Inspirational Sessions (part 1)
Using machine learning techniques for data-driven planning improvements
Marie-Louise has created, for her master’s thesis research, a method to evaluate the production performance on actual planning parameters. With machine learning, we can analyse the production and planning performance in order to develop an overall production improvement strategy which comes from a data-driven approach.
During the session, the results of the model—using real client data—were discussed. After the extraction of production and planning parameters from ERP systems, examples of advanced outlier detection based on cluster methods were presented. By linear programming, we were able to define efficiency improvement steps on product level, after which a strategy is developed for focusing on planning improvements highlighting priority attributes on a tactical level.
By: Marie-Louise Greijmans
How supply-chain people make Friesland Campina more customer driven
Royal Friesland Campina is a cooperative, and as ‘You cannot turn off the milk tap’, always had a very strongly rooted ‘push supply chain’. Pierre Mill shared the Burning platform within RFC SC and explained the approach to change this. He elaborated on the approach, starting with the ‘why’ story and how FC defined what success looks like. After that, he discussed how to upgrade the supply chain to become customer driven: using the people, process, organisation and systems approach with a strong base in change and stakeholder management.
By: André Vriens
How to assess your supply planning maturity?
You know your supply planning processes are not in control, but you cannot really pinpoint where the issues are, how big these issues are, and what the underlying root causes are.
During our interactive session about supply planning maturity, we’ve gone on a maturity-assessment journey together with the participants. In three steps, the participants rated their own supply planning maturity within their organization, from three different angles:
- System, people, process and data
- Based on EyeOn’s supply-planning assessment categories
This self-assessment was really intended to give participants insights in their current status, but also visibility about the maturity levels of their peers.
Now that participants know their own maturity level, of course the most important question for them is: ‘What next step will I make on my journey to world-class supply planning?’
By: Vasco Werners & Eelco Franckaert
Creating transparency within the own brand supply chain of Etos
Next to A-brands, Etos offers customers a wide selection of quality, own-branded products. To get a better grip on this supply chain, more transparency and collaboration with the suppliers was needed in order to improve on-shelf availability and the phasing-in and phasing-out of products.
During the session, Frank Antonis explained how the change of the Etos brand and store concept requires improved supply-chain monitoring and tactics. This resulted in the development of a cloud solution that uses Dataiku as a data platform and Tableau as and alert-reporting tool. A short demo showed how Etos’ flow managers, replenishers, sourcing managers and its suppliers use these insights to prevent potential issues in the own brand supply chain.
By: Willem Gerbecks
Best practices in implementing ERP and APS systems
ERP and APS implementations are often a source of disappointment for many companies. Delays, additional costs and poor user adoption generate frustration and leave corporations wondering how could they have avoided such unwanted outcomes.
During the session, EyeOn shared several insights from its experience in ERP and APS implementations to help companies anticipate the most common issues that arise during implementations. Some of the topics of greatest interest by the audience were: to be completed post session
By: Alvaro Tudanca
Embracing uncertainty: Moving to triple-A demand planning
Increasing demand uncertainty is a fact of life, especially in the consumer products industry. Based on master’s thesis research, a practical framework for triple-A demand planning has been shared to increase the ability to handle, prevent and share demand uncertainty. During the session, attendees acknowledged the different drivers of demand uncertainty in today’s environment. In order to deal with these dynamics, we discussed how to successfully apply agility, adaptability and alignment in the demand planning process.
At the end of session, there was a good discussion on where to put most effort to reach the triple-A status in demand planning.
By: Ruud van der Heijden
Forecast Value Add: Enabling trade-off between humans and machines
Organisations spend quite some time to create the perfect forecast. But what is the best forecast, and what can we best invest our time in? With research and a company case, we showed the effect of adjusting the forecasts and the effect of human behaviour.
The key findings discussed in the sessions were:
- Higher forecast accuracy does not always mean better forecast; always compare to a benchmark to see if you really add value or just add time.
- In most organizations, a lot of adjustments are made to the statistical forecast—up to 98% of all observations. This consumes a lot of time.
- The adjustment is not always better. In particular, upward adjustments tend to be too optimistic.
- You can gain insights into your enrichment by visualising and making a monthly process to review the effectiveness.
By: Koen Koolen
Boost performance while prolonging your APS’s lifespan
Many companies use APS tools that were implemented several years back. Models and parameters were set at the start, but are often not adapted to rapidly changing dynamics. Moreover, everybody seems to embrace advanced analytics to get grip on this evolving environment, and companies are afraid to fall behind. The good news is that they can buy time while getting access to the latest advanced analytics capabilities.
The session covered how companies can bridge the gap to next-level planning instantly by plugging-in proven ‘systems of innovation’. And simultaneously, they can shape their future planning landscape at their own pace.
By: Edward Versteijnen
High-quality forecasting in project-based business by using sales funnel information
Companies dealing with big-sales funnels can increase their tactical forecast accuracy by using available funnel data more intelligently. This session discussed EyeOn’s approach to arrive at better sales-funnel data usage for planning purposes.
One of the major benefits was achieved when one key party in this type of forecasting was included: Sales departments! They have all the relevant market information and can contribute by sharing their CRM data, leading to improvement in short- and midterm planning.
By: Tom Vaessen
Accelerating data-driven demand planning at Elanco
To get forecast accuracy to the desired level, Elanco launched a project to build a forecasting center of excellence. Based on a build-operate-transfer approach, EyeOn designed and implemented the new demand-planning process. This involved running a training programme for the new central analyst team in India and supporting the process operationally in the first phase.
Dylan Lilienfeld (Elanco) elaborated on how Elanco kickstarted this journey to centralised and data-driven demand planning. He shared lessons learned, e.g. on how to convince local affiliates. It is not an easy journey, but certainly worthwhile!
By: Edward Versteijnen
The road towards multi-echelon inventory optimization
The Amsterdome building originally served a different purpose: an aviation exhibition. The first planes of aviation pioneers were shown in the beautiful Amsterdome building. When flying high in the sky, you can overlook the whole world, and that’s exactly what characterises multi-echelon inventory optimisation. You take a holistic view of the supply chain. Can you think of a better place to explain the concept of MEIO to the inventory pioneers?
During the session, the differences between single-echelon, multi-stage and multi-echelon inventory optimisation were explained. As for the first aviation pioneers, the technique was not arrived at in one go. It was therefore discussed what can be done to bridge the time between single-echelon and multi-echelon inventory policy. Lessons from earlier projects were shared to give new insights into how the data quality can be improved and prepared for MEIO. Data-health checks can and will result in quick wins to improve your inventory. It will not only help your company going forward when using multi-echelon policy, but also when using single-echelon policy. Before taking off to an MEIO world, make sure that data quality is brought to the next level.
By: Maarten van Liempd
Information overload vs effective dashboarding
As data is booming, dashboards are becoming increasingly important in our daily decision making. During this session, it was proposed that often additional data actually clouds the picture rather than arms users with key information.
Effective dashboarding has been introduced as a counter solution. The presenters demonstrated the impact of effective dashboarding with two cases. As the participants experienced the process of building effective dashboards, they were encouraged to challenge their own reports and analysis based on the proposed design principles.
By: Eva Ruitenbeek & Pascal Betzel