The notice arrives at the end of the fiscal year: your intervention envelope is shrinking, and the decision has to be pushed down onto a portfolio of thirty, fifty, sometimes a hundred agreements already signed. You open the tracking spreadsheet and the problem is immediately visible. The "committed amount" and "disbursed amount" columns are current, but everything else is stale. Each partner's real burn rate will only be known at the next interim report — four months away for some, six for others. Findings from the last monitoring visit sit in a PDF, in a shared folder, unconnected to the budget line they concern. And the decision is due in three weeks.

This is the moment when grant portfolio triage happens by default: you cut what has not yet been disbursed, you cut the smallest agreements because they are administratively the heaviest to process, and you spare the large ones because they are the most politically visible. This article argues for the opposite approach — a method for deciding on the marginal value of each euro and on actual risk, rather than on administrative convenience. It also explains, at the end, how Abvius equips funders who want to decide on current data rather than on last year's reports.

Grant portfolio triage: deciding under constraint without breaking your partners

Reading time: ~16 min

1. Why 2026 turns triage into a core professional skill
2. The costliest reflex: triaging by partner size
3. Building a defensible triage grid: five criteria
4. Reduce, suspend, terminate: three decisions, three legal regimes
5. What the cut reveals about your information system
6. Abvius: deciding on live data, not on expired reports
7. Five steps to run a portfolio triage exercise
8. Mini FAQ

1. Why 2026 turns triage into a core professional skill

The context is no longer cyclical. Preliminary OECD figures record a historic drop in official development assistance, and projections published in 2026 anticipate a further contraction of roughly 5 to 7 % over the year. The five largest DAC donors account for almost the entire decline. In France, the 2026 finance bill confirms a reduction of around 803 million euros in the official development assistance mission, taking the envelope from 4.37 to 3.57 billion euros — the fifth consecutive cut in under two years, with the humanitarian budget hit hardest.

What these numbers change for you, concretely: revising a portfolio downwards is no longer an isolated incident handled case by case, it is a recurring exercise. And it comes with a contradictory instruction. You are asked simultaneously to shrink the envelope and to hold the line on aid localisation — that is, to maintain or increase the share flowing directly to civil society organisations in the country of intervention. A grant portfolio triage run without method produces mechanically the opposite: it concentrates funding on a small number of international intermediaries, because those are the files easiest to defend internally.

There is a second, less visible and more lasting effect. Every euro withdrawn from an agreement is not neutral: it destroys part of the partner's capacity. A CSO losing 30 % of its funding does not reduce activity by 30 % — it first loses its management accountant, its compliance officer, its ability to produce the supporting documents you will request at the final report. Six months later you get back a less well-equipped partner, on whom your supervision cost rises at precisely the moment your own appraisal staff is shrinking. That is the central paradox of the period, and it does not resolve itself with a spreadsheet.

2. The costliest reflex: triaging by partner size

In most facilities, triage under constraint follows a natural slope. You sort agreements by amount, spot the small lines, and observe that they represent 8 % of the envelope for 40 % of the appraisal workload. The conclusion imposes itself: cut the twenty smallest agreements and you free up budget and staff time. Arithmetically true. Strategically a mistake, for three reasons.

2.1 Appraisal workload is not a property of the partner

It is a property of your facility. If a 40,000-euro agreement costs as much to supervise as a 900,000-euro one, it is not because the CSO is small: it is because the same reporting template, the same validation circuit and the same level of supporting documentation apply to both. Proportionality of requirements is a lever you hold entirely. Until it has been pulled, presenting supervision cost as a triage argument means making the partner pay for an organisational choice that is yours. We develop this point in our article on the cost of supervising a grant portfolio.

2.2 Small agreements often carry the highest marginal value

A euro withdrawn from a 3-million-euro programme reduces a volume of activity. A euro withdrawn from a 60,000-euro agreement can make the organisation disappear. These are not the same decision, and they should not be taken with the same instrument. The relevant question is not "how much does this line weigh in my envelope?" but "what do I lose permanently if I cut it, and at what cost could I rebuild it in three years?". Dismantled local capacity cannot be repurchased at the price it was built for.

2.3 Excluding small organisations weakens your own accountability

If your upstream donor, your board or your supervising authority asks you to report on the share of your funding reaching local actors directly, triaging by size produces a trajectory you will have to justify. It is more comfortable to document at the point of decision why a given agreement was maintained or reduced — against written criteria — than to reconstruct the reasoning after the fact. This is exactly the logic we apply to call-for-proposals appraisal, where selecting must not amount to mechanical exclusion.

One point deserves emphasis: a local organisation's weak tooling is not a lack of seriousness, it is a budget line that has never been funded. A partner keeping accounts in a spreadsheet and producing PDF reports is doing what its funding allows. That weakness is fundable, often for a few thousand euros. Excluding it from the portfolio on the grounds that it is expensive to monitor turns an equipment problem into a verdict on competence.

3. Building a defensible triage grid: five criteria

A triage decision holds if it can be explained to three different audiences without changing version: your management, your upstream donor, and the partner concerned. That requires a grid written before the exercise, not during it. Five criteria, combined, produce a defensible ranking.

• Real burn rate to date. Not the disbursement rate — the rate of justified expenditure. A partner who has received 70 % of its tranche but consumed only 15 % at mid-term is signalling a start-up problem, not an easy saving. Distinguishing the two is the precondition for any serious triage.
• Irreversibility of commitments made. Employment contracts, leases, orders placed, obligations to third parties. Cutting an agreement where 80 % of commitments are legally firm does not free 80 % of the amount: it transfers a liability to the partner.
• Leverage and associated co-financing. An agreement that triggers co-financing from another donor does not cost its face value. Reducing it may collapse the whole arrangement — always check co-financing clauses before deciding.
• Substitutability of the partner in the area. Is there another actor able to take over the activity within the same timeframe, with the same access to populations? If not, the cut is not a reduction in activity, it is a withdrawal from the area.
• Observed management soundness, adjusted for support received. A partner with recurring findings who has never received organisational support is not comparable to a partner with the same findings after three years of accompaniment. Our article on assessing CSO financial management capacity details how to build a grid that incorporates this adjustment.

The operational difficulty is not defining these criteria — most teams have them in mind. It is populating them across fifty agreements in three weeks, when every data point sits in a different document produced by a different partner in a different format.

4. Reduce, suspend, terminate: three decisions, three regimes

A triage decision misqualified in legal terms costs more than the cut itself. Three situations are regularly conflated in internal notes, although they carry neither the same obligations nor the same timelines.

4.1 Negotiated reduction by amendment

This is the normal regime for a budgetary triage decision. The agreement continues, at a revised amount and scope, formalised by amendment. It requires negotiation with the partner on what is dropped and what is kept — a negotiation in which the partner holds information you do not: the exit cost of each activity. Short-circuiting this step systematically produces unusable residual balances and expenditure reclassified as ineligible at the final report.

4.2 Suspension

Suspension freezes disbursements without breaking the contractual link. It belongs to a risk logic, not a budgetary one: you suspend because a finding justifies stopping payment, not because the envelope has shrunk. Using suspension as a cash management tool is a frequent error under constraint, and it exposes the funder: the partner keeps incurring expenditure while awaiting resumption, and the bill eventually comes back. We cover the subject, from the funded organisation's seat, in our article on managing a grant freeze.

4.3 Termination and early closure

This is the heaviest decision and the one least often well prepared. It immediately opens a closure sequence: cut-off of eligible expenditure, inventory of assets acquired on the grant, calculation of the balance and, where applicable, recovery. The point to watch: early closure mobilises more staff time than running the agreement to term. A termination decided to save money can cost, in appraisal workload, the equivalent of what it yields. Closure mechanics are detailed in our project closure guide.

5. What the cut reveals about your information system

A triage exercise is a stress test applied to your information chain. Within days it makes visible what the routine of contracting had masked for years: you do not have a current view of your portfolio, you have a collection of photographs taken at different moments.

The mechanism is always the same. Each partner keeps its accounts in its own system. At the deadline, it extracts data, reformats it into the template you sent, and emails you a file. You re-key that file into your consolidation spreadsheet. At each step information is lost: the link to the supporting document, the exact date of the expense, the original analytical coding. What remains is an aggregate whose reliability depends on the quality of two successive re-keyings, and whose freshness depends on the reporting calendar — never on the decision calendar.

This works well enough in steady state. It becomes blocking as soon as a decision must be taken off-calendar. That is precisely what we describe in our article on harmonising partner reporting: as long as collection is about documents rather than data, no real consolidation is possible.

6. Abvius: deciding on live data, not on expired reports

Abvius is a Finance, Operations and MEAL ERP built for international solidarity organisations and for the facilities that fund them. Two parts of the product bear directly on grant portfolio triage.

The donor monitoring dashboard. Each funded CSO or partner works in its own space: it builds its budget there, records expenditure with supporting documents attached, and updates activity progress. You do not wait for a report to be produced: you have a consolidated, real-time view of your portfolio, line by line, showing the real burn rate rather than the disbursement rate alone. On the day a decision is due in three weeks, the data behind the five criteria in section 3 is already there. You do not launch a collection round: you filter. This is the direct extension of what we describe in our article on the grant portfolio monitoring dashboard.

Capacity strengthening for the organisations you support. Abvius equips partners, not only the funder — and that is what changes the economics of supervision. Your upstream donor's eligibility rules are configured once, then applied automatically across the whole contractual cascade: what used to be checked afterwards becomes blocked at entry. Consolidated reporting happens with no re-keying, at every tier. The audit trail runs down to partner level, including for sub-grants. The consequence: control and capacity strengthening stop being two separate activities funded on two different lines. The partner improves by using the tool, and your supervision cost falls structurally instead of growing with portfolio size.

This does not remove the need to make choices. It changes their nature: you decide on what you observe, not on what you assume. And you can knowingly protect partners whose capacity is critical and non-substitutable. To go further on partner equipping, see our article on strengthening CSO financial capacity beyond training, and the platform overview.

7. Five steps to run a portfolio triage exercise

Step 1 — Freeze the scope before costing anything

Start with an exhaustive list of live agreements, each with: end date, committed amount, disbursed amount, justified amount, legally firm commitments. This photograph must be taken as at a single date and circulated internally. Most triage exercises derail because two departments are working on two versions of the portfolio.

Step 2 — Separate the reversible from the irreversible

For each agreement, isolate the genuinely mobilisable share: what is neither disbursed, nor committed, nor legally due. That is your real pool. It is almost always far smaller than the gap between committed and disbursed amounts — and discovering that gap after announcing the decision is the leading source of dispute.

Step 3 — Apply the grid and document deviations

Score each agreement against the five criteria, produce a ranking, then — this is the important part — document the decisions that depart from it. There will be some, for legitimate political or diplomatic reasons. An owned and traced deviation is defensible; a grid quietly circumvented is not.

Step 4 — Negotiate before notifying

Open the discussion with each partner concerned before the official letter goes out, presenting the target amount and letting them propose how the reduction should be structured. They know their exit costs better than you do. This step adds two to three weeks to the timeline and avoids most unusable residual balances and reclassified expenditure.

Step 5 — Fund the exit when you exit

If you end a partnership, provision for closure costs: final audit, severance, archiving, asset transfer. Making the partner carry these on its own reserves means making it pay for your decision — and it is often the breaking point of an otherwise recoverable relationship. On how organisations absorb these shocks, see our mirror article, written from the funded organisation's seat.

8. Mini FAQ

Isn't an across-the-board cut fairer?

It is simpler, which is not the same thing. The same percentage applied to a 2-million-euro agreement and a 50,000-euro one produces two unrelated effects: in one case a reduction in scale, in the other the crossing of a viability threshold. Fairness here means equalising impact, not percentage.

How do I justify my decisions to my upstream donor?

Through traceability of method, not through the outcome. An upstream donor almost never contests a decision documented by a written, dated grid applied uniformly, with reasoned deviations. What it contests are decisions whose basis cannot be found six months later — at audit time.

A partner has management findings: do I cut or support?

The prior question is: has this partner ever received organisational support, and do the findings persist after it? If support was never funded, the findings measure your appraisal as much as their management. Our article on partner risk management offers a decision matrix between enhanced control, accompaniment and exit.

Is this the right moment to change monitoring tools?

Deploying a platform during the triage exercise itself, no. Deploying it immediately afterwards, yes — because that is when the organisation has concrete proof of what the absence of current data cost it, and when the portfolio is at its tightest, therefore easiest to migrate. Facilities that wait for the next cut to equip themselves experience it exactly as they experienced the last one.

Conclusion

A successful grant portfolio triage is not judged by the amount freed up, but by the state of the portfolio eighteen months later: how many critical partners were preserved, how much local capacity survived, and whether supervision cost fell or rose. Three conditions are well known — a grid written before the exercise, a rigorous distinction between the reversible and the irreversible, and negotiation conducted before notification. The fourth is technical: having a current view of your portfolio at the moment the decision is taken, and not at the rhythm of reports. In a context where cuts are now recurring, that last condition separates the facilities that steer from those that absorb.

To go further, see our articles on managing a CSO support facility and cascading sub-grants, the grant portfolio monitoring dashboard, the cost of supervising a portfolio, assessing CSO financial management capacity, strengthening CSO financial capacity and humanitarian aid localisation. To discuss your portfolio with our team, get in touch.