Predictive Analytics - A Way to Prevent Food Waste

Food waste is a real challenge, and we all have a co-responsibility: The individual consumer, the retailer, as well as the distributor and the manufacturer. By applying business data science on a data foundation established through increased collaboration in the integrated value chain, we can reduce food waste by optimizing the supply chain and still deliver the right products to the right consumers at the right time and place. 

UN's Sustainable Development Goal 12.3

By 2030, halve per capita global food waste at retail and consumer levels and reduce food losses along production and supply chains, including post-harvest losses.

In the fast-moving-consumer-goods industry (FMCG), supply and demand must balance better to reduce food waste. All parts of the integrated value chain focus on producing, distributing, and delivering the right products to consumers in the right amount and at the right time and place. However, the major challenge is the insufficient sharing of data between the parties in the value chain. This gives poor conditions for predicting and making the right decisions. A better data foundation and increased transparency between all parties can reduce food waste.

From a traditional point of view, the parties in the value chain have no interest in sharing knowledge with others. However, knowledge obliges. To meet the UN’s Sustainable Development Goal 12.3 of halving food waste and food losses in production and supply chains, we have to find alternative ways to share data and forecast consumer demand.


The challenge is to identify which data may be shared between the parties - without capitalizing on one another and without compromising private consumer data.

Data Exists but Isn't Connected or Ready for Analytics

Retailers have detailed information on their consumers' purchasing behavior through cash vouchers and their socioeconomic characteristics through loyalty cards. Price elasticity and substitution effects are relevant factors in the customers' buying behavior. This is also the case for the shop's floor plan and other marketing initiatives. 

The manufacturer, on the other hand, prepares demand forecasts on the basis of its customers' (i.e. retailers) sales history, sales orders, and marketing initiatives, potentially in collaboration with the retailer. Calendar activities, such as holidays, vacation time, events, etc., are also applied information.

From a predictive analytics perspective, massive amounts of data exist. It's collected but not fully used by relevant parties in the value chain, and for this reason the entire ecosystem doesn't benefit. In order to significantly reduce food waste, existing data must be available to a greater extent and be useful for larger parts of the integrated value chain.

The large amount of data and the many variables must be incorporated into the forecast and processed using business data science. It combines economic knowledge, statistics and machine learning, and is a powerful weapon for identifying data relationships and finding causal relationships; the key to more accurate predictions. 

Untraditional Initiatives are Needed

Predictive analytics combined with economic theoretical knowledge from the field is undeniably the right tool to solve the problem. The challenge is how to provide and display the information to the parties in the value chain without distorting the market mechanisms or violating the GDPR.

The challenge is to establish a constellation between the parties in the value chain, each contributing with data. By orchestrating a data foundation which can be shared between the parties in the value chain, they can optimize their own activities more accurately.

Consumer to manufacturer consent must be given for certain data while the data for predictive analytics must be compiled in such a way that the necessary degree of data privacy exists. This applies, for example, to data relating to consumers and brands.

Technically, a data foundation must be prepared by associating various variables, and then applied in predictive analytics. This applies to the variables creating the extra information, which makes it possible to perform better forecasts - throughout the value chain.


Are you curious about how we find data correlations and causality in data when working with predictive analytics, contact Giovanni Mellace.

Book a meeting


Rooftop Analytics has the competences to analyze and identify which data makes the difference when forecasting. We utilize statistical analysis rooted in knowledge of economic theory. We handle big data and apply machine learning - all necessary elements for predictive analytics, required for companies to enable the fight against food waste.

Read more

When multiple technologies come into play the synergies arise.

Rooftop Analytics stands out on the market and creates value for companies.