From humble beginnings in 1911 when Frank C. Mars made his first Mars candies in his Tacoma, Washington kitchen, Mars has grown to become a global brand with net sales of more than $30 billion across six business segments. Almost 4,000 of the global workforce work of 70,000 work in the UK manufacturing brands including MARS, SNICKERS, ORBIT, PEDIGREE, DOLMIO, KLIX and FLAVIA. The Mars facility at Basingstoke is responsible for manufacturing KLIX cup drinks and FLAVIA single serve drinks and at the heart of this hugely complex scheduling challenge sits Preactor Advanced Planning and Scheduling (APS) software and the new Preactor Graphical Master Production Scheduler (GMPS).
Mars Basingstoke supplies demand to other Mars business units in the US, EU and Japan which means that it has to manufacture a very large amount of product across a wide range of possible variations. For example, the ‘in cup’ division has 6 dedicated machines that produce an average of 400 cups per minute across a triple shift pattern which commences 7am on a Monday and ends 7pm on Friday. There are 99 permutations of 33 SKUs with some machines being dedicated to soups, others to tea etc. On the ‘single serve’ division there are 10 machines that produce approximately 100 units per minute across a similar range of permutations. To add to the complexity, different countries require different formulations as well as different pallet sizes/amounts.
Another key challenge lies in handling the diverse ways that demand comes from different markets, as Paul Hazelwood, Supply Logistics Manager explains. “When dealing with UK and EU demand, we receive a forecast breakdown to an individual product level within a given time period, typically a rolling 20 week basis. For the US and Japan however we work on a Make to Order basis, again over 20 weeks, but with a 4 week locked window when levels can’t be changed. And whereas orders for the UK, US and Japan are scheduled on a weekly basis, the EU is scheduled on a less regular basis.” All finished products irrespective of final destination are routed via the company’s Andover facility.
Given the scale of production, scheduling this accurately would be a challenge in and of itself but as Hazelwood continues, there are many other factors that have to be taken into consideration. “We have a fixed number of machines running over a 3 shift period so optimising sequencing is critical to maximising capacity usage. “ He continues, “We have other considerations such as product grouping and allergen control to take into account in order to minimise changeover and clean down times and maintain a smoothness of supply. Furthermore, certain products have specific storage requirements.
For example, our fresh roasted coffee has to be packed with a given number of days of the coffee being roasted. Different suppliers have different delivery times and this therefore impacts the start times of such orders and all of these variables have to be factored into both our high level and detailed scheduling.”
With over 26 years’ experience at Mars, Hazelwood was only brought into the project as the company was implementing Preactor APS and GMPS from Preactor International. He outlines how the company used to try and meet the scheduling challenges. “We previously used a system called Schedule X but this dealt with a rolling 20 week horizon, it only dealt with 24 hour operations. The schedule used to be based on buckets of data and was viable because more stock made it easier to adjust the schedule. The combination of Mars IS withdrawing its support for the systems and wanting become leaner, which meant using more accurate data and reducing stock levels, a better solution was called for.”
For Mars, that solution was Preactor APS and GMPS and when Hazelwood arrived on the project in November 2011, various internal decisions left him in his own words, “holding the baby and thrown in at the deep end with the project.” It is a testament to Hazelwood’s dedication, the intuitive nature of the Preactor solutions and the support of the Preactor team that in the 8 weeks remaining before “go live”, that the system was fully developed, tested and implemented for Mars’ unique requirements. “There was a lot of hands-on training and a large amount of scenario testing and I was essentially walked through the project by the Preactor team until and beyond the go-live on February 20th 2012.”
The project is still in its early days and Hazelwood admits there is still work to do given the scale of the implementation but there is already a marked improvement. All demand forecasts and orders are now received and imported into Preactor GMPS along with the relevant sales and stock data from which a rolling 20 week high level schedule is generated. Hazelwood again, “GMPS looks at high level scheduling at a weekly level and ensures that we remain within target levels across all our SKUs, taking into consideration our high level constraints. From here we export the GMPS data into our Preactor APS which then handles the detailed scheduling down to what needs to be done, where and when, taking into account our detailed level constraints. Once the schedule has been exported to Preactor, we generate an MRP report which we use to order our materials.”
The most immediate benefit from Preactor GMPS has been the lowering of stock levels across the entire pipeline. While still too early to quantify the full impact of this, Hazel wood cites the example of one particular SKU where stock levels have been reduced from 10,000 pallets down to 4,000. Applied across all SKUs this represents a very considerable cost saving. Because of the interlinked nature of the Preactor GMPS and APS solution, where the impacts of every decision can be seen on everything else, Hazelwood also says that the company now works much more closely, at a global level. “Anyone involved in the pipeline has visibility of the entire pipeline and this increases confidence and trust at every level.”
Looking to the future, he is adamant that while further ongoing refinement is called for, “the combined benefits will be very good once everything is fully working.” Hence he concludes, “Preactor GMPS is driving internal behaviour change and we are definitely already seeing Lean Improvements. We have a long road ahead but there are undoubtedly big benefits in the future.”