Salad supplier turns over new leaf in fight against waste

One of the world’s largest suppliers of bagged salads, soups and ready-to-eat meals has reduced the margin for error in its daily sales forecast by 10,000 boxes.

That’s according to Fresh4cast, a UK-based technology provider that has worked with the leading supplier over the past year to review its commercial planning.

As chief executive officer Mihai Ciobanu explains, the multinational was able to identify “significant improvements” in its longer-term planning, its levels of collaboration between teams, and the overall accuracy of its data.

“As the business is very complex, their legacy system only revised the medium-term weekly forecast once a quarter, and the daily forecast once a week,” Ciobanu says.

“Fresh4cast updates the weekly forecast once a week and the daily forecast every day, reflecting the latest challenges and opportunities.”

The new tool also allows the group to estimate demand more accurately as far as 52 weeks ahead. The machine-generated baseline for this “planning horizon” is based on a million individual pieces of data that are updated weekly.

A crucial connection

The technology has apparently helped the supplier to adopt a more joined-up approach with the data it already had available.

Previously, a more detailed seven-day demand plan was updated without any affect on an entirely separate eight-week prediction.

“Fresh4cast’s bespoke algorithm converts the weekly demand plan into an input for the sales planning team, who always get the latest numbers and visibility over a longer time horizon,” Ciobanu explains.

“This allows planning processes done by different teams to integrate seamlessly, with updated numbers immediately available and easy to use.”

The speed of the new system has reportedly made for a transformed supply chain. Whereas a change typically took the old system ten minutes to update, the same change now takes ten seconds.

Hands-free and spot on

The need to make extensive manual adjustments to daily sales forecasts has also been removed, with errors in daily numbers generated by Fresh4cast now five times lower than before – equal to a forecast error reduction of 10,000 boxes per day.

“Planners now have more time to see the big picture, look into details and make better judgements,” Ciobanu notes. “For example, the promotional impact had to be quantified manually and applied separately in the legacy system. In Fresh4cast, it is modelled in the machine forecast and displayed alongside a non-promotional baseline, so that planners can see it clearly and make informed, fast decisions.”

The result, he adds, is a more competitive customer and another step forward in terms of sustainability in fruit and vegetables.

The next step for both companies? Time to start planning material requirements with the same speed and precision.