Working in a cluster and coordinating in the field

1The cluster's tools

• The 4W (who does what, where and when; sometimes 5W with “for whom”): each partner reports its activities by area; the cluster uses this to produce presence and coverage maps.
• Gap analysis: comparing needs (people in need, severity) with coverage to steer partners towards forgotten areas.
• The cluster's standards and technical guidance: kit composition, transfer values, protocols, adapted from Sphere to the context.
• The monthly dashboard: people reached against HRP targets.

Filling in the 4W on time is not a formality: it is what makes your work visible and allows you to defend funding for your area.

2A useful coordination meeting

• An agenda sent in advance, with decisions to be taken, not just rounds of updates.
• Data shared beforehand (4W maps, gaps) so as to discuss solutions rather than read out figures.
• Short minutes with the decisions, the people responsible and the deadlines, shared within 48 hours.
• Meetings in a language understood by local actors, with interpretation if needed, at accessible times and places.
• Co-facilitators from NGOs, including national ones, alongside the lead agency.

3Authorities, the military and local actors

Authorities: involve them in meetings, share plans and aggregated data, respect procedures (authorisations, registration), while preserving impartiality (lists are determined by needs, not affiliations).

Civil-military coordination (UN-CMCoord, facilitated by OCHA): dialogue is necessary for security and access, but there must be a clear distinction between humanitarian and military actors. Military assets are used only as a last resort (Oslo Guidelines for disasters, MCDA Guidelines for conflicts), under civilian control, with no escort by default.

Localisation: national and local organisations know the terrain and remain after the crisis. Involving them in cluster leadership, sharing funding (the Grand Bargain set the target of at least 25% of funding to local actors as directly as possible) and strengthening their capacities are part of good coordination.

4Cluster information management: the 4W step by step

The 4W is the most widely used and the most poorly completed tool in coordination. A reliable 4W makes it possible to see duplication and forgotten areas, to calculate coverage against the response plan targets and to make the case for funding. An approximate 4W produces misleading maps that steer resources to the wrong place. Quality depends as much on the partners who fill it in as on the cluster's information management officer (IMO).

What a well-designed 4W contains

Counting people reached: the unique individuals rule

The cluster distinguishes two figures. The number of people reached per activity (the same person who receives a hygiene kit and attends a promotion session counts twice) is used to monitor activities. The number of unique individuals reached, which is the figure compared with the response plan target, counts each person only once per period. The usual method is, for each geographic unit, to take the activity that reached the most people, or to deduct known overlaps between partners. The cluster guidance should specify the rule; if it does not, ask for it, because the differences can be as large as a factor of two.

The typical monthly timeline

1. From the 1st to the 5th of the month: partners enter their data for the previous month in the cluster tool (standardized file, online platform such as ActivityInfo, Kobo form).
2. From the 6th to the 10th: the IMO checks consistency (outliers, duplicates, invalid P-codes, incomplete disaggregation) and sends queries back to partners.
3. From the 10th to the 15th: production of information products (presence map, coverage against targets table, dashboard), shared before the monthly meeting.
4. Around the 15th: submission to OCHA for the intersectoral dashboard and response plan monitoring.

Key point: for a partner, the 4W must be generated from its own monitoring system, not reconstructed from memory on the day of the deadline. If the cluster's categories do not match your internal indicators, map them once and for all at project start-up.

Data responsibility: what to share and what not to share

The IASC Operational Guidance on Data Responsibility in Humanitarian Action sets out a clear principle: share the data needed for coordination, at the highest level of aggregation that allows the intended use. In practice:

• The 4W never contains personal data: no names, no beneficiary lists, no GPS coordinates of households.
• Sensitive data (protection incidents, gender-based violence, precise locations of at-risk groups) follow specific protocols, defined by the protection cluster and its areas of responsibility, and do not appear in public products.
• Each cluster should have an information sharing protocol, which classifies data by sensitivity level (public, restricted, confidential) and specifies who can access what. Sharing of beneficiary lists between organizations (for example for the deduplication of cash transfers) takes place under a data sharing agreement, with a legal basis and security measures.
• Data requests from the authorities or armed forces are handled collectively, at cluster or HCT level, and not by each organization on its own.

5Full example: from the coverage map to the decision to reorient

Fictional context: Sarré district, Haut-Doma province, Republic of Lomandi. The food security sub-cluster brings together eight partners, including two national NGOs. The response plan target for the district is 83,000 people receiving monthly food assistance (in kind or cash) during the lean season. You are the sub-cluster co-facilitator for an international NGO and you are preparing, with the IMO, the gap analysis for June.

Step 1: gather the data

You cross-reference three sources: people in need and severity from the response plan analysis (updated by the season's IPC analysis), the May 4W, and qualitative information from partners (access, supply chain breaks, population movements).

Step 2: calculate coverage and the gap

Step 3: check anomalies before drawing conclusions

A coverage of 115% in Teni almost always signals a data problem rather than genuine over-coverage. After checking with the two partners present, it turns out that one is distributing cash transfers to 9,600 people and the other is distributing seeds to 4,200 people, almost all of whom are also cash recipients. Seeds belong to another activity (livelihoods support). The real number of unique individuals receiving food assistance is therefore 9,600, i.e. 80% coverage. The correction brings the district's total reached down to 37,200 people (45%) and the gap up to 45,800.

Step 4: prioritize

The sub-cluster uses a simple score, endorsed in a meeting at the start of the season: score = severity x (1 - coverage). The score is not an automatic decision; it serves as a basis for discussion.

Step 5: understand the causes of the gaps

• Mboro: access restricted by an armed group since April; one partner has funding but cannot go there. The problem is access, not money.
• Wassa: no partner present; a national NGO from the neighbouring commune knows the area but has no funding.
• Doré: the main partner suffered a funding break in March. The problem is financial.

Step 6: the meeting's decisions

What the example shows: a useful gap analysis does not stop at percentages. It checks the data, distinguishes the causes (access, funding, absence of partners) and leads to different decisions for each cause, each with an owner and a deadline.

Reorienting a partner: what the cluster can and cannot do

A cluster has no hierarchical authority over its members. It cannot force an NGO to change area. It does, however, have real levers: transparency (a map showing aid concentrated in a commune that is already covered), the technical review of proposals submitted to pooled funds, and dialogue with donors, who increasingly ask partners to justify their areas against cluster priorities. A partner that agrees to reorient must do so with its donor's agreement and while informing the communities and authorities of the area it is leaving, to avoid tensions.

6Preparing, facilitating and following up a coordination meeting: the complete method

A cluster meeting that brings together 30 people for two hours costs the equivalent of 60 hours of senior staff time. It is only justified if it produces decisions that no email could have achieved. The method below applies to cluster, sub-cluster and working group meetings.

Before the meeting (D-7 to D-2)

1. D-7: the coordinator and the co-facilitator list the questions that require a collective decision (geographic priorities, endorsement of a standard, response to an alert, review of proposals). Anything that is purely for information is sent in writing.
2. D-5: circulation of the agenda with, for each item, its nature (information, discussion, decision), the time allotted, the person presenting it and the associated documents.
3. D-3: circulation of information products (4W map, coverage table, draft standard) so that participants read them beforehand.
4. D-2: bilateral contacts with the partners concerned by a sensitive decision (reorientation, disagreement over a standard), to avoid surprises in the meeting.

Typical agenda for a 90-minute monthly meeting

During the meeting: facilitation techniques

• Separate presentation from discussion: the data have been read beforehand; the meeting is for interpreting them and deciding.
• State the expected decision at the start of each item: "We need to choose two priority communes out of the four proposed."
• Manage the floor: give the floor first to national NGOs and remote participants, who speak up less spontaneously; limit roundtables to the items that warrant them.
• Decision rule: consensus is sought; failing that, the cluster's terms of reference usually provide for a decision by the strategic advisory group or the coordinator, with any disagreements recorded in the minutes.
• Manage conflicts: bring the discussion back to the data and the endorsed criteria; refer bilateral disputes to a separate exchange; never allow an organization to be publicly called into question without the opportunity to respond.
• Language and accessibility: interpretation or a meeting in the local working language, short documents, an accessible venue, a remote connection for partners in distant areas.

After the meeting: the minutes and action point follow-up

The minutes are circulated within 48 to 72 hours. They are no more than two pages long and include: the list of participants by organization, the numbered decisions, the action point table and, where necessary, any disagreements expressed. The action point tracking table is reviewed at the start of every meeting.

A simple indicator of a cluster's quality: the proportion of agreed action points that are actually completed by the deadline. Below half, the cluster is producing meetings, not coordination.

Alerts and rapid response between meetings

Sub-national coordination must work between meetings. Most clusters have an alert mechanism: a partner or an authority reports an event (influx of IDPs, epidemic, fire in a site); the coordinator verifies the information within 24 to 48 hours, triggers a joint rapid assessment with the other sectors if necessary, then calls on the partners present or a rapid response mechanism where one exists. Trigger thresholds (for example a movement of more than a few hundred households, or an epidemiological threshold defined by the health cluster) are set in advance to avoid debates at the moment of the emergency.

7Authorities and armed forces: rules, criteria and practical tools

Working with the authorities without losing impartiality

National and local authorities have primary responsibility for assisting and protecting the population. They issue authorizations, register organizations, sometimes set priorities and can facilitate or block access. A regular working relationship is essential. The most common points of friction are predictable and can be prepared for collectively:

Civil-military coordination: the texts and the criteria

Humanitarian civil-military coordination (UN-CMCoord), led by OCHA, pursues three objectives: information sharing, deconfliction (preventing military operations from endangering humanitarian workers and civilians) and, in strictly limited cases, the use of military assets. The reference texts are:

• The Oslo Guidelines on the Use of Foreign Military and Civil Defence Assets in Disaster Relief.
• The MCDA Guidelines on the Use of Military and Civil Defence Assets to Support United Nations Humanitarian Activities in Complex Emergencies.
• The IASC Reference Paper on Civil-Military Relationship in Complex Emergencies (2004).
• The IASC Non-Binding Guidelines on the Use of Armed Escorts for Humanitarian Convoys (2013).
• Country-specific guidelines, adopted by the HCT, which in practice take precedence in the field.

The last resort principle breaks down into cumulative conditions: there is no comparable civilian alternative; the military asset is unique in its capability or availability; it meets a critical, life-saving humanitarian need; its use is limited in time and scale, with an exit strategy; it remains under civilian control; and the assistance retains its civilian character. In conflicts, preference is given to infrastructure support (repairing a bridge) or indirect support (transporting goods to a warehouse) rather than direct assistance to populations by military personnel.

Armed escorts: a decision tree

Under the 2013 guidelines, the default rule is not to use armed escorts. An exception can only be considered if four criteria are met, in addition to last resort:

1. Humanitarian need and priority: the convoy meets life-saving needs that cannot be met in any other way.
2. Responsibility of the sovereign authority: the escort is preferably provided by the authority responsible for security, and not by a party to the conflict perceived as hostile by the population.
3. Safety of personnel: the escort genuinely reduces the risk, without exposing the convoy to being targeted because of it.
4. Sustainability of access: the escort does not compromise future access, the perception of neutrality or the security of other organizations.

The decision is taken at the appropriate level (usually by the HC or the HCT on the advice of the security adviser and OCHA), documented and reviewed periodically. An NGO may have its own, more restrictive policy: many refuse any armed escort, whatever the circumstances.

Practical deconfliction tools

• Humanitarian notification systems: organizations notify the parties to the conflict, through OCHA, of the location of their fixed facilities and of their movements, so that the parties are informed of their protected status. Notification is not a request for authorization and does not guarantee safety.
• Designated points of contact: a single civilian interlocutor (OCHA) between the humanitarian community and the armed forces, so that each organization does not deal with them directly.
• Rules of conduct: no transport of armed personnel in humanitarian vehicles, no uniforms or weapons on premises, clearly distinct logos, no participation in "hearts and minds" activities.

Key point: the question to ask before any operational contact with armed forces is not "is it effective today?", but "what will the consequences be tomorrow for the perception of neutrality, other organizations' access and the safety of the population?".

8Common mistakes, contextual variations and the donors' perspective

Common mistakes made by cluster partners

• Filling in the 4W late or incompletely, then disputing maps that do not show your work. Visibility depends on the quality of the data submitted.
• Reporting planned activities as completed, which hides gaps and delays the reorientation of resources.
• Acting alone towards the authorities or armed forces on an issue that affects all actors (taxes, lists, escorts): an isolated concession becomes the norm imposed on everyone.
• Ignoring cluster standards (kit composition, transfer values), which creates tensions between neighbouring communities and puts other partners in a difficult position.
• Using meetings to promote your own organization rather than to solve common problems.
• Neglecting intersectoral coordination: distributing hygiene kits in a site without water, or seeds without coordination with cash transfers.

Variations by context

What donors and auditors look at

• ECHO asks in its Single Form how the action is coordinated with other actors and the clusters, and its field technical experts often attend cluster meetings: they know whether an organization is present and active.
• BHA asks applicants to demonstrate coordination with the clusters and alignment with sectoral standards, particularly for cash transfer values and minimum expenditure baskets.
• Pooled funds generally require projects to be aligned with the response plan and the organization to report to the 4W; the cluster's technical review weighs directly on selection.
• Evaluators check consistency between areas of intervention and cluster priorities, the absence of duplication, adherence to standards and the use of community feedback.

For an audit or verification, keep evidence of your participation: minutes mentioning your organization, 4W submissions, correspondence on geographic distribution. These documents justify your choice of areas and your costs.

9Interview questions and key references

Interview questions

1. "The 4W data show 130% coverage in a commune. How do you interpret this figure?"

A good answer first points to a data problem (double counting between activities or between agency and partner, confusion between monthly and cumulative figures, outdated target), describes the verification with the partners concerned and the unique individuals rule, then, if the over-coverage is real, the discussion on reorientation.

2. "How can a cluster meeting be made more useful?"

The candidate talks about identified decision points, documents sent in advance, controlled duration, action point follow-up, prompt minutes, participation of national NGOs (language, timing, remote connection) and an indicator tracking completed action points.

3. "A local authority insists on approving your beneficiary lists. What do you do?"

The answer explains what is at stake for impartiality, proposes involving the authority in the criteria and in community verification without handing over selection, takes the issue to the cluster and OCHA for a common position, and documents the exchange. It mentions the complaints mechanism as a safeguard.

4. "Under what conditions would you accept military support to deliver aid?"

The answer cites last resort and its conditions (no civilian alternative, unique capability, life-saving need, time limitation, civilian control), distinguishes between disasters and conflicts, recalls the default rule against armed escorts and the four criteria of the 2013 guidelines, and specifies that the decision is collective and that the organization's own policy may be more restrictive.

5. "What data are you willing to share with the cluster?"

The answer distinguishes between aggregated data and personal data, and cites the IASC guidance on data responsibility, the information sharing protocol and data sharing agreements for specific cases (deduplication of transfers).

Key references

• Cluster information management guidance and standards, and 3W/4W tools: OCHA and the global cluster support teams.
• Operational Guidance on Data Responsibility in Humanitarian Action: IASC.
• Common Operational Datasets (CODs) and place codes: OCHA.
• Oslo Guidelines and MCDA Guidelines: OCHA.
• Civil-Military Relationship in Complex Emergencies: an IASC Reference Paper (2004) and Non-Binding Guidelines on the Use of Armed Escorts for Humanitarian Convoys (2013): IASC.
• Humanitarian Civil-Military Coordination field handbook (UN-CMCoord Field Handbook): OCHA.
• Camp Coordination and Camp Management toolkit: Global CCCM Cluster.
• The Sphere Handbook (2018 edition), in particular the standards on coordination, and the Core Humanitarian Standard (CHS).