This blog is an attempt to share the individual ‘understandings’ and ‘expectations’ related to the emerging trends in SAP Forecasting/Supply Optimization vis-a-vis Demand Sensing and Inventory & Service Level Optimization currently projected with SAP’s takeover of Smartops.

Understandings as above relate to outcome based on the work experience in the core ERP system (MM/PP modules)  and

Expectations translate to whether the ways of working will change because of the new developments and lead to additional learning.

    For quite some time we have been hearing about the takeover of Smartops and there would have been people interested to know if it had got anything in store for them. The merger formalities seem to be complete and post merger it has become SAP-EIS (Enterprise and Inventory Service Level Optimization). A team is set to deliver what is called as the Enterprise Demand Sensing (EDS) and Multi-Stage Inventory and Service Level Optimization.

a. Enterprise Demand Sensing: (Positioned as a Cloud-based Analytics solution).

   Traditional forecasting techniques using time series methods predominantly which always have the risk of poor prediction tag leading to lost opportunity.

Loads of sales/consumption history data across years provides forecast results usually on a long term horizon of say a year atleast and when they start creeping in towards the short term because of these sudden changes order execution at the short term level gets tougher and tougher.

More CPG/Retail companies like P&G, Unilever, Kraft/Mondelez to name a few have gone for what is called as demand sensing techniques using predictive analysis/intelligence techniques to make much better forecasting and quickly adapting to the market changes. Companies in other industries/ areas are also seen to catch up this aspect.

    To-the-minute/ current realities of supply chain are considered with an ability to respond to sudden demand spikes/real time demand because of say an instant promotion activity in a networking site for instance or a natural disaster is what is attributed as the USP about this demand sensing. Intelligence to be provided should come via the SAP Demand Signal Management (DSiM) to process the data in a HANA backed environment. Guesstimates say improvements in forecasting accuracy should be seen in limits of 30% northwards.

Take a look at the clipping made available in

The video clearly mentions that this software has the option of integrating with an ERP as well as APO-Demand Planning application.

As a layman with limited understanding of new developments and belief that forecasting is what we use as V* MRP types using a host of forecast models in ERP the question is whether the forecasting algorithms would undergo any changes in the future in the ERP system. If the understanding is right and again if that happens would that be as a Core-ERP offering or would be for something for IS-Retail scenarios only as it is mostly the CPG/Retail customers that are the prime targets?

b. Enterprise Inventory and Service Level Optimization: (Positioned in the SCM suite as an Integrated S&OP offering again within the Supply Planning space)

    At the very first instance hearing about the name Enterprise /Multi-stage/Multi-Echelon Inventory and Service Level Optimization itself was very ambiguous tasting like a bitter gourd forget about the association part. Some one prunes it Service level optimization in discussions and there is a sincere belief that the term had been heard definitely before. Some more tries to associate the relation and the material master MRP view specifying the Service level field flashes across. Bit of reading here and there then brings us closer to the association part.

   What we actually try to do with these service levels in ERP Forecasting is to fix up a percentage based on defined KPIs/criteria. We come across terms like  a ‘normal distribution’ usage to take into account probabilities etc to meet optimized service levels. Finally things start falling in place and we understand that what we are trying to achieve as optimization the service levels is at a single location/ single-stage namely the manufacturing plant. This is called ‘Single-Stage Optimization’ and this is what we on a maximum work in the core-ERP side.

   In any Supply chain, viewed from the point of view of a plant we see two things:

  1. 1.Upstream view–> Represents all partners above the plant and supplying raw materials/components pushing out raw material inventory. Covers the procurement network and called ‘Tiers’. Multiple layers then become ‘Multi-tiers’.
  2. 2.Downstream view–> All partners downstream of the plant involved with fulfilment aspects with FG inventory. Covers the distribution network and termed ‘Echelons’ with multiple levels becoming ‘Multi-Echelons’.

(Side note being what is in Plant or the centre point is the manufacturing network covering WIP/In-transit Inventory).

Fulfilment (Customer is King) forces Echelons to take priority over other two letting us know why it is named ‘Multi-Echelon/Multi-stage Inventory and Service Level Optimization’.

   Any Supply Chain experiences pains/issues in one form or another and they get consolidated at the very high level as ‘complexity’ issues and ‘variability’ issues.  Again ‘variability’ breaks up into ‘demand’ variability’ and ‘supply variability’.

   Looking at the earlier SCM SNP space and with the offerings provided earlier the ‘Supply variability’ was definitely a pain point and with the new offering, SAP is set to address the issue with the new setup/merger. Handling supply variability and optimizing is expected to be the USP here. And you can see that Demand sensing is fitted in to handle the Complexity issue part.

In parallel during these ‘inferring’ days, on the SCN site we started to see a host of materials being uploaded and made available in this same forum.

Has real good materials set to explain Safety Stock Fundamentals and Stochastics, Cycle and Pipeline Stock: The Deterministic View etc., and what Smartops does differently there. Read and get benefited.

In the materials you can see a mention about usage of ‘Gamma distribution’ compared to the traditional ‘Normal Distribution’ methods we know of in our ERP system today and the advantages etc., making us to think again if any new changes will come over. With the SCM SNP space opening up to inventory modeling and optimization there seem to be more opportunities for specialists concerned with the appropriate and new skills in this space.

Will utilise this blog to look for relevant additions/corrections from experts working in the Forecast related space with forecasting models/time series data/Inventory Modeling techniques in Supply Chain and share their thoughts. Also useful would be pointers to other targets/forums who might be able to help further.

May be premature days to expect still but any information provided will be useful to understand things better and share it back in the forums for general benefit of all.

Thanks for your precious time now for reading and also for your future time to write back.



To report this post you need to login first.


You must be Logged on to comment or reply to a post.

  1. Srini Ravindra Kumar Post author

    MM experts in forecasting and inventory modeling background- Please let me know of your thoughts/corrections/suggestions.

    MM Moderators, Your inputs are also very much appreciated.

    1. Joao Sousa

      A bit of complex landscape no? It’s not very normal to see a customer with APO, in that diagram he needs APO, EDS, S&OP…. auch.


Leave a Reply