How GUD Holdings saved planning time by moving beyond Excel
Case summary
Before this project, planning at GUD Holdings was largely manual and time consuming.
Demand planning relied heavily on Excel spreadsheets maintained by planners. Forecasts existed, but there was no single, trusted view. Different versions circulated, assumptions were hard to trace, data analysis limited, and planners spent a lot of time reconciling numbers rather than improving them.
Capacity planning was largely Excel-driven and highly time consuming. Complex spreadsheets were used to model varying production line requirements, changeovers, and shift patterns. Due to this, the MPS Planners had to work on plans for weeks at a time and often worked additional hours in order to load the factories on time. Simulations or changes to the plan were not easily available, to make quick decisions and would generally require the business to wait for the new planning cycle to run.
A rolling average forecasting approach did not account for trends, seasonality, outliers, or promotional activity. Safety stock policies were also Excel-based, which limited how frequently they could be reviewed or adjusted. As a result, planners were often forced to intervene manually where possible, rather than steering inventory through a consistent, data-driven logic. Inventory buffers were sometimes increased, particularly on imported items, to manage uncertainty and lead time risk.
The objective of this project was not to replace people or processes overnight, but to automate planning in order to balance the planners’ workload, provide simulations when required promptly for decision making, have better control of inventory changes resulting in reduced working capital, therefore improving inventory health overall and optimizing production through the factories.
Company context
GUD Holdings operates in a manufacturing environment where demand variability, capacity constraints, and inventory trade-offs are closely linked.
In this context, planning is not just about producing numbers. Demand signals need to be reliable enough to support production and inventory decisions. Capacity plans must reflect real constraints rather than theoretical limits. Inventory policies need to balance service and working capital without relying on excessive buffers.
The people driving the change
This was not a tool-first project. Progress depended on the people involved and their willingness to question existing assumptions.
On the GUD side, the core team included supply chain leadership, demand and supply planners, and IT, each bringing a different perspective:
- Marizka Nothnagel, supply chain director, set the direction and priorities.
- Josiah Pillay & Rickardo Naicker, both being supply planners, focused on capacity and production feasibility.
- Thiloshinie Reddy, demand planning manager, worked hands-on with forecast generation, validation and improvement.
- Shane Nair, Software Services Manager, supported data flows and integration.
- James Naidoo, Information Technology Director, supported the process and helped clear system bottlenecks and connectivity issues.
- Gervaise Roberts, Group Demand and Supply Planning Manager, contributed from operational and planning roles.
The starting point: heavy manual work and limited visibility
Demand planning
Demand planning was largely Excel-based. Planners manually built and adjusted forecasts, often without a clear view of what was driving changes.
Planners spent significant time cleaning historical data, removing the impact of promotions, and trying to smooth peaks to avoid overreacting to short-term spikes. In practice, many of these adjustments happened after peaks had already occurred. This led to one-time inventory increases that were difficult to unwind, because the same peak did not repeat or demand shifted to different locations.
Effort went into maintaining spreadsheets rather than providing data-driven insights or consistently improving forecast quality.
Capacity planning
Capacity planning relied on complex Excel spreadsheets to develop a capacity model that accounted for varying production line requirements, changeovers, and shift patterns.
MPS planners were required to ensure the files were accurately updated each month and that the correct data flowed through multiple tabs to appropriately constrain the plan and load the factory. Due to the volume of data and the number of formulas required, the spreadsheets became increasingly slow and difficult to manage.
The files also included unconstrained planning views, further increasing complexity. This placed significant pressure on the planning team and often required planners to work additional hours beyond their normal schedules to meet deadlines. Scenario planning was particularly time consuming, as it relied on manual manipulation of data, resulting in delays in decision-making.
Inventory planning
Variability in forecast accuracy, combined with a rolling average forecasting methodology, made it difficult to consistently align inventory levels with changing demand patterns. Demand forecasting relied on a rolling average methodology, with no management of outliers or adjustments for trends, seasonality, or promotional activity, largely due to the Excel driven process. Frequent updates to safety stock policies were constrained due to it also being an Excel driven process. This meant the planners had to manually intervene were possible.
Phase 1: establishing a structured planning foundation
The first phase focused on three tightly connected areas: demand planning, inventory and capacity planning.
Demand planning: building a reliable baseline
A system-generated demand baseline was introduced and benchmarked against the previous manual approach using GUD’s own historical data.
Initial results indicated meaningful improvement versus the Excel-driven forecasting setup. However, the most important shift did not come from the baseline but from having one system in which all of the data is used to generate the best possible forecast. Whenever somebody makes a change, this is directly tracked instead of being lost in an Excel.
With a clearer demand signal in place, the team could start structuring capacity and production planning in a more practical way. The key improvement is not in loading less capacity, but in reducing the effort required to balance and validate the plan.
Inventory planning: aligning buffers with reality
With demand and capacity plans aligned, inventory planning could be revisited on a more structural basis.
Safety stock calculations were introduced using actual demand variability and supply behavior, rather than relying on static or infrequently updated Excel-based policies. This removed the need for planners to manually adjust buffers each time forecasts or capacity assumptions changed.
More importantly, inventory decisions are now consistently linked to the same demand and capacity assumptions used elsewhere in the planning process, rather than being managed.
Implementation reality: iteration before automation
The project did not start with full integration.
Initial testing was done using manual files to validate logic, assumptions, and outputs. This allowed planners to understand how demand, capacity, and inventory interacted before automating anything.
Where GUD is today, and what comes next
GUD now has the structural foundation in place across demand, capacity, and inventory planning. Improved transparency and reduced manual effort are key benefits for the planners.
The improvements are still materializing as the organization progresses through phased implementation, but the direction is clear: less manual effort, stronger data control, and faster scenario evaluation.
The next phase will expand this foundation into distribution planning and detailed scheduling, continuing the same phased approach.
Rather than attempting a big-bang transformation, GUD is building planning maturity step by steps validating each layer before moving to the next. That is what makes the change sustainable.
