1. Implementation Guide
  2. Appendix: New Item Configuration Guide

A Appendix: New Item Configuration Guide

This appendix describes how RDF Cloud Service performs the batch process for New Items and alternative implementation approaches for the New Item module.

New Item Batch Tasks

New Item batch performs the following tasks:

  1. Identify the current New Item/stores that will turn into existing items. This identification is based on the approve substitution method, sales history start date and time series duration. If a New Item/store is going to turn into an existing item/store, its approved Like Item/like stores, approved substitution method, and other settings will be automatically wiped out.

  2. Automatically identify New Item/stores in the system based on the user loaded New Item forecast start date

  3. Calculate eligible Like Items for the auto-identified New Item/store. The eligible criteria is based on the previous approved base demand, existing items's sales history pattern, active forecast item, and if the existing item is under the same user specified level as the New Item. The Like Item eligibility criteria also includes a Like item exclusion mask that can be customized based on business needs. The user specified level is an input in New Item plug-in. The sale history used in the eligible like item calculation can also be specified in New Item plug-in.

  4. Calculates the similarity score between the New Item and eligible Like Items based on product attribute information and provides three Like Item recommendations per the New Item/store that is based on the similarity score rankings.

  5. Automatically approves system recommended Like Items, substitution methods, and parameters if user specified.

For details about New Item substitution methods and settings, refer to the Oracle Retail Demand Forecasting Cloud Service User Guide.

New Item Forecast Approaches

In RDF Cloud Service, New Item forecasting is implemented differently from previous versions. A baseline forecast for the New Item is generated based on the seasonal curve level. The seasonal curve for the New Item is produced through seasonality escalation like other existing items. The level for the New Item can be produced based on settings in new item module.

Implementation Options

In a normal RDF Cloud Service implementation, there are four possible approaches to setting up the New Item module.

First Approach

Ignore

The customer does not plan to use it and chooses to simply ignore the New Item. For this approach New Items need to be set up for the RDF plug-ins to run successfully. Do the following to ignore using the New Item:

  1. Leave all the GA New Item plug-in inputs unchanged.

  2. Use the GA prod attribute hierarchy file for the domain build.

  3. After the domain build, open the Batch Flow Management workbook and set up the boolean switches related to the New Item as follows:

    1. Set runnewitembatch to false

    2. Set runnewitem_CF_ to false

    3. Commit the workbook changes

Second Approach

Use New Item to Manually Set Up New Item Forecasting Parameters

The customer does not have any attribute information, but would like to use New Item to manually set up New Item Forecast Parameters such as Like Item, Like Store, Substitute Method, user provided Base Demand, and so on. The New Item batch is still necessary to run in this case, because the first step of New Item batch is necessary to automatically turn off New Item/New Store functionality once a New Item/Store matures. However, the Attribute Maintenance workbook will not be able to be created because there is no valid attribute information.

For this approach, the implementor need to set up the New Item using the following steps:

  1. Go over the following New Item plug-in inputs one by one:

    • New Item data source

      The intersection of the specified measure decides the level where the New Item and New Store will be set up. The sales history from this measure will be used to decide if a new item/store is matured.

    • Product Map

      Allows the user to specify which level that the New Item and eligible Like Item should be under. A Like Item can only be selected from the items that roll up to the same level.

    • New Store level

      Allows users to specify Like Stores for the whole level

  2. Use the GA attribute hierarchy file for the domain build.

  3. After the domain build, open the Batch Flow Management workbook and set up the boolean switches related to New Item as follows:

    Note:

    New Item forecasting will generate unpredictable results if the final level intersection's prod and loc dimensions do not match what is configured in the New Item module.

    1. Set runnewitembatch to true.

    2. Set runnewitem_CF_ to true if the level needs to enable New Item forecasting. The final level forecast intersection must have the same product and location level as the New Item data source measure.

    3. Set runnewitem_CF_ to false if the level needs to disable New Item forecasting or the final level intersection 's product and location level were different from the New Item data source measure .

    4. Commit the workbook changes.

  4. Use New Item Maintenance workbook and the New Store Maintenance workbook to set up New Item and New Store settings in the New Item Basic Parameters worksheet of New Item Maintenance Workbook and make sure nitautoapp is set to false.

  5. This will turn off autoapprove in the New Item batch process. Set nitdtsdur (time series duraton) to the user desired value. This control how many sales periods a New Item/store needs to mature. Within the New Item Maintenance Workbook, you can also play with the New Item forecast parameters to review the What-if New Item Forecast.

Third Approach

With Attribute Information, Use New Item to Get Recommended Like Items and Manually Set Up Like Item and Like Stores

The customer has attribute information and would like to use New Item to get recommended like items and manually set up like item and like stores. Full fledged New Item functionality is enabled in this case and data preparation is very important. In this case, the implementor needs to set up the New Item using the following steps:

  1. Go over the following New Item plug-in inputs one by one:

    • New Item data source

      The intersection of the specified measure decides the level where the New Item and New Store will be set up. The sales history from this measure will be used to calculate eligible like item and decide if a new item/store is matured.

    • Product Map

      Allows the user to specify which level that the New Item and eligible Like Item should be under. A Like Item can only be selected from the items that roll up to the same level.

    • Attribute Weight Location

      Allows the user to specify which location level that the attribute weight used in similarity score calculation is based on. The attribute weight measure intersection is to be on product Map/attribute-weight-loc/patt. In GA, it is clss/chan/patt.

    • New Store level

      Allows users to specify Like Stores for the whole level

  2. Prepare the following data files for the RDF environment build and ongoing data load in the future. Product Attribute hierarchy (PATR), product attribute measure (prdattT) and attribute measure (nitdattwgt) should all come from RI. If they are not available from RI for some reason. They need to be in the following files:

    • PATR.csv.dat—attribute hierarchy file

    • prdattT.csv.ovr—product attribute value data file

    • nitdattwgt.csv.ovr—attribute weight data file (without this data file, there is no similarity score and Like Item recommendation)

    The New Item Forecast start date override (nitfcststovr) and user provided New Item Base Rate of Sales (nisros) should be provided by implementor. The following file should be provided if possible:

    • nitfcststovr.csv.ovr— New Item forecast start date data file (without this data file, there is no New Item identification)

    • nisros.csv.ovr—user provided New Item base rate of sales data file.

    Without proper data, the Attribute Maintenance workbook can not be created in the future.

  3. After the RDF environment build, open the Batch Flow Management workbook and set up the boolean switches related to New Item as follows:

    Note:

    New Item forecasting will generate unpredictable results if the final level intersection's prod and loc dimensions do not match what is configured in the New Item module.

    1. Set runnewitembatch to true

    2. Set runnewitem_CF_ to true if the level needs to enable New Item forecasting. The final level forecast intersection must have the same product and location level as the New Item data source measure.

    3. Set runnewitem_CF_ to false if the level needs to disable New Item forecasting or the final level intersection 's product and location level were different from the New Item data source measure

    4. Commit the workbook changes.

  4. Use New Item Maintenance/Management workbook and New Store Maintenance workbook to set up New Item and New Store settings. In the New Item Basic Parameters worksheet of New Item Maintenance workbook, set up nitautoapp to the user desired value. This will enable or disable autoapprove in the New Item batch process. Set nitdtsdur (time series duraton) to the user desired value. This control how many sales periods a New Item needs to mature

  5. If the loaded New Item forecast start date and New Item base rate of sales were not available, the following business flow must be performed:

    1. Load the New Item base rate of sales by loading the customer provided nisros.csv.ovr file

    2. Open the New Item Maintenance workbook to input nitfcststovr without running any approve custom menu and then commit the change.

    3. Run newitem_batch to automatically identify New Items and generate system recommended like SKU.

    4. New Items can be approved automatically in Step 5.c. The user can also create the New Item Maintenance workbook again to manually approve the New Item settings.

Fourth Approach

Bypass New Item Workbooks and Set Up New Item Forecasting through Direct Measure Loading and Calculation

Use this approach when you want to use New Item Forecast, but do not have any attribute information and you do not want to populate New Item parameters manually. You can load information into the New Item Parameter measures directly and avoid using any New Item workbook. It is still necessary to run the New Item batch because the first step of New Item batch automatically turns off New Item functionality once a New Item matures.

For this approach, the implementor needs to set up the New Item using the following steps:

  1. Verify the following New Item plug-in inputs one by one:

    • New Item data source—the intersection of the specified measure decides the level where the New Item and New Store will be set up.

    • Product Map—allows the user to specify which level that the New Item and eligible Like Item should be under. A Like Item can only be selected from the items that roll up to the same level.

    • New Store level—allows users to specify Like Stores for the whole level

  2. Build the domain using the GA attribute hierarchy file.

  3. After the domain build, open the Batch Flow Management workbook and set up the boolean switches related to New Item as follows:

    Note:

    New Item Forecasting generates unpredictable results if the final level intersection's prod and loc dimensions do not match with what is configured in the New Item module.

    1. Set runnewitembatch to true.

    2. Set runnewitem_CF_ to true if the level needs to enable New Item forecasting. The final level forecast intersection must have the same product and location level as the New Item data source measure.

    3. Set runnewitem_CF_ to false if the level needs to _disable the New Item forecasting or the final level intersection 's product and location level were different from the New Item data source measure.

    4. Commit the workbook changes

  4. In the New Item Maintenance Workbook, ensure that nitautoapp is set to false to turn off auto approve in the New Item batch process. Set nitdtsdur (time series duration) to the user desired value. This controls how many sales periods a New Item needs to mature.

  5. Prepare the custom batch control file that loads settings into measures such as nisros (user provided base demand), nitdtsdur (time series duration), nitappsubm (substitution method), nitapplkitm1 (like item), nitappadj (like Item adjust percentage).

    Note:

    In the weekly forecast batch, the New Item batch is run before the RDF batch forecast. The initial step in New Item batch identifies matured New Items and clears out its substitute method. RDF batch checks the substitution method to decide if an item goes through a normal forecasting process or New Item forecasting.

    If an implementor wants to load a substitute method or other new item parameters, it is important to decide when it should be loaded, either before or after the New Item batch runs. If a substitute method is loaded before the New Item batch runs, it may be automatically wiped out depending on how long its sales history is. If a substitute method is loaded after the New Item batch, then the item will be subject to New Item forecasting forever.

Troubleshooting Tips for Common New Item Issues

The following tips will provide troubleshooting solutions to common issues for New Items in RDFCS.

Issue 1

Issue: The New Item Maintenance Attribute Maintenance Workbook Cannot be Opened Due to an Empty Pre-range Solution

Solution

If you are running into this issue you have by-passed Oracle's new item batch or data loading process.

The New Item Maintenance workbook is pre-ranged by these measures:

  • nitwkrange

  • nititm2clsr

The nitwkrange measure ranges the CLND hierarchy. The nititm2clsr measure ranges the pror hierarchy. The nitwkrange measure is calculated in the new item batch, where as the nititm2clsr measures are calculated as a post-process after the data load. The Calendar has been pre-ranged to bring in weeks > (TODAY - TS duration). The nititm2clsr measure is calculated based on nititm2itr. The nititm2itr data file is generated by RDF data load process based on product hierarchy and loaded automatically. These logics were implemented through batch control file and OAT.

Users should not encounter any issue if the standard RDF process is followed.

If you encounter this issue after by-passing your domain, make sure that nititm2itr and nititm2clsr are populated in your domain.

Issue 2

Issue: The New Item Attribute Management Workbook Can Not be Opened Due to an Empty Prerange

Solution

Follow the same solution as for Issue 1.

Issue 4

Issue: You Can Not Generate a Forecast for a New Item Even If There are Valid Like Item and Substitute Method Settings

Solution

The New Item forecast is produced by looking up the seasonal curve through escalation and getting a base rate of sales from the like item/user input/base rate of sales escalation.

If it is not populated, that means there is no seasonal curve for all of the escalation intersections at this particular item/store. An implementor should consider adding a new catch-all escalation level. The other possibility is base demand is zero. You can further looking into the reason by base rate of sales cannot be generated.

Issue 5

Issue: The New Item Forecast is Setup in the New Item Maintenance Workbook, but No Forecast is Generated for the New Item and Base Demand Measures have a Zero Value

Solution

Check the runnewitem_CF_ measure which is visible in the Batch Flow Management workbook. If the measure is false, set it back to true, commit and rerun the forecast.

If runnewitem_CF_ is true, then check the nitappsubm substitution method.

If nitappsubm is set to: Then:

None

No New Item forecast will be generated.

Like Item

At least one valid existing item must be assigned as the Like Item.

The Like Item itself must have a forecast. The Like Item assignment measure is:

  • nitapplkitm1

The nitappadj (total adjustment percentage) can not be zero.

User Input

The user must provide the base rate of sales for the New Item. The value will be stored in nitappros after the New Item approval process. When the substitute method is user input, then nitappros can not be zero.

Issue 6

Issue: When Trying to Approve Like Item Recommendations, Received this Approve Error Message: Can Not Approve, Check New Item Error Message

Figure A-1 Error Message


Error Message

Solution

This error is due to a problem in your New Item set up. On the worksheet, display show/hide for select and approve. For the New Item error message measure, ensure that measure is visible on the worksheet. It will provide additional clues for the issue.