Promotions are the lifeblood of so many grocery items, especially perishables. But they’re also the bane of demand planners.
The whole point of a promotion is to grab the consumer’s attention. But it doesn’t translate into increased sales unless the manufacturer and retailer collaborate to get the right amount of product on the shelf at the right time.
It’s a lesson that Granarolo S.p.a. knows all too well. The Italian producer of dairy and milk products runs thousands of promotions each year, generating some 34,000 forecasting combinations. As a result, it might trigger demand up to 30 times greater than baseline sales. In a given year, promotions can account for up to 75 percent of sales, making accurate forecasting essential.
Founded in 1957 as a milk producer’s consortium, Granarolo has since grown into a major force in the Italian and international grocery market. In Italy, it’s ranked first for dairy products, first for fresh milk, second for ultra-high-temperature (UHT) milk and third for yogurt and cheese with long shelf lives. With nine production plants, it produces around 750 million liters of milk per year, according to demand planning manager Dario Pettinato. Annual turnover is about €1bn.
The perishable nature of Granarolo’s products means the company has little time to react to shifts in consumer buying habits. “Our main challenge is to plan production and demand in a really fast-moving scenario,” says Pettinato.
No More Spreadsheets
Rapid growth in sales has made the task of demand planning all that more difficult. Spreadsheets used to be sufficient for getting the job done, says Pettinato. But they didn’t lend themselves to an integrated approach, and they couldn’t respond quickly enough to upticks in sales. In addition, they weren’t designed to handle the large numbers of SKUs that the company has had to juggle.
Even as a smaller cooperative, Granarolo was making use of a demand-planning application from ToolsGroup, known at the time as DPM. That application later evolved into the vendor’s SO99+ software, which Granarolo opted to implement on a more systematic basis. Says Pettinato: “We decided to take advantage of the knowledge that the company has already shared with our group and their consultants, in order to make the system and process easier.”
The timing was fortuitous. Not only was Granarolo coping with dramatic growth in its markets, it was also experiencing a shift in the profile of its customers. Italy, says Pettinato, “comes from a history of small retail shops, which in the last 20 years have been aggregating. We’re moving to big stores.” The change has been motivated in part by an incursion of French distribution companies, which have imported a business model tailored to their own country’s network of hypermarkets.
SO99+ is especially well-suited to the needs of perishable products. Rather than simply minimize stock or maximize margins, the tool employs an algorithm that allows for mixed optimization, says ToolsGroup chief executive officer Joseph Shamir. Service levels can be adjusted by product category. As a result, product flows more smoothly and in the right quantities, lengthening residual shelf life.
Implementation of SO99+ took place in stages. By the time Pettinato joined Granarolo in 2010, the basic shaping of the system was complete. His challenge for the next three years was to move from the forecasting of 300 to 400 major SKUs to cover the company’s entire product line of some 1,000 items, with an eye toward addressing replenishment and fulfillment as well. In addition, he needed to configure the tool to analyze the impact of promotions.
For trade-promotion forecasting, Granarolo combined S099+ with Rulex machine-learning software. The latter “clusterizes” promotions to give an optimal picture of sales uplift, Pettinato says.
A Learning Experience
The new technology could learn the correlation between all of the variables that affect increased sales driven by a promotion, including advertising, flyers, coupons and other tactics. Some of that functionality was adopted by ToolsGroup specifically for Granarolo. “We were able to take this learning technology and transform it into a time-phased capability,” Shamir says.
Acquisition of the system turned out to be a learning process for both people and machine. “The deeper you can analyze a product, item or process, the more you can understand it and manage it,” says Pettinato. “That’s the main goal that a demand manager or planner has to face day by day.”
Forecast validation is essential. “Every demand planner has to help to make the forecast more accurate,” Pettinato says. Granarolo worked hard to simplify the tool and related reports so that it could provide production and sales with one clearly understood number for demand. It incorporated key performance indicators that keep all departments within the company on track.
Granarolo has seen results on multiple fronts. With the help of the ToolsGroup application, it has raised average forecast reliability from 80 to 85 percent, peaking at 95 percent for fresh milk and cream, and 88 percent for yogurt and desserts. Inventory levels have dropped by more than 50 percent, halving capital and lead time. Delivery time has shrunk by 50 percent, prolonging product freshness on the shelf. The company has also seen significant increases in customer service and sales, while transportation costs have dropped.
There’s more work to be done. Pettinato sees room for improvement in the areas of fulfillment and replenishment. His group is looking to simplify the modeling of some product hierarchies, in order to obtain even more accurate results. In addition, Granarolo wants to get a better grasp of the downstream process, with an eye toward teasing out further efficiencies.
For its part, ToolsGroup is moving toward increased use of trade-promotion optimization, which helps a producer to make better decisions about its trade spend. TPO allows other functions within the business to draw on the same demand-uplift data for planning their own activities, says Shamir. Another ToolsGroup customer, Danone, is already using that capability.
The ultimate goal is better service with less waste. “With such a perishable product,” says Shamir, “if you’re able to deliver it fresher, thanks to better forecasting and shelf-life optimization, then that translates into economic benefits in terms of the price you’re able to get for the product.”
Timely, incisive articles delivered directly to your inbox.