Generic selectors
Exact matches only
Search in title
Search in content
Search in posts
Search in pages
Filter by Categories

Planning Architecture in a Omni-Channel World

The Problem Statement

Ever since the early 80’s we have worked with a traditional 4 tier planning and execution architecture incorporating all the major elements of a supply chain, namely demand, distribution, production and suppliers. The 4 tiers include strategic planning, tactical planning, operational scheduling and execution or also referred to as the transactional backbone.

The key question is that with the advances in supply chain and technology, what, if any chances are required to the basic and traditional planning architecture? In order to answer this question, we need to review the functional components and mapping of the technology to those functional components.

Technology and Functional Mapping

The functional components are mapped to the planning architecture below, they include all functions of an enterprise which are required to operate the supply chain, they exclude planning functional included in the production development process such as global line planning or merchandise planning. The biggest difference in the traditional model was the technology and business process driven difference between integrated supply or master planning approach and the sequential approach.

From the time, when this planning architecture was developed, a number of developments have taken place in the supply chain world:

  • The supply chain has changed, mostly due to the diversification on the market channel side with the emergence of omni-channel structures. This is in particular true for traditional wholesale oriented companies, which now operate their own retail and e-commerce channel and hence have to manage multiple fulfillment models with complex inventory allocation rules.
  • The focus on retail has changed, for many wholesale company the challenge has become to adopt traditional retail only process and functions such as merchandise planning (category planning, assortment planning, price planning, promotions planning and space planning (macro and micro), store level forecasting and replenishment.
  • The signal has changed, the traditional history based signals have changed to more real-time usage of sales and demand information, sometimes referred to as demand sensing or demand driven supply chain, to make the supply chain faster and more responsive to demand.
  • The technology has changed, the ability now to process big data volumes, application of artificial intelligence to planning and the ability to plan with real time interactions around the globe.
  • The processes have changed, the drive to implement sales and operational planning at strategic, tactical and operational level has forced more integration of the traditional sequential and function oriented decision-making process.
  • The interaction has changed, internal and external collaboration has become embedded in many process and functions.


As a consequence, the planning architecture is impacted in three key ways:

  • additional functional components are required
  • the content of existing functional components needs to be expanded and adjusted
  • the volume handling and intelligence applied changes the decision-making cycle

Changes to the Planning Architecture

The impact to the functional components of the planning architecture are driven mostly:

  • by the inclusion of retail planning capability,
  • the drive to deploy technology to support integrated sales and operations planning processes
  • by the inclusion of development and marketing centric processes such as global line planning
  • the introduction of attributes to describe planning problems item independent (see also the ebp Global article on attribute based supply planning)

Based on the impact, the traditional 4 tier architecture needs to be expanded with the channel and product layer. The channel layer includes all distribution channels served: wholesale, direct retail, e-commerce or other channels such as QVC. The product layer includes all planning related to the product range as part of the product development and marketing process.

Collaboration, as referred to above, is embedded in many of the functional components such as sales and operations planning, and hence does not require separate functional components.


However, the core functional components are still valid as ever and will not change. There is an impact on certain functional components to handle the additional complexity managing a omni-channel channel structure. One of the key areas is inventory deployment and inventory allocation. Traditionally they dealt with customer in a wholesale setting, now they also need to manage the inventory deployment and allocation across wholesale, retail and e-commerce channel. Another change is impacting available to promise, which needs to cater for the e-commerce needs to allocate product from different channels to the customer order.

One of the changes in the technology provider landscape is the emergence of players such as Anaplan or Vanguard, both of which come from a business intelligence background and have expanded successfully into the planning space. This shows that the traditional set of players are increasingly pressurized to changed their architectures from the traditional functional building blocks into big data and real-time processing driven setup.

The application of machine learning and artificial intelligence to planning is another area which changed the traditional building blocks of planning systems.


ERP versus APS (Advanced Planning and Scheduling)

The basic distinction developed in the 80’s between ERP and APS system is still valid. The shortcomings of material requirements planning (MRP) as the core planning logic of any ERP system or traditional planning architecture and its assumptions and impact to the overall lead-time is also still valid.

MRP still does more harm than good, because MRP completely disregards the first two laws of supply chain physics (see Hopp & Spearman). The first law states that variability always degrades the performance of a production system. The second law states that variability in a production system will always be buffered by some combination of inventory, capacity and time.

Most planning systems are item dependent and aggregate at some level to allow the system to operate effectively in a certain layer. The layers must go in the future. Planning needs to go from concepts used in the development process based product attributes to allow effective mapping to capacity and materials, seamless to purchase or product orders ie the transactions. Within ONE planning system all elements need to be handled included the transformation and mapping from aggregated elements to the final transaction. This also means that the boundary between planning systems (APS) and transactions systems (ERP) must disappear.

The next phase of development in the APS space must address some of those issues again to help support the pressure on responsiveness, flexibility and speed required by omni-channel and demand driven supply chain of today.