The software, supplied by SAS, is designed to put an end to under or overstocking around key events and holidays by predicting customer shopping habits.
It enables factors such as the weather, the economy, market conditions, customer demographics and marketing activities to be quantified and used to forecast stock levels, rather than basing ordering purely on historical data.
Rob Thompson, head of supply chain at Waitrose, said event forecasting would enable it to predict demand for thousands of products more quickly and accurately. "We anticipate that the return on investment we will achieve through reduced wastage and increased revenues will be significant," he added.
The system is currently forecasting for 20,000 SKUs in five stores. It is being expanded to a further 20 before Christmas, with a plan to roll out the software to all shops next year.
A fruit and vegetable manager in one of the trial stores said his order amendment rate had been halved, while one fresh food manager added that stock levels had significantly reduced with a much higher "direct to shelf hit rate" now achieved. A spokeswoman for SAS said the system was able to take into account if a particularly hot day had been forecast. It would then predict, for example, how many ice creams a store would need to cope with the likely extra demand.
Waitrose initially trialled the system between July and November last year, monitoring what stock levels SAS would predict against the supermarket's traditional ordering methods and actual sales.
Weather forecasting has been crucial this summer, with record temperatures in July. Kevin Hawkins, director general of the British Retail Consortium, said it was essential big stores took advantage of the best forecasts on offer. He added: "This helps retailers plan several weeks in advance."