| Oracle® Retail Size Profile Optimization User Guide for the RPAS Fusion Client Release 14.1 E56467-01 |
|
 Previous |
 Next |
| Oracle® Retail Size Profile Optimization User Guide for the RPAS Fusion Client Release 14.1 E56467-01 |
|
Previous |
Next |
This chapter provides an overview of the Oracle Retail Size Profile Optimization (SPO) application, the products that are integrated with the application, and the features, and concepts of the application.
Oracle Retail Size Profile Optimization (SPO) creates optimal profiles of size distribution, and prepacks by merchandise area for each store and at higher levels of aggregation on both the merchandise and store dimensions. SPO augments/enhances the analytical selection of merchandise at size-level by determining optimized size profile and creating prepacks to match them. This helps downstream applications such as Assortment Planning, Inventory Planning, and replenishment systems to translate style/color, or similar aggregate level plans or replenishment strategies, into optimal order or buy quantities by SKU at the store level.
SPO reduces stock outs and end of season markdowns due to size limitations while maximizing assortment productivity, ensuring that the store's shelves are stocked with the right merchandise in the right sizes to meet customer demand. Additionally, SPO allows users to view and approve these optimal size profiles within the RPAS interface.
SPO is usually an independent activity performed to generate, review, and approve size profiles, and prepacks. The frequency of creating size profiles varies from retailer to retailer, and from product area to product area. Optimization runs for size and prepack are typically processed by season or quarter and may be run at shorter frequency or on ad-hoc basis. Frequency may vary depending on retailers, for different geographies and for different product areas.
The Size Profile Optimization solution includes:
Data preprocessing to correct for stock outs and other data anomalies.
Size profile generation per store.
Richness versus reliability tradeoffs to determine optimal size profiles.
Automated sizeprofile escalation and exception-based approval of size profiles.
Manual review, analysis, and adjustments of size profiles.
Creation of location clusters based on size profiles.
Creation of prepacks aligned to cluster size profiles.
The sections below describe the various concepts behind SPO. The various concepts are:
A size range is a collection of types of sizes and is an attribute of a particular collection of merchandise manufactured by a vendor (for example, S-M-L-XL). While at the lowest level in the Product dimension, such as the Style-Color, size range is a unique attribute for each Style-Color, at higher levels in the Product dimension a group of merchandise can have multiple valid size ranges. For example, a category may contain merchandise that includes multiple types of size ranges, such as both S-M-L-XL and 4-6-8-10. Thus, a given category/store can have multiple size profiles for each valid size range.
|
Note: Users can view, analyze, and approve size profiles at different product/location levels as well as browse through profiles for different size ranges. |
SPO utilizes size profile escalation logic to determine the optimal size profile for a level in product and location hierarchies.
The escalation logic used is a predefined traversal path along hierarchies and can be defined through a combination of configuration time decisions and user inputs from within the Administration activity of SPO.
|
Note: For details on how to specify the escalation path, refer to the "Escalation Path Management Step" section in Chapter 2, "Administration". |
After the escalation path has been specified, SPO uses this information to determine the optimal size profiles associated with any merchandise/location/size range combination.
The concept of escalation becomes relevant due to the automated pruning or rejection of invalid size profiles by SPO at various levels. When a size profile at a given level is pruned, SPO traverses the escalation path to determine the valid or optimal size profiles associated with that level. This does imply that higher level optimal size profiles are resolved (by approval) before lower level size profiles are resolved.
The following diagram represents an example escalation path:
SPO traverses the predefined escalation path to resolve the size profile for each style/store. If the system does not find a profile associated with a given level in the escalation path, it traverses to the next level on the assumption that a valid profile exists in at least one level on the escalation path.
For example, for style/store, SPO attempts to look up the size profile distribution at style/store. If it does not find a valid profile at that level, it moves to department/store, then category/region, and so on.
For certain businesses, it may be necessary to further classify size profiles by season code in addition to product, location, and size range.
Season code is a merchandise attribute that classifies merchandise based on launch date or start date. Examples of season code are spring, summer, fall, and so on.
SPO can calculate and present size profiles by season code. SPO can thus determine optimal size profiles that also take season code into consideration while traversing an escalation path.
Kinks are anomalies in a size range. When SPO checks generated size profiles for accuracy, it looks for anomalies in size profiles such as a size that sells for a considerably lower price than its neighbors. These are considered kinks in a size profile, and the system can prune or reject a size profile if it has kinks.
However, certain odd sizes also exhibit kinks. For example, it is acceptable or even expected that an odd size sells significantly lower than its neighboring sizes. Kink exceptions are mechanisms for users to indicate or tag these odd sizes to the system where the kinks are acceptable. By specifying these odd sizes as kink exceptions, users can indicate to the system the sizes that need to be exempted from checking for kinks when SPO conducts its checks for accuracy of calculated size profiles.
|
Note: Kink exceptions or odd sizes within a size range can be specified from within the Size Range Management. |
SPO utilizes the following hierarchies:
Calendar (CLND)
Product (PROD)
Location (LOC)
Size (SIZH)
The size dimension uniquely identifies a size within a corresponding size range; for example, a size of "S" comes from the "S, M, L, XL" size range. The size dimension displays the size in relation to the size, size range, and size length.
Attribute (ATTR)
Size Range (SIZR)
The size range dimension displays information based on the size dimension, including the size range and size range length.
Escalation Levels (ELH)
The escalation levels dimension groups escalation levels together along a single dimension.
Generation ID (GIDH)
The generation ID dimension groups profile generations in a dimension.
Cluster (CLSH)
The cluster dimension has the positions of all clusters that can be used in location clustering.
Pack (PCC)
The pack dimension has the positions of prepacks.
There are three types of parameters in SPO: Pre-Processing, Profile Generation, and Post-Processing. These parameters are used at various stages in the SPO batch in the calculation and pruning of size profiles.
Pre-Processing parameters can have both global and override versions. The global versions are scalar measures which can be set in the Administration, and the override versions are based on SKU/store and can be set in the Pre-Processing. The values in both versions are merged prior to running preprocessing, and are then saved in generation ID-specific measures. The SPO parameters can be specified from within the Administration activity, and the preprocessing overrides can be specified in the Pre-Processing Analysis.
The Profile Generation and Post-Processing parameters are both escalation level-based or scalar measures, and neither has an override version. These parameters can be set in the Administration activity, and their values are stored in generation ID-specific measures prior to each batch run.
The SPO process flow consists of the initial setup, executing preprocessing, generating size profiles, the pruning of size profiles based on accuracy criteria, and finally the analysis and approval of size profiles. The manual analysis and approval of size profiles can be restricted to exception or alerted size profiles. Alerts can be configured into the SPO system.
The following diagram depicts which tasks are used for the steps in the SPO process flow:
SPO contains five templates that support the following processes:
Administration Activity
Used to set up the preprocessing, profile generation, and postprocessing parameters.
Preprocessing Analysis Activity
Used for what-if analysis of preprocessing. This is an analytical tool available for the user to adjust parameters and observe the results.
Size Range Management Activity
Used to manage size ranges (including sub-size ranges) and product-to-size-range relationships, to specify kink exceptions and kink exception ratios, and to request size profile runs for specific size ranges.
Size Profile Approval Activity
Used to adjust and approve the generated profiles. Only approved profiles can be used in the escalation and export.
Location Clustering and Pack Definition Activity
Used to create a specified number of location clusters based on similarity in their size profiles. In addition, there is a new functionality of prepack definition added in this, which allows users to define prepacks using cluster size profiles.
The SPO user workflow comprises the following steps:
Use the Administration activity to set up the parameters.
Use the Preprocessing Analysis activity to perform what-if analysis and tune the parameters.
Use the Size Range Management activity to specify the subset of size ranges for which the size profile batch needs to be executed.
Run a batch to generate size profiles.
Use the Approval activity to view the generated profiles and exceptions, adjust and approve the profiles, and review the resolved profile at the lowest level.
Use the Location Clustering activity to create clusters.
Create prepack containers using computed cluster profiles.
Some steps of the Size Profile Optimization require certain planning actions to be executed. They are available on the menu bar as shown here. The planning actions for SPO are available on the menu bar for some activities. The menu options varies depending on the currently open step.
The planning actions available for each of the Size Profile Optimization steps are as follows:
Preprocessing Analysis Activity
Preprocessing
Size Profile Activity
Approve Level
Location Clustering and Pack Definition Activity
Location Clustering
The Preprocessing planning option is used to initiate Preprocessing. The Preprocessing planning action is a batch process. This planning action cleanses historic data based on user set parameters and thresholds.
An Approve Level option is available for every size profile level displayed in the activity, based on the selections made in the Size Profile Approval process.
The number of Approve Levels displayed depends on the number of escalation levels chosen in the Size Profile Approval activity. The Approve Level that the user approves becomes the escalation level at which the dimension is maintained.
The standard release of SPO is optimized for typical workflow of the intended users. It can be customized further based on retailer policies or user preference using the following options:
To save the format in which each is displayed, use the Save Format feature. This ensures that when this is retrieved or when similar s are created in the future, the user is presented with the same format. To ensure that the data entered is saved in template format, click the View menu and select Save Format.
For information about creating, modifying, and editing users in SPO, see the Oracle Retail Predictive Application Server Administration Guide for the Classic Client or the Oracle Retail Predictive Application Server Administration Guide for the Fusion Client.
This is used to setup alerts for SPO. Alert Manager provides user-defined and user-maintained exception reporting. You can define a measure and have it evaluated to see if any of its values fall outside an acceptable range or do not match a given value. When this happens, an alert is generated to let you know that a value may need to be examined and possibly amended in a workbook. For more information about Alert Manager, see the Oracle Retail Predictive Application Server User Guide for the Fusion Client or Oracle Retail Predictive Application Server User Guide for the Classic Client.
