The World Food Programme moved from manually combining fragmented operational data to a connected analytics system used across emergencies and country operations. Across this wider programme, WFP reported more than $150 million in savings, an amount it estimated could fund food assistance for around two million people for a year.
Why was fragmented information making humanitarian decisions harder?
WFP operates across more than 80 countries, where teams make decisions about food, procurement, transport, funding and delivery under rapidly changing conditions.
The information needed for those decisions was often spread across different systems. A planner assessing one operation might need to combine data on food requirements, procurement contracts, transport costs, available funding, nutritional values and local market prices before understanding the full picture.
Early analytical tools could bring this information together, but doing so was labour-intensive. Preparing a dashboard or running an optimisation analysis could require at least a week of data collection, cleaning and validation. This meant the tools were initially concentrated on a small number of major emergencies and relied heavily on specialist teams at WFP headquarters.
The challenge was therefore larger than developing individual analytical models. WFP needed a way to make reliable data and analysis available more consistently across the organisation.
Why does this matter from a prescriptive analytics perspective?
Understanding what is happening is only the first stage of a humanitarian decision. WFP also needed to anticipate problems and decide what to do next. Whether stock should be moved, which food should be purchased, which transport route should be used or how limited funding should be allocated.
WFP therefore developed three connected capabilities:
the Supply Chain Management Dashboard, which brings together information about current operations and emerging problems;
Optimus, which uses prescriptive analytics to compare food-basket and supply-chain decisions;
DOTS, which integrates data from different systems so that these tools can operate at a larger scale.
Together, they connect three questions: What is happening? What is likely to happen next? What action should be considered?
That distinction matters because a dashboard alone can show a future shortage, but prescriptive analytics can also compare ways of responding to it.
How did WFP and its partners build the solution?
The transformation developed over more than a decade through collaborations involving WFP, Tilburg University, the University of Amsterdam, Northeastern University and Georgia Institute of Technology, alongside private-sector partners such as UPS.
Researchers and WFP teams explored applications including network design, routing, food-basket planning, price forecasting and optimisation. These collaborations gradually moved from research prototypes towards tools used within WFP operations.
Private partners collaborated with WFP on DOTS, the organisation's data-integration platform. One of its first applications connected the information required by Optimus from more than a dozen systems. The resulting pipeline contained more than 260 linked tables that were synchronised each night automatically.
UPS and the UPS Foundation contributed to the wider rollout of Optimus, including change-management experience as WFP expanded the use of optimisation beyond specialist teams.
The collaboration was therefore not centred on transferring one technical solution into WFP's ways of working. Rather research, operational knowledge, software development and organisational change developed together over several years.
What changed when analytics became part of operational planning?
The biggest change was that analysis could be used repeatedly rather than assembled from scratch for individual crises.
The Supply Chain Management Dashboard created a common view of each operation, including planned food distributions, available commodities, funding gaps and emerging delays. It could flag issues such as stock expected to expire or shortages likely to affect future distributions and identify possible actions for teams to consider.
DOTS reduced the amount of manual work required to prepare the underlying data, while Optimus allowed teams to compare alternative supply-chain decisions.
WFP also changed its internal decision processes. Cross-functional groups brought together colleagues from areas such as logistics, procurement, nutrition and resource management around the same operational information rather than having each function analyse the situation separately.
By early 2021, more than 40 Optimus power users were distributed across WFP regional bureaux, creating capacity to run and interpret optimisation analyses closer to the operations where decisions were being made.
What did this mean for humanitarian operations?
The wider analytics programme produced results across several different emergencies.
In Iraq, optimising food baskets reduced monthly costs by 12% without compromising nutritional value, generating more than $25 million in savings over two years.
In South Sudan, analytics informed changes to procurement, transport and pre-positioning, placing food closer to where it would be needed before seasonal roads became inaccessible. WFP reported savings of more than $100 million from changes to the supply network.
Reliance on costly airdrops fell substantially. Food delivered by airdrop decreased from around 65,000 metric tonnes in 2018 to 22,000 tonnes in 2019, as more supplies were moved in advance by road or through river routes.
That change was not simply a transport saving. During the rainy season, large parts of South Sudan can become inaccessible by road for months. Expanding river transport and pre-positioning allowed food to remain available in locations that could otherwise become cut off.
What does the $150 million figure represent?
Across the wider analytics programme, WFP reported more than $150 million in savings. Back in 2024, WFP estimated that this was equivalent to the resources required to provide food assistance to around two million food-insecure people for an entire year. This achievement was only possible because of the impact of multiple tools and operational changes, rather than one sophisticated optimisation solution.
Its significance is that operational savings translate directly into additional capacity when humanitarian resources are limited. Lower procurement or transport costs can allow existing funding to cover more food assistance, reduce the effects of funding shortfalls or release resources for other operational needs.
Why was change management as important as the technology?
WFP's experience also showed that a mathematical model does not automatically become part of everyday decision-making. Early versions of Optimus required specialists to prepare the data and interpret the results. WFP therefore invested in a web-based interface, automated data pipelines, training and a distributed network of users who could apply optimisation within their own regions.
The tools were also embedded in cross-functional decision forums. Logistics, procurement, nutrition and resource-management teams could examine the same evidence and understand how changing one part of the operation affected another.
This organisational change is an important part of the impact. Analytics moved from being an occasional technical exercise towards becoming part of how operational alternatives could be discussed and compared.
What can this mean for future humanitarian operations?
The WFP experience shows that scaling analytics requires more than increasingly sophisticated models. Reliable data infrastructure, tools that non-specialists can use, people who understand both the analytical methods and operational context, and decision processes that bring different functions together are all part of the same system.
The next analytical questions also involve greater uncertainty. Food prices, transport availability, conflict and climate conditions can change after a plan has been developed. Later research and solution development are therefore examining how food-assistance plans can remain flexible as new information becomes available.
The broader direction is towards analytics that can be updated as conditions change while remaining integrated with the judgement of the teams making operational decisions. For humanitarian organisations working with limited resources, that can mean moving from analysing individual problems after they emerge towards identifying risks earlier and comparing possible responses before resources are committed.
