SYMP 10-4
Taking the long view: Mars Corp. as a business use case for a new sustainability indicators platform
The sourcing of agricultural raw materials impacts natural systems and human wellbeing in many ways. Conversely, natural and societal drivers can lead to vulnerabilities in food supply chains. Efforts over several decades have identified many indicators associated with these impacts and vulnerabilities. However, these have suffered from the lack of a unified theory of sustainability resulting in confusion over the best suite of indicators for different sustainability concerns. To address this problem, we developed a graph database tool for use by Mars Inc and other stakeholders to help select useful indicators to quantify sustainability goals. We have included in the database more than 2,000 indicators that we identified from global assessments. Each of these indicators was linked with one or more issues (e.g. biodiversity) spanning natural, social, and economic systems. To test the efficacy of the database, we conducted analyses to identify a minimum covering set of indicators for the identified issues using both Marxan software and linear programming (LP), comparing the differences between algorithms. We have also begun linking indicators to datasets that measure them and developed a prototype process for Mars on groundnuts sourced from West Africa as a use case for these tools.
Results/Conclusions
We identified 44 major sustainability issues in our reviews, and developed 344 component issues to provide details for these issues, totaling 388 issues in our database. When we ran covering set analyses with linkages between issues and any indicators that provided information about them, minimum covering sets of 23-31 indicators were found to provide full coverage of the issues using the Marxan approach. The LP approach found 15-21 indicators that provided the same full coverage. Using stricter linkage criteria, the Marxan approach identified 141-203 indicators that provided full coverage of all issues while the LP approach identified 137-191 indicators. We expect future work to correlate and structurally link issues will further reduce the required number of indicators to a more manageable set. While these results can provide useful information to Mars and other users regarding the best indicators to track, the specific context of a given query will undoubtedly shape the results. Linking GIS data to some of these indicators for groundnut-producing regions of West Africa is an example of how a user such as Mars could best utilize this platform to further their sustainability goals. Future work will also address the usefulness, credibility, and legitimacy of the database indicators.