Let the technical logic allocate the money
Weak budgets often begin with a target share for each partner. The consortium negotiates percentages and later tries to fit tasks around them. This produces artificial symmetry: institutions with very different experimental burdens, facilities and validation responsibilities receive similar amounts because equal distribution feels politically easier.
A strong budget begins with the proof-of-principle logic. Where is the breakthrough created? Which experiments carry the highest uncertainty? Where will integration fail if the team does not allow enough iteration? Which evidence must be independently reproduced? Resources should concentrate around those questions.
This does not mean that every technical partner receives a large budget or that management and impact work are unimportant. It means that the allocation can be explained causally. The consortium should be able to point from each substantial cost to a responsibility, from the responsibility to an output and from that output to a project decision.
Budget by blocks of responsibility
Collaborative projects are described through several overlapping views. Administrative forms show cost by beneficiary. Implementation tables show effort by work package. The actual work, however, is performed through responsibilities that often cross several work packages. Treating one partner as one work package rarely reflects reality.
A responsibility block captures the real scope owned by an organisation: the tasks it performs, the deliverables it helps create, the facilities it provides, the interfaces it must support and the decisions for which it supplies evidence. Once this block is clear, the team can estimate people and costs coherently before reconciling them into the required administrative views.
This approach also exposes hidden gaps. A partner may appear well funded at beneficiary level but lack effort in the work package where its decisive validation occurs. Another may have person-months but no visible cost for the specialised service or consumables on which its result depends.
Make person-months tell the truth
Person-months are not an abstract unit of ambition. They represent people performing specific work over time. A credible estimate connects roles, seniority, task difficulty, coordination load and the likelihood of iteration. Integration-heavy and high-risk experiments rarely succeed as a single linear pass.
Teams often underestimate the effort around the experiment: preparation, calibration, sample exchange, data cleaning, failed runs, documentation and joint interpretation. These activities may not sound as exciting as the core mechanism, but without them the evidence will not be reproducible or comparable across partners.
The opposite problem is generic inflation. Large effort numbers with no connection to task volume make the plan look unexamined. Review each block by asking what the people will actually do, when they must be available and what happens if the first route fails.
Expose the cost of credible evidence
A proposal can promise advanced characterisation, multi-site validation, field-mimicking tests, data stewardship, safety review, life-cycle assessment or stakeholder validation, yet forget to resource them. Evaluators notice when an ambitious evidence plan has no visible means of execution.
Cost items should therefore be traceable to evidence. Consumables belong to a defined experimental series. Travel supports a replication visit, shared test campaign or necessary portfolio activity. External services address a capability that the consortium cannot reasonably hold internally. Equipment is justified by additional project need, not by a general desire to upgrade a laboratory.
Specificity does not require an exhausting shopping list. It requires enough information to understand why the cost exists, why the consortium needs it and which result becomes less credible without it.
Reconcile every view before submission
The final budget appears in several places, and each must tell the same story. Partner totals, work-package totals, person-month tables, task descriptions and narrative justifications should reconcile. Small contradictions create disproportionate doubt because they suggest that the implementation plan and the financial plan were built separately.
A useful final review follows one responsibility block from beginning to end. Can the reviewer identify its tasks, timing, people, facilities, purchases and deliverables? Does the same effort appear in the beneficiary and work-package views? Is the cost proportionate to the uncertainty the partner resolves?
A winning budget is not necessarily lean, equal or simple. It is believable. It shows that resources sit where the project can fail, that support functions remain proportionate and that the consortium has translated scientific ambition into an execution model it can defend.