SAP Retail integration painpoint

Hello everyone,

Today i will talk about my experiences on SAP Retail Integration “Pain-point“. This topic i think it isn’t new for you guys who do retail integration.

We saw a very nice integration Landscape with data transfer smoothly from system to system, from SAP to vendor systems. But behind the screen, there are many pain-point you will face to make sure the technical system does not affect business daily basic.
This topic will show my opinions about SAP retail integration, you can share your idea by a comment on below.

1. Real Time Inventory

Real-time or near real-time inventory is important in retail.

Customer matter: for online sales channel where system/website can show available stock to a customer, customer can check store location nearest actively. With the walk-in customer, the cashier/salesman can help check available stock to increase sales opportunities.

Example: Walking customer goes to store A asking for new release Phone, but Phone sold out store A, a salesman will check stock stores nearest customer location and suggest delivery to a customer at some point of time.

Store and Distribution Center matter: Store manager and Distribution Center (DC) operation can see actual inventory in their location and ready for fulfillment and other purchasing, plaining inventory activities. Moreover customer booking then negative stock and can not deliver on-time will be negative impact with customer and retailer consequence.

Inventory is impacted from many sources, one of the main sources is from sales out. As you see in the picture above, retailer systems own different system landscapes, technologies, infrastructure. Inventory data flow will go from system to system before going to SAP, which means integration design and monitoring have to authentic and regular. So the challenges come from vendor systems like POS, E-Commerce ..etc. In this case, SAP side can not control.

Development effort matter: As you see in the landscape above, CAR now is one of the important systems to audit, aggregate, and analyze data. Data available on HANA views, where you can perform visualization by development.

So, pain-point here:

– To get inventory exactly, make sure that no down-time interfaces which impact inventory

– Vendor system dependency

2. Offline System Mode

We have to accept that networking and infrastructure now are perfect, but it isn’t mean no down-time.

Online and Offline modes are regularly question of retailer.

Online System Mode help data transfer consistency reduce impact business.

Offline System Mode help store can run with own master data when network down.

Online-offline combination: the combination between online and offline modes of vendor’s system will help improve data real-time and reduce business impact.

Example: Electronic store POS system still can sales with own master data when system down.

Online-offline combination, we have to design integration more flexible to adapt. Not only SAP but vendor system also have been adapted.
To run offline mode, for example, POS has to store Material, Sales Price, Promotion on local PC. Transaction data will be synced to POS HQ when the network is up.
Asynchronous integration messages are recommended for online-offline mode.

So, the pain-point here:

– Make sure that vendor’s system can run on both mode

– Integration objects may need more to adapt offline mode

3. System Distribution

Vendor systems may distribute depend on their infrastructure, location, and data volume. Data transfer from front-end to back-end thought steps. SAP Integration services like PI/PO itself designed should not access data to each terminal. So the question is how to collect data from each server and send it to SAP ? and how to distribute data (example: master data) to each terminal ?.
Normally, the vendor’s system will have “Integration Tools”, it looks like middleware component where it can collect and distribute data to servers/terminals

So, the pain-point here:

– Vendor system need build / adjust “integration tools” to adapt interface objects.

– Pay more effort for interfaces monitoring.

– Complexity of system landscape

Thank you,

Cuong Dang