Hema
Reduce safety stock in supply chain by managing & improving demand forecast performance
Client & Challenge
HEMA is maybe the strongest retail brand in the Netherlands. It is a trusted household name for affordable, unique HEMA branded, retail products & food. It has a large presence in the Netherlands with hundreds of stores, and a growing online footprint online.
The great strength of HEMA is also a source of great complexity in the supply chain. HEMA has to stock hundreds of stores with, with uniquely branded products (no substitutes), sourced across the globe, and with a large amount of new launches each year. Getting this right, means getting stock levels right, as well as your demand forecasts on the short- and mid to long term.
HEMA has invested into this topic using, among other things, state-of-the-art forecasting engine. There was a hypothesis that more is possible.
This raised a crucial question: How can HEMA manage and improve the demand forecast performance across a million product-store combinations and reduce the capital in stock as a consequence?
Our Approach & Solution
Borg led a project team of HEMA supply chain specialists, data science and forecasting expert to find out how we could create “grip” on, and structurally improve its performance to reduce safety stock.
The solution consisted of three parts.
- Make it easy for management to monitor and invest resources on specific areas of the demand performance that needs to be improved.
- Measure the quality of data, feeding into the forecasting engine, to surface the reason why the model does not perform.
- Couple the performance, to the “safety stock setting” – so that better performance forecast reduce the amount of safety stock
Measure and monitor
In order to monitor performance, we need to collectively define it. In workshops with supply chain we mapped all the “horizons” on which sourcing decisions are made for the different product groups and agreed on metrics of performance.
The second step is, making the scope smaller and top-down to make it manageable. Understanding which forecasts are more important to get right compared to others.. E.g., slow moving products on store level cannot be forecasted. Secondly, there is a great difference in revenue contribution across products.
Contact us to learn how we can help your business harness the power of data and AI to drive innovation and achieve tangible results.
Other Cases
FD Mediagroep
How can FDMG create strategic impact with AI over the next two years, and how should it be organized?
Sunweb
How can we make smarter, more dynamic decisions on when to rely on internal stock and when to switch to external flight capacity — in a way that balances margin, availability, and risk?
BridgeFund
How could BF use its wealth of data and technological capabilities to enable faster, more accurate and responsible credit risk assessments?
