Inspirational Sessions (part 2)
Evolution of forecasting in Jumbo’s fast growing supply chain
Piet Verbakel and Jochem Westeneng described how, in an environment with more stores, more channels and more countries, forecasting and replenishment contributes to customer satisfaction.
The F&R team maximises availability, quality and freshness at optimal costs. In a demand-driven supply chain for bricks and clicks, they explained how they used concepts such as cross-chain info sharing, multi-echelon thinking, advanced systems and models, CPFR and integrated business planning. The result: Exceeding customer expectations!
By: André Vriens
Implementing IBP in a nuclear medicine context at Curium
Nothing is straightforward in an IBP process at a nuclear medicine company. Stock evaporates in days, there are very few suppliers and it is nearly impossible to shape the demand. In this session, we’ve shown that, with the right flexible tooling, the right people and a properly managed data set, you can make IBP work even in a tough environment. By building the Anaplan model while telling the Curium story, we demonstrated the complexity of the process and the ability of Anaplan to deal with it. For most people, this is pure magic made possible, and that’s why we ended the session with a magician.
By: Bram Kurstjens
Inventory insights: Maximise the value of inventory reporting
Inventory control, as an important building block of a health inventory management, is maintaining the right balance of stock in your warehouses while providing a better service level and real-time monitoring of the stock dynamics.
Inventory reporting is a basic but very effective way of inventory controlling, and combines the power of visualisation together with analytics and a call for action.
Successful inventory reporting should include stock decomposition and monitoring. Understanding the purpose of the inventory, or stock decomposition, can be done on a lowest SKU levels, as well as total region level.
Monitoring the stock components and their drivers leads to fact-based decisions and proactive management.
So, the simple recipe to a worry-less inventory management is: visualise, analyse, act!
By: Bohdana Shumanska
NPI planning at Abbvie
New product introduction (NPI) is an integral part of Abbvie’s strategy. During the session, which was hosted by Conno Hendriksen, some aspects were highlighted on Abbvie’s approach to prepare for successful product launches, both from an organisational and process point of view.
The first part of the presentation was focused on a general introduction of the company. Afterwards, Conno explained how NPI’s fit in Abbvie’s differentiated planning approach. Dedicated supply planning teams, cross-departmental collaboration, ramp-up and scenario planning are some examples of NPI specifics that were detailed out.
By: Manu Nuyts
Enjoy making Heineken customers win by planning excellence
This session focused on a roadmap story on aiming for planning excellence by the largest Heineken export operating company, based in the Netherlands. Heineken prepared itself to be ready for the future by finding the right balance between processes, tools and people. Paul Groen (Head of Supply Chain Planning, HNS) acknowledged that the road to planning excellence is one of success, but that there are many lessons to learn along the way. During the session, he shared three things he learned are needed: happy and healthy people, data insights and prescriptive planning.
By: Ruud van der Heijden
Implementing inventory settings via machine learning
Having an inventory policy dedicated for each individual product is the optimal solution. However, in the case of thousands of products, dedicated product policies are extremely hard to maintain without any advanced tooling and a dedicated analytical team. To reduce complexity, product categories are introduced such that inventory policies are defined per product categories instead. However, products within the same category may have quite diverse inventory setting configurations, and having those products being represented by the same category policy may lead to bad performance. We use machine learning to create better segmentation strategies with the goal of reducing inventory costs under service-level constraints.
By: Francesco Russo
Experiment: No-touch planning vs. some-touch planning?
Forecast adjustments are a time-consuming component of the forecasting process, but do these judgmental adjustments actually increase the forecasting accuracy? Is no-touch planning the future or is a more optimal, hybrid combination between human planner and data-driven support system more effective (a procedure that we label ‘some-touch planning’)? In our experiment, we tested the potential of such a hybrid forecasting procedure as a more effective alternative to a procedure-based on ‘no touch planning’.
During the session, we were confronted with our own cognitive shortcomings when it comes to planning. By applying a hybrid procedure using human knowledge when required it was shown that the quality of the forecasts and plans can be increased.
By: Freek Aertsen
Improve capacity planning in labour-intensive industries
It’s time to stop fighting fires; it’s time to get rid of your Excel sheets; it’s time to start proper capacity planning. This session shed a light on one of the most poorly managed processes in project-driven companies.
Most of the attendees were not happy with their existing resource planning processes and solved ad-hoc questions in Excel. Therefore, project managers and resource managers did not have the same view on whether the projects would have the right resources with the right skills at the right time. Anyone can understand that lacking the right resources leads to expensive delays. In this session, we solved that equation with a fit-for-purpose, drum-beat process and an Anaplan tool that provides the right data at the right time. A bit of an effort, but ultimately it will save cost and time on your projects.
By: Sander Hulsen and Bram Kurstjens
Next Generation APS
The APS landscape develops rapidly. ‘Traditional’ vendors provide ever-deeper, industry-specific solutions while leveraging analytic capabilities for faster decision making and improved visualisation. The cloud-based, analytics-driven contenders, meanwhile, provide ever-more ‘template-based’ process solutions. What is the current state of the landscape? What are best practice roadmap strategies given these fast developments?
The APS systems’ traditional core added values are in aggregation and visualisation, optimisation, and integration, delivered in well scoped in functional areas like forecasting, supply network planning, production planning and scheduling. New technologies and capabilities push data-driven, cross-boundary and planner-first improvements like cognitive automation, enhanced ‘search-bar’ user interfaces and digital twins.
The session explored how to these cross-boundary, cross-process capabilities are best deployed to drive maximum value. Whether they are delivered through an entirely new system or an upgrade of existing APSs, the core is to focus the application to expert decision making, whether it is in faster decision making at S&OP level, improved responsiveness at demand execution, continuous supply chain improvement through deep data analysis or perhaps offering schedulers tips to improve their schedule further.
By: Bram Bongaerts
De-mystifying machine learning for time series forecasting
Machine learning has the promise of changing forecasting and improving forecast accuracy. However, as we have seen in recent research on timeseries forecasting, pure machine learning models still do not beat traditional statistical models. But by combining machine learning and traditional methods, a performance increase can be reached.
Key finding discussed during the session were that in order to unlock the potential of machine learning in your organization you need:
- Skilled people
- Quality data
- A good business case
- A good setup in IT
By: Koen Koolen
Student Hackathon: The sustainable supermarket
A group of students was given the task to define a supermarket’s strategy and translate that to a sensible inventory policy. Will they be the cheapest or the freshest? And did they manage to control the data in EyeOn’s data science toolkit to get the desired results?
With the help of EyeOn and EyeOn’s software partner, Dataiku, students were quickly able to prepare and transform the given input data and create meaningful insights to make effective decisions. The winning team, chosen by the PID attendees, was true to its given strategy and has shown nice results: balancing costs of lost sales, inventory, and waste.
By: Arthur Nazarian
Planning at war? Key challenges to achieve optimal inventory health in conflict areas of the ICRC
Non-governmental organizations (NGO’s), like the International Committee of the Red Cross (ICRC) are considered different, but are they really? In this session Ruben Naval showed that there are quite some similarities in inventory health management, though the places in the world and the circumstances in which are different. Many hospital operations are predictable, though some events (disasters/war strikes) are not. This asks for a differentiated planning policy to ensure the right and timely availability of relief supplies with minimal losses.
By: Ruben Naval (ICRC)