Humboldt Fellowship Research Proposal 2027: Five Pages, Your Idea, Not Your Thesis
The research outline is capped at five pages including the bibliography. Reviewers judge it on originality, method, focus and feasibility at your chosen host. Here is a structure that follows the foundation’s own questions.
ScholarshipUnion Editorial
3 min read
Updated Oct 6, 2026
Your research outline is the one document in the file where you speak directly about the future. Everything else — the CV, the publications, the reviews — is about your past. The foundation’s guidance for hosts sets the brief: roughly five pages in total, including the bibliography, developed by you and coordinated with your host.
The three questions the outline must answer
- 1Max. 1 pageWhat is the current state of research?
Evidence it with about five relevant publications from the field. Show the gap, not a literature review.
- 2CoreWhat questions will you examine, and why do they matter?
State the questions precisely and explain their significance for advancing the research area.
- 3CoreWhat methods will you use to reach the objectives?
Name them, justify them with references where useful, and tie them to the host’s facilities.
What reviewers score
The selection criteria on the programme page say the proposed research is judged on its “originality and innovative potential”, broken down into:
- Significance for the further development of the subject area
- A convincing selection of scientific methods
- Potential for your scientific development
- Clear focus and feasibility within the requested sponsorship period
- Viability at the host institute
Note what is missing: impact statements, societal outreach plans, budgets. This is not a Horizon Europe proposal. Spend the pages on the science.
A practical five-page structure
| Section | Length | What goes in |
|---|---|---|
| Title and summary | ¼ page | One paragraph a non-specialist committee member can repeat. |
| State of research | ≤ 1 page | About five key references; end with the open question. |
| Research questions and aims | ¾ page | Two to four precise questions; why answering them matters for the field. |
| Methods and work plan | 1½ pages | Approach per question; data, material or equipment; a month-by-month plan matching the months you request. |
| Why this host, why Germany | ½ page | What the host group, archive or facility adds that you cannot get at home. |
| Your development | ¼ page | How the project moves you beyond your doctorate or current line of work. |
| Bibliography | ¾ page | Counts toward the five pages. |
The section split is our suggestion, built around the foundation’s three required questions and its published criteria. It is not an official template.
The originality trap
The foundation lists “lack of originality” and “lack of academic development” among the most frequent reasons for rejection. In its words, the project “should not be a direct continuation of the doctoral thesis or of the academic work carried out to date.” If your outline reads like chapter six of your dissertation, reviewers will notice.
“I will extend my PhD model of X to two more datasets.” Routine, predictable, and the host adds nothing essential.
“My PhD showed X under condition A. Whether it holds under B is unknown because B needs method M, which the host group developed. I will test it.”
Independence: write it yourself
The host is asked directly: “did the applicant prepare the submitted research project independently? To what extent did you or others contribute?” Coordinate with your host on feasibility and access, but draft the idea yourself. A proposal that obviously comes from the host’s grant pipeline weakens both statements.
Fit the plan to the months you request
Postdocs can request 6 to 24 months; experienced researchers 6 to 18. Requested time can be split into up to three stays of at least three months each, used within three years. Reviewers check that the plan is feasible in the period you ask for, so a 24-month request needs 24 months of real work, and a 9-month request should not promise a five-year programme. If part of the work must happen abroad (field work, archives), plan it in the middle: stays outside Germany are capped at usually 25% of the funding period and need at least six months at the host.
Generative AI
The foundation publishes guidance on dealing with generative AI in selection and requires applicants to follow the rules of good scientific practice. Treat any AI-assisted drafting as you would co-authorship: the ideas and the responsibility must be yours.