Waitrose improves stockholding, reduces waste

The ‘Waitrose’ name has become a byword in the UK for the highest quality, freshest produce and widest product range in-store. This leading supermarket brand is now using SAS® to forecast customer demand, helping ensure the right products hit the right shelves at the right time.

Part of the John Lewis Partnership, Waitrose is one of the UK’s best-known grocers, with a brand promise based on high quality products, from fresh fruit and vegetables to packaged goods, combined with outstanding customer service. Waitrose is a unique retailer for two reasons. First, it’s the only supermarket to hold a Royal Warrant from HM Queen Elizabeth II. Second, the business is co-owned by its 39,000 employers or ‘Partners’ rather than shareholders.

With some 190 branches carrying over 15,500 product lines - and 30% of the range new each year – Waitrose needs to ensure it predicts demand accurately to supply the right products to the right stores in a timely way. Ordering the correct stock is essential, to maximise sales and ensure the least waste possible. On a regular basis, it’s also important to ensure the right stock is in place for specific events and promotions: chillers full of special-offer ice cream as the heat wave hits, shelves stacked with mince pies as Christmas approaches. Now, with the help of a SAS® demand forecasting solution and support from SAS partners, Waitrose is achieving more accurate store replenishment, satisfying the needs of its valued customers and balance sheet alike. Since the system rolled out, the branches have seen a reduction in stockholding of at least eight percent and a fall in wastage of up to four percent.

Matching supply with demand
“My team is responsible for developing new ways to support stores and their ordering, including predicting demand and enabling accurate replenishment,” says Gail Richmond, Manager - Branch Ordering Development, Waitrose. “We work with the systems team to create solutions then hand them over to the ‘business-as-usual’ people.” Waitrose had forecasting capabilities in its branch ordering system since the late 1980s that, Richmond says, “were pretty good at normal forecasting but not as good for promotions and events. So many events now affect retailing, from Pancake Day to Halloween via Mother’s Day, while promotions change things every three weeks.  She says the critical issue is “to make sure we’re hitting the shelf at the right time with the right amount of stock, not holding too much stock in the back room, and helping stores be really productive.”

Waitrose looked at various vendors in a formal tender process, narrowing six down to two. “We saw solutions like SAS that could do the whole job for us: using historical data to improve normal forecasting, observing sales and reacting to seasonality, as well as improving event forecasts,” says Richmond. Waitrose was not previously a SAS user. “One reason we chose SAS was that the other solution was a ‘black box’. You can use SAS as a ‘black box’ but it’s also very flexible: you can continue developing it to meet specific forecasting needs. We also believed we could have a good working relationship with SAS - a partnership.”

“We wanted to continue using our branch ordering system, developed in-house, but with a new approach to demand forecasting,” says Kim Newark, Forecasting Systems Manager. However, problems emerged with the system performance and the forecast models during trials says Newark. “So we involved StatApp, a SAS partner with expertise in retailing, to improve the model. The project ended up very much a partnership between Waitrose, SAS, StatApp and another SAS partner, Amadeus, the latter helping us sort out performance. After all, we were forecasting for every product line, every day, for every branch overnight.” This additional development led to further trials, which delivered greatly improved results, and the system was rolled out across the store estate. SAS models now run overnight, with forecasts automatically feeding into the Waitrose stock control and replenishment system.

More in store
“If we’ve seen an event before we can forecast demand using our historical data,” says Richmond. “If we haven’t seen it before, we still can model using a line or product in the same grouping.” Feedback from stores soon indicated that, for example, stock levels in chiller units were more efficient, with improved stock rotation. In dry goods, stores also reported higher direct-to-shelf hit rates. Improved accuracy meant further productivity gains, with order amendments in stores falling by 40 percent; and managers now had more time to spend focusing on customer service and satisfaction. Results like these, together with other measures, are having a significant impact on the bottom line.

We have achieved increased productivity, efficiency and financial gains across our 190 or so branches,” says Richmond. “With the help of SAS we have been able to reduce stockholding by at least eight percent and reduce wastage by three to four percent. We’d originally estimated a two percent drop in wastage, so this is far better than predicted. In addition, we can look at past events and see if mistakes were made, at Christmas say, and put them right next time. Or look at successes and try to replicate them. SAS is very good at providing visibility of the forecasts, to measure accuracy.”
 
A further benefit, she says, is that the knowledge gained by Waitrose and SAS, particularly in addressing a shared infrastructure, can be passed to other retailers in the UK and overseas. “From SAS’ and our own point of view, we’re happy to talk to other retailers to help. We now have a far better idea of what you need to do to forecast on this scale, turning over millions of SKUs (Stock Keeping Units) every day, and have a good working model of what the infrastructure should be, to help avoid some of the issues we ran into.”

The next step, says Richmond, is to push forecasting further back along the supply chain to predict what Waitrose orders from suppliers. “In terms of forecasting, we wanted to look at where we could achieve the biggest benefits first, which meant the branches – we knew we could make big gains there,” Kim Newark concludes. “As we move forward, we would say ‘where do we go next?’ using the same data and our forecasting in different ways.”

Copyright © SAS Institute Inc. All Rights Reserved.