Humanitarian Trendlines Data to Sep 2026

A lens on the humanitarian sector

What the humanitarian sector hires for, year by year

Every job post on ReliefWeb since 2011, read by a language model and tagged against 65 topics. Topic trends are shown as a share of all posts in each year, so the overall rise and fall in the number of job posts does not read as a rise or fall in a topic.

–job posts
–topics
–largest organisations tracked

All ReliefWeb job posts per year

Hot since Jan 2025

Topics
Organisations

Topic trends

Pick a group, then up to six topics.

Measure Count

Core means the topic or theme is central to the role, usually named in the title or purpose. Core or duty adds jobs where the topic is a substantial part of the work but not its main subject: broader, and less reliable. Core is the safer default. Four topics are counted as core only: Partnerships & partner capacity, Needs assessment & analysis, Safety & security management, and Compliance, audit & risk. They are part of so many roles that counting them as a duty would add noise. Interoperability & data sharing is mostly a duty rather than a job in itself, so read it as core or duty.

Part-years are drawn dashed.

Biggest movers

Four ways to slice the jobs

Junior or senior? Staff or consultant? UN or NGO? Experience, job type and career category are what the advertiser chose on ReliefWeb's posting form. Organisation type starts from ReliefWeb's label for each organisation and splits NGOs into international, locally based and regional, using each NGO's home country and where it hires.

Countries Beta

Where jobs are advertised, and how that has shifted. An advert listing several countries counts in each. This is a first version: breakdowns by topic and organisation for each country will be added.

Measure

Biggest movers

Organisations

The 150 organisations that post most on ReliefWeb. Posts by several organisations count once for each.

Job posts on ReliefWeb by this organisation

Topic concentration, recent

Topic share in this organisation's posts ÷ share across all posts. 2.0× means twice the sector rate. Topics with fewer than 10 of its posts are left out.

    Who's hiring for what: last 6 months

    Who's hiring since 2025

    Select a row to open it above.

    How to read this

    The short version. The full methodology covers sources, checks and how to cite; the glossary defines every term and topic.

    What is counted

    • Advertised posts on ReliefWeb, including expired ones. Not hires, and not all humanitarian hiring.
    • Every ad was read by a language model (OpenAI GPT-6 Luna) and tagged against a list of 65 defined topics, in English, French and Spanish. A topic counts when the job itself involves it, not when the word appears in the organisation's description of itself or in standard recruitment wording.
    • Each tag has a level. Core: the topic or theme is central to the role. Duty: the topic is a substantial part of the work but not its main subject. Skills asked for and passing mentions are not counted.
    • A specialist topic also counts under its parent: sexual and reproductive health, epidemics and MHPSS under Health; MPCA, cash coordination and group cash transfers under CVA. Do not add a parent to its specialists.

    Cautions

    • Use shares for trends. Total posting fell sharply in 2024–25, around the US funding cuts, so counts fall for almost every topic.
    • Unusually quiet months are flagged in the data. The cause cannot be told from this data.
    • Some organisations moved recruitment to their own sites. A collapse in one organisation's volume may mean that rather than a hiring freeze.

    Topics to read carefully

    • The tags are a model's reading, not a hand count. In checks, 98% of its core tags were confirmed by a stronger model, and it found 87% of the core topics that model found. Counts therefore run a little low; trends are more reliable than levels.
    • Core or duty is noisier: about three in four duty tags were confirmed.
    • Research, studies & evaluation covers research and study roles and evaluation consultancies in general, not only rigorous impact evaluations.
    • Ads with almost no text (about 0.3%) were not read and carry no topics.
    • Exploratory themes are small topics. They are shown to two decimal places, left out of rankings, and should be read as signals.
    • Rankings need at least 20 matching posts in each period compared.