“Our very first forecast told us that we could hold £2 million less in safety stocks of fast-moving products.”
“The whole world was saying that we can’t forecast this business,” says Anne Brug- gink, General Manager Supply Chain at Elec- trocomponents.
Most statistical forecasting systems require expert forecast plan- ners to constantly tune the best fit algorithms to produce a rea- sonable statistical forecast and wind up either being ‘dumbed down’ to the point that they are ineffective or even shelved because they are too difficult to use. SO99+ uses a unique demand sensing approach to continuously and automatically tune the forecasting method for each item. The result is an extremely reliable statistical forecast before any user intervention – essential to Electrocomponents for forecasting so many dfus and their long tail of slow moving items. After SO99+ produced the most accurate forecast in the plan- ning tournament, Electrocomponents knew it could trust and work with the system. According to Bruggink, “A crucial factor for us was the way in which SO99+ handled order-line data.
It
differentiates between 10 orders for one product and one order for ten products and this has a major impact on the forecasting accuracy.”
Electrocomponents has now rolled out the SO99+ software in Europe, where it is integrated with Manugistics, and it is imple- menting it in the US, where it will be integrated with SAP. Elec- trocomponents has further plans to implement SO99+ in Asia.
tomer demand, which is characterised by geographical diversity, small order quantities and specialised products. Because the sys- tem couldn’t detect this type of demand, it reported zero demand for 30 per cent of its inventory which people knew was not the case.
In 2009, Electrocomponents embarked on a comprehensive search for SIOP-compatible software to manage the forecast and safety stocks. It needed to be powerful, highly scalable to cope with millions of demand forecasting units (DFU’s), and highly reliable. The company opted to enlist a third party consultant to manage the vendor selection process to ensure the decision would be completely objective and fact-based. Bruggink elaborates: “We decided to conduct our evaluation through a third party to avoid being influenced by personal biases, large or incumbent vendors and winding up with a ‘me too’ solution. Given all our challenges and business goals, we were looking for a very specialised, very sophisticated ‘Formula 1’ forecasting tool, ideally from a company whose main develop- ment and support resources were dedicated to solving this spe- cific problem.”
Planning tournament 38
The exhaustive 14-month selection process short-listed seven vendors who were invited to take part in a planning tourna- ment. They were asked to forecast demand and propose safety stocks from historical data. The results were compared to actual demand in order to evaluate the forecast accuracy and service level which would have been achieved. The software SO99+ from ToolsGroup provided the most accurate forecast (beating even the company’s own heavily modified forecast) and achieved service levels which were not possible with the current system.
Exceptional outcomes
By taking advantage of the burning need for change during the global economic downturn, completing the Site Specialisation in the UK, transforming forecasting and initiating other projects, Anne Bruggink’s supply chain operation delivered a whopping £36 million to the Electrocomponents bottom line, in terms of lower stock levels, over 2009 and 2010.
The most important long-term outcome of the new forecasting system is simply that it works. Its consistent forecasting reliabil- ity has overcome people’s initial scepticism and it effectively sup- ports the new input from exceptional planning processes. The system frees up forecast planners to concentrate on what they do best; contributing market and business intelligence from the field to make the system even smarter. In terms of savings, Electrocomponents has identified £9 mil- lion of savings by deferring or cancelling outstanding purchase orders and by reinvesting additional inventory into the business. Crucially, the SO99+ tool paid for itself very quickly. As Bruggink explains, “Our very first forecast told us that we could hold £2 million less in safety stocks of fast-moving products. By simply reducing those safety stocks, SO99+ paid for itself in only two months. We had promised the business that the tool would pay for itself in the same financial year, but we were thrilled to have been able to return the money to the business so quickly.” Probably the most important outcome, however, is that a cus- tomer survey held in November 2010 reported the highest ever satisfaction rates. This meant that despite the fact that the com- pany was coming out of the worst recession since the 1980s, had had to navigate ash clouds and earthquakes, and was holding considerably less inventory, it was still able to meet its high level service delivery commitments.