the configuration parameters, click Create Scenario (button) to create the scenario. The new scenario will now display in the Scenarios table below. At this point, optimization aggregation and statistics calculations will begin. Once these stages have completed, select the scenario in the table and click Run in the Status column. Once the run is complete, the status will change to Complete. Click Complete to review recommendations.
   4. On the Results screen, review each of the three recommendations: Highest Growth, Lowest Risk, or Balanced, and approve the one you want to work with by clicking Approve and submitting the Approval form. Upon approving a recommendation, the scenario will be approved, and the portfolio aggregation type will be activated. At this point, the recommendation may be imported by MFP or another application that is integrated with Portfolio Optimizations.

Reviewing Optimization Errors

Error logs for scenario optimization aggregation and statistics calculation stages are stored in the RSE_LOG_MSG table. Optimization run execution logs are stored in RSE_ANALYTIC_LOG_MSG. These can be reviewed if the cause of the scenario error status is unclear in the UI.

Example SQL for Optimization aggregation and statistics calculations stages

select * from RSE_LOG_MSG where PROGRAM_UNIT like '%PORTFOLIO%' and LOG_LEVEL='ERROR' order by MSG_TS desc;

Example SQL for Optimization Run

select * from RSE_ANALYTIC_LOG_MSG where PROCESS_NAME like 'RSE PORTFOLIO RUN JOB';

Exporting Portfolio Optimization Results

Export is supported via a web service that a client application such as MFP can call to retrieve approved results.

Service Name: rase/resources/portfolio/v1/MPOptExp

The following table shows the key fields present in the web service request.

Table 12-2 Web Service Request Input Key Mappings

Key Name	Level	Description
requestID	User	Unique Request ID
prodLevel	User	Product Level can be DEPT, CLS, SBC. MFP GA uses DEPT for Merch Target and SBC for Merch Plan.
locLevel	User	Location Level can be COMPANY, PLANNING_CHANNEL, AREA, LOCATION. MFP GA use PLANNING_CHANNEL for Merch Target / Merch Plan.
clndLevel	User	Calendar Level can be YEAR, QRTR. Only two are supported in MPO. MFP GA defaults to 'YEAR'.
prodScope	User	Optimization Product Scope level can be CMP,DIV,GRP,DEPT. It must be the level above prodLevel. For GA it will use DIV in MT and DEPT in MP.
locScope	User	Optimization Location Scope level. In initial release, Location Scope is defaulted to same Location level.
keyList	dept/chnl/year	Collection for List of Product/Location/Calendar key positions for which data is required.
prodKey	dept/chnl/year	Product Position ID (Department ID if Product Level is DEPT)
locKey	dept/chnl/year	Location Position ID (Channel ID if Location Level is PLANNING_CHANNEL).
clndKey	dept/chnl/year	Calendar Position ID (YEAR ID if Calendar Level is YEAR).

The following table shows the key fields present in the web service response.

Table 12-3 Web Service Response

Key/Parameter	Required	Intersection	Comments
MPOptExpHeader	Yes	User	Collection for the Main Request Response Header
responseCode	Yes	User	Example: 100 - For Successful Processing of Data 200 - For Any Key Syntactical Failures 300 - For Any Header level Validation Failures 400 - For Any Key level Validation Failures 500 - Warning for Any Keys Missing Optimization Results
responseCodeMessage	Yes	User	Detailed Response Message for Request
MPOptExpDetail	Yes	User	Collection for the Main Request Response Details
requestID	Yes	User	Same requestID provided in the input request
keyList	Yes	User	Collection for List of Products/Store Cluster
prodKey	Yes	dept/chnl/year	Same Product ID provided in the input request
locKey	Yes	dept/chnl/year	Same Location ID provided in the input request
cldnKey	Yes	dept/chnl/year	Same Calendar key provided in the input request
optNSlsAmt	Yes	dept/chnl/year	Optimized Net Sales Amount
optGSlsAmt	Yes	dept/chnl/year	Optimized Gross Sales Amount
keyRespCode	Yes	dept/chnl/year	Optional Response Code at Key Level  400 - Requested Key List not Found 500 - Optimization Results not available
keyRespMsg	Yes	dept/chnl/year	Optional Response Message at Key Level.

Data Requirements for Portfolio Optimization

Portfolio Optimization requires at least three years of quality sales transactions to perform meaningful growth calculations. For each year of sales history, the number of weeks with positive sales per year must be at least 5% of all possible sales weeks in the year.