The Hidden Millions in Humanitarian Logistics

Logistics and supply-chain activities are estimated to account for 60–80% of humanitarian intervention expenditure [1]. At this scale, even modest efficiency gains could free up hundreds of millions of dollars for frontline assistance. And this is more than a theoretical possibility.

The opportunity is not simply to make humanitarian logistics more efficient. It is to reach more people with the resources already available.

The scale of the opportunity

According to the United Nations’ Global Humanitarian Overview 2026, approximately 239 million people need humanitarian assistance across the plans and appeals covered by the overview. Humanitarian organisations aim to reach 135 million people, requiring approximately US$33 billion [2].

The scale of humanitarian logistics is equally striking. WFP alone supported 121 million people in 2025, operating supply chains across more than 100 countries [3]. Meanwhile, UNHCR reports that 117.8 million people were forcibly displaced worldwide at the end of 2025, including 41.6 million refugees and 68.7 million people internally displaced by conflict and violence [4]. In 2024, UNHCR dispatched 5.1 million emergency relief items, supporting approximately six million people affected by displacement and emergencies [5].

Behind these figures are vast supply networks operating under extraordinary pressure. Every decision about procurement, storage, transportation and distribution affects how efficiently assistance reaches those who need it.

From predicting needs to optimising decisions

Commercial logistics companies have used advanced analytics for decades to optimise warehouses, inventory, transport routes and delivery schedules. Humanitarian organisations face similar challenges, but under far more difficult conditions: uncertain demand, disrupted infrastructure, security risks and competing humanitarian priorities.

Predictive analytics can anticipate where assistance will be needed and how demand may evolve. Prescriptive AI goes further, identifying how to deploy available resources most effectively within operational constraints.

Predicting increased demand for medical supplies is valuable. Determining where to pre-position those supplies, how much stock to hold and which routes will reach communities most effectively is what turns prediction into impact.

At Analytics for a Better World (ABW), we bring together humanitarian organisations, academic researchers and industry partners to turn these analytical capabilities into practical solutions.

What we are learning from practice

Our collaborations with leading humanitarian organisations demonstrate how optimisation can improve decisions across different parts of the humanitarian system.

Our collaboration with the World Food Programme (WFP) demonstrates what optimisation can achieve at scale. ABW supported the development of the Route the Meals toolkit, using advanced analytics to optimise food sourcing, procurement and supply-chain logistics while meeting nutritional requirements at the lowest possible cost. In 2021 alone, WFP’s optimisation efforts reportedly generated over US$150 million in savings, enough to provide food assistance to an additional two million people for an entire year [6].

With Médecins Sans Frontières (MSF), ABW developed a simulation model to assess alternative configurations of its international supply network. By testing warehouse locations, inventory policies and distribution strategies before implementation, MSF can identify opportunities to reduce costs while safeguarding the availability of critical medical supplies [7].

Together with Amref Health Africa, ABW developed a decision-support tool for mobile clinics in Kenya. Combining geographical information, predicted healthcare demand and route optimisation, it helps identify how mobile clinics could reach underserved communities more effectively [8].

With Fleet Forum, we explored opportunities to optimise humanitarian personnel mobility. In Lebanon, ABW and Fleet Forum worked with eight humanitarian organisations to develop a mobility planning tool, exploring how shared transport and alternative mobility options could reduce costs, vehicle movements and environmental impact [9].

Four lessons emerge from these collaborations:

• Start with the operational problem, not the technology. Focus on concrete decisions that need improving. Technology should serve the challenge, not drive it.

• Involve practitioners in designing the solution. Combining analytical expertise with operational knowledge is essential. At Amref, practitioner feedback helped refine the model to prioritise communities with greater healthcare needs.

• Don’t wait for perfect data. Start with available information and improve iteratively. With Fleet Forum, we developed an initial solution despite fragmented data and different systems across participating organisations.

• Optimise humanitarian outcomes, not just costs. The cheapest or fastest solution is not always the best. Models should also account for reliability, accessibility, equity and other humanitarian priorities.

Together, these projects illustrate the breadth of optimisation opportunities, from global supply networks to last-mile healthcare delivery. While their results are at different stages of implementation, they demonstrate how better decisions can improve humanitarian efficiency, coverage and effectiveness.

The next frontier: making optimisation accessible to all

The greatest opportunity is to make advanced analytics and AI-powered optimisation part of everyday decision-making across the humanitarian sector. Better warehouse locations can reduce transport distances, smarter inventory management can prevent shortages and waste, improved routing can lower delivery costs, and coordinated procurement can eliminate duplication. Ultimately, connecting these decisions could enable organisations to optimise entire supply chains, balancing cost, speed, reliability and humanitarian impact.

If logistics represents 60–80% of humanitarian expenditure, a hypothetical 5% efficiency improvement could release resources equivalent to approximately 3–4% of total humanitarian expenditure. Across a sector spending tens of billions of dollars annually, this could mean hundreds of millions available for additional assistance. These figures illustrate the potential, rather than forecast actual savings.

The next challenge is to democratise access to these capabilities. WFP is exploring with ABW how to make the Route the Meals toolkit open source and accessible to humanitarian organisations worldwide. Similarly, ABW has developed PISA, an open-source geospatial decision-support platform with the World Bank, helping organisations optimise the location of essential services and improve accessibility.

But technology alone is not enough. Humanitarian organisations also need a new generation of in-house analytics and AI specialists who can identify opportunities, apply these tools and embed them in daily operations.

Making optimisation mainstream requires overcoming several barriers: fragmented and often poor-quality data, limited funding for digital infrastructure, a shortage of specialised talent, and organisational cultures in which operational decisions are not yet routinely supported by advanced analytics. Equally important is ensuring that optimisation respects humanitarian principles, including equity, accessibility and the needs of the most vulnerable.

This is central to ABW’s mission: developing open analytical tools, applying them to real-world challenges and building lasting capabilities within humanitarian organisations, together with academic and industry partners.

As humanitarian funding comes under increasing pressure, optimisation should become a core investment priority, not an optional technical exercise. The technology exists, and its potential has been demonstrated. The next frontier is ensuring that every humanitarian organisation can benefit, so that every euro or dollar, every vehicle and every shipment can help more people.