How to Describe Data, Impact and Ethics in an Innovation Grant Proposal

Start from one story, not three forms

Reviewers typically pass through a proposal quickly and judge it as a whole: funders assess not only the quality of the work but also how likely it is to produce the expected impact, and they expect impact goals to be visible across all sections rather than isolated in one paragraph. If you promise to open your data, explain in the data plan the conditions of reuse; if you name a result, show in the impact narrative who it changes and how.

The most common reason proposals fail is internal contradiction. A typical case: an applicant promises full open data while the ethics section concedes that part of the data is personal and cannot leave the organisation. A reviewer spots the mismatch and loses confidence in the entire application. Before writing, articulate a short storyline: what you will create, what change it causes and for whom, what risks arise, and how you will manage them.

• What data and results will the project produce, and in what form?
• What change will they cause for named groups — users, industry, regulators, society?
• What ethical and legal constraints could block the result, and how will you remove them?
• Which documents (patent, publication, partner letter, prototype) back each promise?

Describe the data: a data management plan grounded in FAIR

Data planning should start at the design stage and cover both how data will be managed during the project and how they will be shared and preserved afterwards. A good plan surfaces resources, roles and costs early and supports compliance with ethical and legal duties. The common benchmark is the FAIR framework: data should be findable, accessible, interoperable and reusable.

Be honest about what limits sharing: sensitive personal data, commercial interests, intellectual property and security requirements. Rather than hiding these restrictions, show the steps that reduce them — anonymisation, agreed embargoes, selective release of de-identified sets, or publishing code and methodology. Applicants who describe only an idealised open-data picture look less credible than those who name the boundaries and their workarounds.

Treat the data management plan as a living document reviewed at regular team meetings and attached to reports. Assign named owners, estimate storage and backup costs, and specify repositories and licences. Free planning tools with funder-specific templates, such as the DMPonline family of services, help you match the structure a particular grantor expects.

• Types, volume, format and documentation standards (metadata);
• Storage, backup and security during the project;
• Sharing and archiving strategy afterwards, with repository and licence choice;
• Restrictions on sharing and measures to reduce them;
• Roles, budget and a schedule for revising the plan.

Describe expected impact: from outputs to lasting change

Impact is the demonstrable benefit of your research and its activities, and a strong statement contains three elements: a description of the benefit with its significance and reach, a causal link between your work and the change, and evidence. Distinguish outputs (publications, prototypes, datasets), outcomes (changes in knowledge, behaviour or practice) and long-term impact. Definitions differ between funders, so follow the template and terminology of the specific scheme you are applying to.

Make the impact statement specific: not "this work will change practice", but "this method will be adopted by three named organisations, as evidenced by…". State who benefits, roughly how many people or organisations are reached, and how much the situation shifts. Reviewers are rarely experts in your exact niche, so write in plain language, keep paragraphs short and use subheadings to guide them.

For an innovation grant, show a realistic route to commercialisation or adoption, even if it spans several years. Remember that funders verify claims: there are documented cases of teams overstating their readiness stage, offering work already paid for from other sources, or inflating budgets — all detected during due diligence. An honest statement of current stage with clear milestones beats exaggerated maturity.

Write the ethics section as a risk-management self-assessment

The ethics section should read as a self-assessment: identify potential ethical issues, explain how you will address them, and list supporting documents such as approvals or permissions together with their expiry dates. In major international programmes, all proposals undergo ethics appraisal, and additional requirements may be written into the grant agreement as a result; a substantial breach of ethical principles or applicable law can lead to reduced funding or termination.

Documentation typically covers research aims and design, who participates (inclusion criteria, numbers, recruitment), vulnerable individuals or groups, data collection and analysis methods, and an assessment of benefits and risks to participants and third parties. Describe freely given, informed consent procedures, how data will be kept secure and transferred within and outside the team, and any activity outside your home country. Advisory members should declare conflicts of interest.

Where the project touches artificial intelligence, take concrete steps to assess and reduce bias: training data should reflect the diversity of the target population as closely as possible, and a formal AI impact assessment is a useful tool. For health and clinical work, consider diversity in research design and the specific approvals and regulations that may apply to biosamples, devices or interventions.

Approvals do not normally need to be in place at the point of application, but you should describe the plan to obtain them; research must not begin before the necessary approvals exist. Distinguish general good practice from mandatory rules in your jurisdiction, and when in doubt consult your local ethics committee, research office or a qualified specialist on data protection and legal matters.

• Participants and recruitment, including vulnerable groups;
• Freely given, informed consent and how it is documented;
• Benefits and risks to participants and third parties, and how they are maximised or minimised;
• Security, confidentiality and transfer of data, including sharing beyond the team;
• Non-UK or cross-border activity and applicable approvals;
• AI-related bias assessment where relevant;
• Conflicts of interest among advisers and partners.

Run a reviewer's-eye self-check before submission

Before you submit, read the proposal as an external reviewer would and ask five questions. Does it read as one story, or do data, impact and ethics contradict each other? Is it clear what results you will deliver and in what form they will be accessible? Is the expected impact demonstrably achievable, with named beneficiaries and a causal route? Are risks to people, data and reputation explained with mitigation measures? Do the promises match the current stage of the project and available evidence?

These questions are essentially the content of a good proposal. The checklist below lets your team pass the three blocks through a single consistency check before submission, so you fix weaknesses early instead of after a rejection.