Drugstore chain Matas automates order process
Danish drugstore chain Matas has automated its forecasting and replenishment process. “Ordering was previously done manually, often based on gut instinct rather than hard facts,” said Johnny Rolsted, supply chain analyst at Matas, during the SAS Analytics conference in Cologne, Germany, on 14 June 2012. This led to a work-intensive and hence time-consuming replenishment process, a high inventory value and many out-of-stock situations for crucial items.
Having started out as a cooperative in 1949, Matas has since grown to a chain of 296 stores in Denmark and two in Sweden which together generate an annual turnover of half a billion euro. The retailer sells a wide range of both branded and own-label goods in four market segments: beauty products, vitamins and dietary supplements, household products, and pharmaceutical products. All of its stores are supplied from the 15,000-square-metre distribution centre in Allerød, which is also the home of the company’s headquarters. The drugstore chain’s ownership structure changed in 2007 when it was acquired by CVC Capital Partners, a private equity firm, and the focus has increasingly been on standardisation ever since.
Before the acquisition, Matas was a cooperative with 180 owners who managed the stores individually, and they were also responsible for placing their own orders. “The orders were predominantly based on gut instinct and sales figures from the previous month, while the distribution centre’s order proposals were based on sales from the distribution centre and then adjusted by hand,” explained Rolsted.
Matas calculated how much turnover it was missing out on due to products simply not being on the shelves in its stores. When it emerged that 40 percent of out-of-stock situations involved fast-moving items, the retailer decided to automate the ordering process. “We were keen to obtain forecasts based on Point Of Sale (POS) data and to reduce out-of-stocks, especially on key items,” Rolsted continued.
Matas called in outside help from SAS, a specialist in business analytics software and services, and implemented a new system which would generated the orders automatically based on POS data, stock levels, orders and product range. “With the SAS Forecast Server, we are now generating two million forecasts for SKU/store combinations in the first echelon every week, and 26,500 forecasts per SKU in the second echelon every day,” stated Anders Richter, senior analytical consultant at SAS in the joint presentation with Rolsted.
By using SAS Inventory Optimization, Matas aims to ascertain the optimum size of repeat orders and to identify the minimum and maximum stock levels. “However, Inventory Optimization cannot aggregate orders at supplier level. Hence if there are any restrictions, such as a minimum order size, we use SAS Operations Research,’ said Richter.
“Our expectations were really high, but this…” said Rolsted, his words trailing off yet leaving no doubt that the solution had exceeded them. The total inventory value was reduced by ten percent and the number of out-of-stock situations declined by two percentage points, while the number of man-hours spent on replenishment was cut by 70 percent. The replenishment flows have become coherent and the DC order proposals are now semi-automated. “We now have access to facts rather than having to rely on gut instinct,” concluded Rolsted.