What the question asks, and what the Criteria page tells you
Two tasks in 200 words: say why this specific course at Cambridge , and say how it facilitates your future plans . Both halves are marked. A beautiful account of the course with no plan fails, and so does a clear plan that could be pursued anywhere.
What the Criteria page tells you
- Make a strong case for a particular postgraduate degree at Cambridge.
- Show the academic background and skills to complete it.
- PhD and most research MPhil applicants contact a potential supervisor and summarise a research project.
- Taught MPhil and other applicants show how the course advances their career plans.
- Gates wants an "academically transformative experience," so you must research and justify the course in full.
The test is whether a reader finishes thinking: this person has found the exact place where this question can be answered.
The capability gap, and the swap test
The strongest Question 2 answers are built on one sentence: "I cannot yet ___." Everything else follows from it. The course exists to close that gap.
| Weak gap | Strong gap |
|---|---|
| "I want to learn more about development economics." | "I can describe where social schemes fail, but I cannot evaluate whether a fix works before a state spends money on it." |
| "I want to deepen my knowledge of genetics." | "I can build a field test, but I cannot yet predict which virus strains will beat it." |
| "I want a global perspective." | "My legal argument works in one jurisdiction, and I need the comparative case law that shows where it has already succeeded." |
If you cannot finish the sentence in one line, you are not ready to draft.
The swap test
Replace "Cambridge" with another university. If most of your answer still works, it is generic and it will not score.
| Generic, survives the swap | Specific, does not |
|---|---|
| "Cambridge is a world-leading university with outstanding faculty." | "Dr X's group has spent a decade on the exact failure mode I study, and her hybrid-wall results are the closest evidence to my houses." |
| "The interdisciplinary environment will broaden my perspective." | "The course pairs a causal inference paper with a policy practicum, which programmes in my region teach separately." |
| "The reputation of the degree will open doors." | "The dissertation lets me test one design on administrative data I already hold." |
| "I will learn from top academics and peers." | "My cohort will include officials who have run schemes under different legal systems, and portability is an inter-jurisdiction problem." |
Avoid rankings, prestige and praise of the University. The panel already knows. Gates is judging whether you know why you need it.
Researching specifics
| Source | What to extract |
|---|---|
| The course entry in the Course Directory | The structure, the funding deadline, any prerequisites |
| The department's course page | Papers or modules, assessment, dissertation or thesis rules |
| The staff list and recent publications | Two or three people, and what each recently found |
| Research centres and facilities | What exists there that does not exist where you are |
| Your own supervisor emails (PhD) | Their advice on the project and the department |
Before you write, collect: one supervisor or group with a recent result, one structural feature of the course, one resource, and one reason each links to your gap. Every name costs characters, so choose the three that do the most work.
Structure for a PhD or research degree, and for a taught degree
| Move | Words (of 200) | What it does |
|---|---|---|
| 1. The question and your gap | About 40 | What you want to answer and the specific thing you cannot yet do |
| 2. The place | About 70 | A named supervisor or group, what their work gives you, plus one resource |
| 3. Readiness | About 30 | Why you can begin in month one: skills, prior results, a pilot |
| 4. The plan | About 50 | The first role afterwards, then a dated, measurable aim tied to improving lives |
| 5. A closing contribution | About 10 | One thing you give back, if there is room |
Structure for a taught degree
| Move | Words (of 200) | What it does |
|---|---|---|
| 1. The career gap | About 45 | What you do now, where it stops, what you cannot yet do |
| 2. The course design | About 70 | Two or three specific features, each linked to part of the gap |
| 3. Readiness | About 30 | The background that makes this a step, not a leap |
| 4. The plan | About 55 | The role you return to, a dated aim, the measurable change |
The future-plans clause, and special cases
The second half of the question needs more than "I will make a difference." Use three levels:
| Level | Example |
|---|---|
| The first role after Cambridge | Lead a technical unit in the housing ministry |
| The output you will produce | An open retrofit standard municipalities can adopt |
| The dated, measurable aim | 5,000 homes covered by 2034 |
The plan must match Questions 3 and 4. A reader who sees three different futures in three answers will trust none of them.
Special cases
| Case | What to do |
|---|---|
| A second master's | One clause on what your first degree did not give you, and why this one does |
| A switch of field | Name the bridge between your old and new field, and show you have the prerequisites |
| No supervisor reply yet | Say what you know of their work. Gates does not require a confirmed supervisor, but department research is expected |
| Professional degree with a research element | Focus on the part that closes your gap, usually the dissertation |
| Applying for several courses | Make each answer a real case. You will be interviewed once, for one course |
The writing process
- Finish the sentence "I cannot yet ___."
- Gather your three specifics (section 3).
- Write one line per move, then check it against the structure.
- Draft to about 250 words.
- Cut every sentence that could appear in another applicant's answer.
- Check every name against the department's website.
- Count characters against the 1,400 cap.
Three worked examples (all invented)
Example A: PhD in Engineering (about 185 words)
Informal housing on the hillsides of Valparaíso is built without engineering input and will be tested by the next major earthquake. I have designed low-cost retrofits that held in two pilot houses, but I cannot yet predict how mixed masonry and timber walls fail, which is what I need to prove a retrofit is safe at scale.
That is the question behind a PhD in Engineering at Cambridge. Dr Esther Marlowe- Quint's group has spent a decade testing how unreinforced masonry behaves under cyclic loading, and her recent work on hybrid walls is the closest existing evidence to the houses I study. The group's testing programme would let me build and break full-size wall panels, and the department's work with seismic risk researchers would tell me where to retrofit first.
My two years designing retrofits and a Master's thesis on timber-masonry joints give me the modelling and fieldwork to begin in month one. Afterwards I will lead a technical unit within Chile's national housing ministry, publishing an open retrofit standard that municipalities can adopt without a consultant. By 2034 I expect it to cover 5,000 homes.
Why it works. The first paragraph states a gap, not a love of the subject. The supervisor is named with what her work is and why it matters to this problem. "Build and break full-size wall panels" shows the applicant knows what the work physically involves. The plan has a role, an output, a year and a number, so it can be checked in 2034.
Example B: taught MPhil in Public Policy (about 200 words)
For six years I have designed state social protection schemes in India, and I know where they fail: a migrant worker loses her ration entitlement when she crosses a state border. What I cannot yet do is evaluate whether a fix works before a state spends its budget on it, and I lack the quantitative training to defend that evidence to a finance department.
The MPhil in Public Policy is built for this gap. Its policy analysis course and quantitative methods paper would give me causal inference methods I have so far used only through other people's reports. The dissertation would let me test one portability design on the state administrative data I already hold. And a cohort drawn from governments in several regions matters because portability is an inter-state problem, and I need peers who have run schemes under different legal systems.
I come with six years of administrative experience and a first-class econometrics course. On return I will become evaluation lead in my state's social justice department, and within three years I aim to have scaled one tested portability rule to all 11 districts that send workers out of state.
Why it works. It opens with a professional's problem and a clearly scoped inability. Each course feature is attached to a specific part of the gap. The readiness paragraph answers the Criteria page's request to show the background and skills to complete the course.
Example C: LLM in International Law (about 190 words)
I represent communities displaced by dam construction in the Brazilian Amazon, and I keep losing the same argument: courts accept that people were harmed but not that international law obliges the state to consult them before construction begins. What I cannot yet do is build that argument from the case law of other regions and the doctrine of free, prior and informed consent, where it is strongest.
The LLM in International Law is the course that teaches this. Its option on indigenous peoples and international law covers consent doctrine through comparative case law, the one place the argument has been won. The research paper would let me write the comparative analysis I need as a brief for my clients. And a cohort of practising lawyers from several legal systems would test my reading against people who have argued it elsewhere.
I come with seven years of litigation, including two appeals before the federal court, and a published article on consultation duties. On return I will lead the strategic litigation unit at my legal-aid organisation, and by 2031 I aim to have filed consent-based claims for 12 affected communities.
Why it works. The gap is a legal skill, not a topic. Each course feature (option, paper, cohort) answers part of it. Experience is stated as readiness, in numbers. The plan is a role, an output and a count. Check that any option or paper you name actually exists in the current course handbook.
Common failures, and the test suite
- Prestige instead of reasons: rankings, history, "the best in the world."
- A list of modules with no link to a gap.
- A supervisor named as decoration, with no reason attached.
- Repeating Question 1 or the research proposal.
- A vague future: no role, no year, no number.
- A second master's with no explanation.
- College or student-life reasons, which belong nowhere in this answer.
- Naming facilities or papers that are out of date.
The test suite
- The swap test. Change the university name. Does it still work?
- The gap test. Can you finish "I cannot yet ___" in one line, and does each feature answer part of it?
- The name test. Is every name attached to a reason?
- The 2034 test. Could someone check your plan in ten years?
- The thread test. Does the future match Questions 3 and 4?
- The five-minute test. Could you talk about every feature you named for five minutes in the interview?
Checklist
My gap is a specific thing I cannot yet do
The answer fails the swap test if I change the university's name
Every name is real, current and linked to a reason
I show I can begin the work now
My plan has a role, a year and a measurable outcome
It does not repeat Question 1 or my research proposal
It stays under 1,400 characters including spaces