PAL & ASSORTMENT

[Former Member]

PAL & ASSORTMENT

Retail Industry is going through dynamic changes – one, primarily due to shifts in economy and other due to shifts in technology. Brick and mortar commerce is merging with e-commerce, m-commerce and ultimately with budding Omni Channel scenario which include multiple online sales platforms, New Cross Sell and Up Sell Models, Online Auction Engines and Social Media. All along these changes have been pushing and introducing different directions and business process requirements, which are currently not met by traditional ERP/POS and Planning Solutions for Retailers.

However, the key change and impact – which is evidently due to the incremental Omni Channel - is on Demand Planning accuracy and Right Assortment planning. We have experienced this during the past years Thanksgiving and Christmas where the retailers were unable to meet with the customer service expectations.  The cause has been highlighted as shipping issues. However, the issue is of stock availability at right place. As these Omni Channels are new in existence and there is not much historical data to support supply chain decisions - Predicting Virtual consumer behavior has become a bottleneck for the Retailers. This situation is expected to continue as the shopping destination for products are no more definite or limited to stores.  The efficiency of service would be the key to success for Retailers.

In one of my recent Assortment solution assignments with a Global Fashion Retailer, I can see this dynamic shift – the process of local assortment and allocation has to change,  as the Omni Channel demands are creating the imbalance in stocking strategy. In current scenario one Channel is moving unpredictably - suddenly faster/slower than other. This becomes a critical issue and to plan Stock availability at right place becomes a huge challenge – agile quick decisions are required as Stock pools have to pushed and pulled based on dynamic demands of the consumer behavior.

The introduction of the Omni channel in the Retail environment is creating a new need for Retailers. They need to gear up with sophisticated tools like SAP HANA - Predictive Analysis. SAP HANA - PAL ( Predictive Analysis Library) provides a fantastic solution for Retailers to tackle these complex stock assortment scenarios and ensure the Channel / Location predicts the right assortment, right quantity at right place including the virtual locations.

In SAP HANA, release of PAL includes classic and universal predictive analysis algorithms in the eight data mining categories:

• Clustering
• Classification
• Association
• Time Series
• Preprocessing
• Statistics
• Social Network Analysis
• Miscellaneous

(One can get great details here http://www.saphana.com/community/learn/solutions/predictive-analysis)

Few of the Key algorithms available in PAL are - Outlier algorithm / K- Means algorithm /Apriori algorithm / Multiple Liner and Polynomial Regression / ABC Analysis. These algorithms can play a very critical role in setting the right and profitable Assortment for the Retailers.

In order to quickly identify the Stock Out and Over Stock Situation - Outlier algorithm is a smart way. By using Outlier algorithm, Retailers can get information on their stock status in fraction of seconds. Traditionally achieving clear information about Stock Out / Over Stock and Loss of Sales information has been a handicap.

Currently, merchandise category performance matrix are viewed at standard grouping or static clustering level. K-Means Algorithm has the potential to be a great tool for Merchants and Stock Planners to analyze the fast moving and slow moving products and locations.

Concept of cross sell and up sell can be easily identified with Apriori algorithm.

(check this link for an informative blog by Hardik Patel http://www.saphana.com/community/learn/startups/news-views/blog/2014/02/27/hana-curious-apriori-algo...

In my current POC, we are implementing the PAL library in our Assortment solution.  A few interesting facts we are observing are:

• Algorithms are extremely powerful in nature,
• Easy to deploy in a complex solution
• Speed and Performance is great

As I continue to explore PAL library on SAP HANA, it is becoming evident that PAL is meant to play a vital role for Retailers in the current complex Omni channel environment.