Overview
ITE Consult is a consultancy company that delivers cloud planning solutions and supports decision tools merging planning, forecasting, simulation, and optimization features.
A fast-growing company in the hair and body care industry needed to implement and support a sales and operations planning (S&OP) platform in order to improve and integrate their demand and supply planning processes.
Problem
A company experienced difficulties with a production plan because they had only two weeks projected using Excel. Supply planning was done manually at different aggregation levels.
They wanted to bring raw materials, finished goods, and packaging materials to a safety stock level. Also, the company started to have some labor restrictions as they expected an increase in their retail business of around 50%.
In addition, the company needed to set its production capacity and production rates. They had two production models, (DTC) make to order and (Retail) make to stock, with different demand trends and production restrictions but sharing the same production process. On top of that, the customer wanted to include inventory costs in the optimization process.
Solution
ITE Consult implemented a sales and operation planning platform with all the functions for planning the different custom aggregation levels.
Consultants created a SPOT (single point of truth) with all relevant data (historical data, budget, and forecast). They transferred governance of master data to the business user so they could create data for new products and clients.
ITE Consult developed the AnyLogic model that allowed the company to run a rough-cut capacity plan (RCCP) and play different scenarios. They wanted to handle alternative demand inputs because they had different projections.
The S&OP process cycle
The processes in the S&OP technical platform were repeated monthly. This cycle started with a demand planning process. They added “building blocks” to the statistical forecast. Then there was a demand planning review. Once demand planning was approved, the next step was the simulation and optimization cycle, where different scenarios could be run.
The engineers added constraints regarding labor and production lines. Over a few days, they ran scenarios to choose the most appropriate. Then they moved on to the supply planning review. They discussed on the site what plans they wanted to push up to the executive S&OP meeting. Once they got to that point, it was the end of this process. The stakeholders agreed on sales, operational, and financial plans.
The S&OP process cycle is demonstrated below.
The blue cycle was powered by SAP Analytics Cloud. The orange cycle was supported with the AnyLogic production optimization software and Gurobi. The process was embedded into the system, so the system informed managers about actions and when it should be ready every single month.
They ran four typical scenarios for demand and labor. The client usually had two demand forecasts: with marketing lifts and with a bigger raise in “make-to-stock” because the company was negotiating with new retailers.
On the operational side, the biggest constraint was the labor. Thus, the developers were running two operational scenarios: three shifts for the entire projected period and one shift due to labor shortage.
There were five steps to exploring these business scenarios and getting the outputs of these few runs with the use of AnyLogic as demand planning software.
Step 1: Demand forecasting and planning
Modelers used predictive analytics to generate the forecasting baseline for a demand planning process based on historical data and “outside-in” influencers. They just needed to click on the button, and the system would create their statistical baseline.
The stakeholders had the ability to create the “building blocks” with different departments. The information remained in the system, so the changes that were made one month ago in the statistical model could be seen.
Step 2: Supply network optimization run
Once they had their demand planning in place, the modelers opened the simulation model and selected the forecast version and how many months they wanted to run the model.
Usually, they ran the simulation and optimization process for a year or 18 months. After the data was loaded into the system, they had to select the initial inventory date for simulation. Then the engineers selected parameters for the scenario.
For DTC and the retail processes, engineers could change the lines and rates per month. Also, they could change the shifts, for example, the number of shifts by month and hours per shift, considering breaks and weekends.
The model developers wanted to keep their inventory between two levels: the minimum (safety stock) and maximum WOS (weeks of supply), thanks to the optimizer and simulation. Depending on the demand, weeks of supply were calculated.
After they set all the parameters, the modelers ran the solver. The AnyLogic production simulation software sent parameters to Gurobi, and it sent back the solution in two or three minutes. Once the process was completed for 12 or 18 months, the modelers could export the data back to SAP.
Step 3: Review supply planning scenarios
Then the rough-cut capacity planning (RCCP) scenarios, generated by the simulation and optimization tool, could be analyzed. These scenarios included a finished goods production plan, raw material requirements, and inventory level projections.
In the picture below, the upper chart shows the demand and production for each month. The information was processed by day, even for the monthly plan. The inventory level was always between the minimum and maximum WOS. There were several months when it was at the top of the safety stock.
In the lower chart, the raw material projection is shown. This chart starts with a high amount of inventory. The production risk was growing slowly until there was a safety stock level of raw materials.
Step 4: Review inventory projections
The next step was to analyze inventory scenarios, including finished goods, raw materials, and packaging materials, comparing historical and projected inventory cost levels. The system was reducing the inventory to safety stock levels, which was a big saving for the company.
Step 5: Review Labor & Machine Usage
After that, the specialists compared different scenarios, running the model, to see what happened with one or three shifts. Engineers also had labor information to identify how many people were involved in each production step for 18 months. Also, it was possible to observe machine usage and see how many hours every single machine was being used. The bottleneck in production was not capacity but labor.
Results
Thanks to AnyLogic’s production simulation software, the company significantly improved forecast errors for demand planning. In addition, they got a single point of truth for analytics and KPIs.
On the supply side, the company had a weekly production plan for 18 months with no stockouts. Also, by using AnyLogic models, the stakeholders had material requirements for this period, considering lead time and minimum batch sizes.
The reduction of inventory to safety stock levels was a huge saving for the company. The stakeholders were able to look at the labor restriction projection and how it would impact inventory, costs, and service if there were fewer shifts on the lines. These were the biggest savings for the customer, thanks to simulation modeling.
The case study was presented by David Kennedy, of Ite Consult, at the AnyLogic Conference 2022.
The slides are available as a PDF.
