Hempel: Improving Forecast Accuracy

Hempel, a world-leading supplier of trusted coating solutions, significantly improves forecast accuracy with sofco Demand Planning.

As a world-leading supplier of trusted coating solutions, Hempel is a global company with strong values, working with customers in the protective, marine, decorative, container and yacht industries. Hempel factories, R&D centres and stock points are established in every region.

Across the globe, Hempel’s coatings protect surfaces, structures, and equipment. They extend asset lifetimes, reduce maintenance costs, and make homes and workplaces safer and more colourful. Hempel was founded in Copenhagen, Denmark in 1915 and is proudly owned by the Hempel Foundation, which ensures a solid economic base for the Hempel Group and supports cultural, social, humanitarian, and scientific purposes around the world.

Why sofco?

sofco are able to support Hempel globally, with an excellent understanding of the business model and issues. The software is scalable and adaptable, providing the visibility and extensive, reliable functionality that Hempel needed.

The Challenge

Hempel required a single world-wide platform and model solution to forecast requirements from all customers (both project and maintenance) into a consistent demand stream. As part of the project, we had to validate and control which SKU’s should be, “make to order”, and which “make to stock”, and provide an accurate forecast to Microsoft AX, and then D365 for “make to order” items.

As the project developed, Hempel embarked on an internal re-structure of their business and thus the reporting regime required changes. The addition of the new attributes and the modification of the affected hierarchies had to be carried out without significantly changing the existing forecasts and associated adjustments to minimise the disruption to the business. Furthermore, the removal of the Business Unit required the statistical cleaning and forecasting to be carried out at the Company level instead.

The Solution

The initial project at Hempel involved producing a forecasting solution that would provide reliable, understandable forecasts that could be compared to actual sales, budgeted volumes, and project volumes from the CRM system, and future orders for the bigger volume items, whilst allowing the business to identify and refine the changeover point between ‘make to order’ and ‘make to stock’ items. Whilst this model was initially to cover the Europe Region only, there was also an extra requirement to create a generic solution that would cover the other Regions as well so that it could be rolled out world-wide.

This global roll-out followed within three years as the additional three Regions of Asia, Middle East, and Americas were added. We implemented the model structure as originally developed, replicated across the three additional databases / cubes. The additional needs of the extra Regions were addressed using new worksheets.

A subsequent internal re-structure of the Hempel business to split it between Marine and Industrial, and of the reporting regime from Country-based Business Units to Cost Centres, required changes to the structure of the Market dimension.

Additionally, the removal of the Business Unit level of the existing hierarchies required the statistical cleaning and forecasting to be modified so they are carried out at the Company level instead, and all the screens in the model had to be reviewed and, where necessary, to change the filter selections and any Row and Column use from Business Unit to Company. The removal of the BU and the modification of the affected hierarchies had to be carried out without significantly changing the existing forecasts and associated adjustments to minimise disruption.

This process was executed in parallel with the other internal system changes that Hempel were making, including the transition from Microsoft AX to D365, and had to be ready for use when the data was available and not stop the availability of the forecasts when the data changed in the external systems.

As the business, and world trading conditions, have evolved further amendments have been made to the model during the last two years.

For example, the basis of the statistical cleaning and forecasting processes was changed from being SKU (Item) based to Saleable Item (Multi-Component) based to address some inconsistencies between Base and Cure forecasts in less frequently sold combinations.

In addition, we had to set the default forecasting process in Europe post-Covid to address the huge variations in past sales histories and introduce a separate global BI Reporting dataset to allow drill-down to the lowest level of data across and Market, Multi-Component, Item, Time combination without slowing the performance of the forecasting system to allow business-wide visibility of the latest forecast.

We also had to provide the ability to create and maintain a ‘frozen period’ in the statistical forecast to prevent any changes within the production window but allow visibility of changes in the forecast, then allowing manual changes if authorised.

After 7 years and numerous fundamental changes to both the model, the world trading environment, and the Hempel business, the sofco solution has proved to be as adaptable and scalable as they hoped and is the world-wide standard for all Industrial (non-Deco) forecasting.