Which Ivy League universities appear in your published cases?

Four of the eight. The twelve detailed case records on this site include Brown, Cornell twice, the University of Pennsylvania and Harvard, alongside Caltech, Duke, Carnegie Mellon, Johns Hopkins, Emory, Tufts, Rice, UCLA and Oxford. MIT and the University of Chicago appear in the shorter admission highlights. We publish what each file contained rather than a count of offers, because the contents of the file are the only part a family can still change.

Is there a formula for a highly selective application?

No, and our own records are the evidence. Across twelve files the subjects run from physics to gender studies, the academic records sit on five different scales, two students submitted no SAT or ACT at all, and the activities have almost nothing in common at the level of what was actually done. What repeats is structural, not topical: in every file there is one question the student owns, evidence that they pursued it themselves, and a reason for each university that could not be swapped for another name. That is what this guide sets out.

Do you plan around Ivy League acceptance rates?

No. A published rate is an output of whoever happened to apply that year; it is not comparable across universities or across years, and no round or ranking compensates for a thin file. It also tells a family nothing they can act on. What we plan around is the five things listed in this guide, each of which is something a student can actually build between Grade 10 and the autumn of Grade 12.

How early does this work have to start?

Our roadmap runs across three years because the five things cannot be assembled in one autumn. Grade 10 is for becoming a problem finder and forming a first version of a central question; Grade 11 turns that question into evidence through research, a core project or an advanced competition, and begins the essay and school-list work; Grade 12 is presentation and precision. One of our case records states the start date explicitly: service began in the spring semester of Grade 10.

What if the grades are strong but the activities look like everyone else's?

That is the first of the three dilemmas our method is built to answer, and the second cornerstone exists for it. The work is to reorganise the record around the real change the student brought about rather than to add more entries to it. One of our published records describes exactly this starting point, conventional humanities activities lacking distinctiveness, and what followed was a single year-long independent study of cultural-heritage policy in Beijing's urban renewal, reviewed by experts, which gave the whole file a direction it had not had.

Ivy League admissions: what a highly selective application has to establish

Ivy League admissions — and admissions at the other universities read in the same highly selective pool — turn on something a transcript cannot carry: whether the file adds up to a person with a question of their own. The Ivy League is eight universities (Brown, Columbia, Cornell, Dartmouth, Harvard, Pennsylvania, Princeton and Yale), four of which appear in the twelve detailed admission cases published on this site, alongside Caltech, Duke, Carnegie Mellon, Johns Hopkins, Emory, Tufts, Rice, UCLA and Oxford. Read across those records, the same five things are established every time: an academic identity rather than a set of scores, a thread the activity record actually has, evidence the student produced rather than attended, a fit argument that cannot be swapped for another school, and one recognisable person across every document. This guide defines the terms, sets out those five, shows what the spine of each published file was, and gives the order the work is done in.

Key definitions

The Ivy League. Eight universities in the north-eastern United States: Brown, Columbia, Cornell, Dartmouth, Harvard, Pennsylvania, Princeton and Yale. It began as an athletic conference and became a shorthand for selectivity. It is not a ranking, and it is not the boundary of the highly selective pool — Caltech, MIT, Duke, Chicago and others are read the same way.
The highly selective pool. The group of applicants at a university where nearly everyone is already academically qualified. Once that is true, the academic record stops separating people and everything else starts to.
Holistic review. Reading the whole file — transcript and school profile, letters, activity list, essays, any interview — as one document set, rather than ranking applicants by a number. It is why the parts of an application have to agree with each other.
Core narrative. In the Visible Growth system at Mellin International Education, the storyline and values distilled from a student's experiences through in-depth conversation. It settles who the student is and how they want to be seen, and it is what gives each document its purpose.
Central question. The one line of inquiry a student owns, distilled from a crowded record rather than invented for the application. In our three-year roadmap it has a first version in Grade 10, becomes evidence in Grade 11, and is presented precisely in Grade 12.
Strategic fit. The three-way analysis of student traits, school culture and the competitive field that positions each application, so that every Why School essay locks onto the core narrative and cannot be handed to a different university.

One thing we deliberately leave out: acceptance rates. They are an output of whoever applied in a given year, they are not comparable between universities or between years, and they give a family nothing to do on a Tuesday evening in Grade 11. Everything below is something a student can build.

The five things a selective file has to establish

#	What the file has to establish	Which cornerstone of our method carries it	How it fails
1	An academic identity, not a set of scores	Academic depth, built and presented	Strong grades with no subject the student is visibly about
2	A thread the activity record actually has	Leadership, reshaped and narrated	Ten impressive entries that could belong to ten different people
3	Evidence the student produced, not attended	Our research, programme and personal-project work	A list of programmes joined, with nothing to show from any of them
4	A fit argument that cannot be swapped	Strategic fit, precisely positioned	A Why School essay that still works with the school's name changed
5	One recognisable person across every document	Whole-person growth, recorded and guided	Essays, activities and letters each describing a slightly different student

Read down the last column and the pattern is clear: every failure is a failure of coherence rather than of effort. The files that do not work are rarely thin. They are usually crowded and unaligned — which is why our preface puts it the way it does: now that software can polish expression and tidy up experience, the surface craft of an application no longer separates anyone. What is left is what only time and sincerity build, and the three failures our method is named for are exactly the ones a last-minute effort cannot fix.

A note on the third row, because it is the one families most often misread. Depth is not the same as access. Our research and programme work connects students to laboratories and researchers at Harvard, MIT, Stanford, Columbia and other leading institutions — 55+ frontier opportunities across STEM, the humanities and business — but the point of a placement is the recommendation letter or academic report that comes out of it, which is evidence of a core academic interest. The same logic rules out competing for its own sake: selective programmes, competitions and journals are chosen because they strengthen a particular narrative, and a personal project is incubated from idea to delivered impact. Choosing those experiences well is a separate piece of work with its own guide.

What the spine of each published file was

Below is what the twelve detailed records on this site say held each application together. Nothing here explains why any student was admitted; it is what the files contain.

Case	Admission as recorded	The spine of the file
Caltech and Oxford, physics	Caltech (REA) and Oxford	Physics and philosophy pursued together; a main essay about keeping an aquarium through repeated failure, and ten supplements reading physical phenomena from one point of view
Cornell, information science	Cornell (ED)	Research on AI and privacy, FRC robotics captaincy and varsity badminton, with a main essay about one match going from leading to losing and what failure finally meant
UPenn, arts and sciences	University of Pennsylvania	A quantitative, humanist physics researcher guided by light: mainland China's first campus light-pollution monitoring station, built with partners and published
Cornell, astronomy	Cornell (ED); Cambridge interview invitation	Exoplanet and black-hole research alongside public science teaching — an online astronomy course built for 89 children in Gansu after being told the gap in prior knowledge made it impossible
Duke and Rice, biology	Duke and Rice (RD)	A hands-on problem solver: first-place grand award at Regeneron ISEF in biomedical and health sciences, and a main essay that opened with repairing the mouse his mother had used for twenty years
Duke, CMU and UCLA, computer science	Duke, Carnegie Mellon and UCLA	Three fields fused on purpose: computer science, East Asian studies and philosophy, carried by a quantum-education organisation across 15 partner schools and a heritage-tea ambassadorship
Emory, neuroscience	Emory (ED)	A recorded chain from personal passion to scientific inquiry to community application to future vision, around the dual identity of music-cognition explorer and community builder
Brown, gender studies and history	Brown (RD)	A future lawyer who speaks for the vulnerable: a restlessness taken from Jane Eyre as the essay's thread, an anti-bullying policy advanced as Head Senator, and feminist legal theory in research
Johns Hopkins, East Asian studies	Johns Hopkins	One cultural tradition entered from every available direction — a shadow-puppetry heritage alliance, first authorship of two papers, performance, digital design and teaching
Harvard, mathematics	Harvard	Competition mathematics at the top end next to a published collection of 80 prose poems: the same mind in two registers rather than a mathematician with a hobby
Carnegie Mellon, mechanical engineering	Carnegie Mellon, College of Engineering (RD)	Drone design from sketch to IEEE paper, with an 800-member aviation club across four cities — a top-30% record and no SAT or ACT at all
Tufts, biotechnology	Tufts (ED)	Biomedicine plus an argued fit: a Why Tufts essay built from values, practice and research interest, and a main essay about being helped by children with disabilities rather than helping them

Four readings follow, and all four are observations about what the records contain.

No two spines are alike, and none of them is a score. Not one of the twelve is held together by an academic result. Where the record is strongest the subject is unusual; where the subject is ordinary the angle is not. The academic file is the entry ticket to the pool, and the spine is what happens after everyone in the room has one.
The evidence is nearly always something that now exists. A monitoring station, a paper, an alliance, a curriculum, an algorithm, a club with members, a poetry collection. Eight of the twelve records name an artefact that outlived the application. That is what row 3 of the table above means in practice.
The main essay is usually small. An aquarium, a badminton point, a broken mouse, a line from a novel. None of the recorded main essays is a summary of the applicant's achievements, which is the single most common thing a strong student wants to write. Our essay guide deals with why.
The fit argument is written school by school. The Emory profile matched that university's emphasis on interdisciplinary research, community service and collective well-being; the Why Tufts essay argued from three angles, each tied to one goal; the Cornell supplements answered that university's own prompt about any person and any study. Ten school-specific answers is a separate discipline, covered in our supplemental essays guide.

Where the three common dilemmas get answered

Our method names three problems before it names any solution, because they are the three families arrive with.

The dilemma	What it sounds like	Which cornerstone answers it, and how
Activities that look like everyone else's	Similar backgrounds everywhere — how does this child stand out?	Leadership, reshaped and narrated: a narrative-leadership framework reorganises and integrates the activity record around the real change the student brought about
Essays without a spark	Excellent grades, yet the personal story falls flat	Academic depth, built and presented: research, deep reading and critical-writing training give the essays something specific to be about
School-list anxiety	Beyond the rankings, which schools truly fit, and which are worth a reach?	Strategic fit, precisely positioned: student traits, school culture and the competitive field analysed together, so every application has a reason

The fourth cornerstone sits under all three. A growth portfolio kept across the three years records the breakthroughs and the reflections as they happen, which is what makes a file about real growth rather than about a reconstruction performed in Grade 12. It is also the only defence against the fifth failure in the first table: a record written down as it happens cannot drift into describing a different student.

Two mentors do most of the work that joins these pieces, and the division matters when a family is deciding who they are hiring. A Chinese planning mentor follows the student across the whole of high school, keeps the timeline on track, handles school communication and guarantees the accuracy of every document. An international essay mentor guides ideation and drafting and advises on the list. Behind them sit young alumni mentors and academic mentors from Harvard, MIT, Stanford and elsewhere. Several people on our mentor team have read files from the other side of the desk — one worked in the University of Chicago admissions office reviewing applications and making decisions, one is a former Whitman admissions officer and Princeton programme mentor, and one spent four years as a Harvard alumni interviewer — and our senior planning mentor works backwards from the admitted-student profiles of target universities rather than forwards from a template.

Six steps, in order

Find the question before building anything. In Grade 10 the task is to stop completing tasks: notice and record what puzzles or frustrates you in coursework and activities, and write a first version of a question that is yours. Everything in steps 2 to 6 is paid for by this one.
Make the academic record support that question. Take the strongest available version of the two or three subjects the direction depends on, keep one quantitative subject whatever the field, and let the subject set read as a sentence rather than a list.
Produce one thing a year that did not exist before. A paper, a project, a programme with participants, a dataset, a curriculum. Choose the research placement, competition or journal because it strengthens this particular narrative, and then finish the output — a placement with nothing written up proves attendance.
Reorganise the activity record around change, not around prestige. Ten slots, one thread. Rank by what each entry proves about the student, say what changed because they were there, and cut the entries that only prove they were busy.
Build the list from fit, then choose the round. Three-way analysis first: traits, culture, competitive field. Our guides on building the list and on ED, EA and RD take this in detail; the rule of thumb is that if the fit paragraph survives having the university's name swapped out, that university does not belong at the top of the list.
Make every document say the same thing. In Grade 12 the essays, the activity list, the letters and the interview all have to convey the same core traits. Read them together in one sitting, out loud, and fix whatever contradicts.

What we recommend

Treat the five requirements in the first table as the specification and everything else as implementation. Settle the central question in Grade 10 rather than in Grade 12; make the transcript and the subject set serve it; produce evidence rather than attendance; reorganise the activity record around real change; build the list from fit and let the round follow; then spend the autumn making every document describe one person. Break all of that into stage goals and weekly tasks so the three years stay solid and controllable instead of becoming a sprint in October — that is what our milestone management exists to do, and it is the difference between a plan and an intention. We do not plan around acceptance rates and we would be suspicious of anyone who offers to. If you want a student's current file read against these five requirements, tell us the grade, curriculum and interests and a consultant will reply within one working day.