Case studies
How specific applicants got into specific schools. Profiles are redacted, but the strategic decisions are real.
How a first-generation Asian male public school student got into Harvard
- Year
- Class of 2023
- School
- California public
- Background
- First-generation, Asian, male
- Stats
- 4.45 unweighted GPA, test-optional submission, 5 AP scores (English Lang 5, Art History 5, US History 4, Statistics 5, Physics 5)
- Major applied
- Economics
- Spike
- Three founder roles (Stats Club 100+ members, Hospital Music Program, e-commerce business) plus a National Literary Journal Publication
- Anchor identity
- Translator and caretaker for non-English-speaking parents and younger sibling
Outcomes
- Harvard (RD)
- Amherst, Bowdoin, Hamilton, Middlebury, Tufts (RD)
- Emory, USC (RD)
- UC Berkeley, UCLA, UCSD, UCSB, UCD, UCI (RD)
- Cal Poly SLO, Cal Poly Pomona
- CSU Long Beach, CSU Northridge
- UPenn (ED)
- Stanford (RD)
- Columbia (RD)
- Brown (RD)
- Yale (RD)
- Pomona (RD)
- Notre Dame (RD)
- Villanova (RD)
- Loyola Marymount (RD)
The diagnosis
He didn't have elite stats for a Wharton-track profile. No AP Calc BC, a five-AP-score profile without the calculus that Wharton-leaning readers reward. What he had: a translator-for-family identity that anchored every essay, three founder roles signaling sustained initiative across all four years of high school, a writing spike with a national publication, and a coherent economics-plus-writing dual-spike narrative. Harvard's holistic review valued the combination. UPenn ED, looking for Wharton-track quantitative readiness, did not.
What we brought out in his essays
Threaded one identity across every essay.
He had a powerful identity anchor sitting at the bottom of his activities list, in the Family Responsibilities slot most applicants leave blank: translator and caretaker for his non-English-speaking parents. We made this the load-bearing thread across his supplements. The translator-for-family arc appears in the personal statement, Bowdoin, Tufts, USC, Pomona, Yale, and Emory. Same person seen through seven different lenses. By the time the third reader saw the anchor refracted again, it had become who he is, not a story he tells.
Surfaced three founder roles to lead with initiative.
He founded three organizations in high school: a Stats Club (100+ members, 16 hours per week, 40 weeks), a Hospital Music Performance Program (25 student volunteers reaching 1,000+ community members), and an online e-commerce business. Founding one thing is normal. Founding three is a signal. We ordered his activities so the founder roles led. Harvard's review weights initiative heavily, and 'he started three things' reads differently than 'he has good grades and joined some clubs.'
Built writing distinction as a real second spike.
He had a National Literary Journal Publication, which is rare. Combined with AP English Language 5 and a Departmental Award in English, we positioned writing as a serious second domain alongside economics. Not 'he also writes' but 'he writes at a published level.' Two real spikes (economics + writing) are stronger than one big spike at this tier because they signal range, and Harvard specifically values that combination.
Repeated one image across multiple essays.
We anchored a cultural-bridging-through-food motif across essays at Tufts, Emory, Pomona, Yale, and Stanford. A middle school lunch exchange. A friend's Syrian dish. The same image, reframed for each prompt. Memorability at this tier is built by repetition across multiple essays in the same application file. The second AO who reads his file recognizes the image and the writer behind it.
Anchored the economics spike to lived experience.
He didn't write 'I love economics' essays. The economics interest is anchored to a specific event: during COVID, his father's video production business was disrupted, and he stepped in to help run it. That experience seeded the peer-led stats-and-economics community he then founded. Personal stake on the spike makes the spike feel inevitable, not constructed for the application.
Cast wide at the LAC and Ivy tier.
He applied to 28+ schools spanning Cal State, UCs, mid-tier privates, LACs, and Ivies. Wider than we'd recommend for most students. For a first-gen applicant without elite quant stats, casting wide at the LAC and Ivy tier maximized variance. The kind of admissions process where Harvard says yes and UPenn says no is best handled with a broad list, not a narrow ED-only bet. The downside is essay load; we managed it by anchoring multiple supplements to the same identity threads (translator, food, founder energy) so the marginal essay cost stayed manageable.
Why the denials make sense
Stanford, Columbia, Brown, Yale, Pomona, and UPenn (ED) are all single-digit acceptance rate schools where decisions at this tier are largely noise. Even Harvard-admitted students get rejected at other Ivies. The UPenn ED denial is the most informative: ED to UPenn typically signals Wharton-track even when applying to CAS, and his profile lacked AP Calc BC — the single course Wharton-leaning reviewers most reward. AP Stats and AP Physics helped but did not substitute. Notre Dame, Villanova, and Loyola Marymount are less about fit and more about how quickly a Catholic-school reader knows he wasn't headed there. School-specific positioning matters more than absolute quality at this tier.
The takeaway
At Harvard-tier admissions without elite quant stats, the bar is: a lived-experience identity anchor, sustained initiative across years, a coherent dual-spike narrative, and a writing voice earned through actual writing recognition. His academic profile would have failed at any calc-driven Wharton-style admit. Harvard's holistic review rewarded the combination. The same profile, rejected by UPenn ED and accepted by Harvard RD, shows that holistic and quant-driven schools see the same applicant very differently.
How a public school CS applicant without elite hooks got into UC Berkeley CS
- Year
- Class of 2026
- School
- California public
- Stats
- 4.85 weighted GPA, high-1500s SAT, 12+ AP scores of 5
- Major applied
- Computer Science
- Spike
- Two university research projects (healthcare ML on a large multi-institution patient dataset, renewable-energy siting abroad) + nationally-ranked chess competitor
- What she didn't have
- No USACO Platinum, no Intel/Regeneron finalist, no patents, no FRC captaincy
Outcomes
- UC Berkeley (CS, RD)
- UCLA, UCSD, UCSB, UCI, UCD, UCSC, UCM, UCR (full UC sweep, RD)
- UIUC (EA)
- University of Wisconsin-Madison (EA)
- Northeastern (EA)
- USC (EA)
- Arizona State (RA)
- Stanford (RD)
- Georgia Tech (EA II)
- Cornell (RD)
- Northwestern (RD)
- Rice (RD)
- UT Austin (EA)
- Purdue (RD)
The diagnosis
She didn't have the elite-CS-applicant signals: no USACO Platinum, no Intel/Regeneron finalist, no patents, no FRC captaincy. What she had was an academically saturated profile (every AP at 5, top-of-class), two research projects with personal stakes, and a memorable Common App anchor that wasn't about CS. We organized the application so all three signals reinforced each other.
What we brought out in her essays
The Common App essay was about chess, not CS.
Her activities list already proved she was a CS applicant. The Common App didn't need to re-prove that. We anchored it on chess instead, specifically the moment she taught a non-English-speaking family elder to play, which became the origin story for a multi-week program she ran for children. The essay positioned her as solver, competitor, creator, collaborator (character traits AOs read as engineering-ready) without telling the AO she liked CS. Most CS applicants write Common App essays about the moment they fell in love with code. Hers didn't. That's what an AO remembers a week later.
Both research projects were anchored to personal stakes.
Without personal stakes, research projects read as resume-padding. The healthcare ML project on a large multi-institution patient dataset was originally 'I did machine learning on healthcare data.' We brought out the family member's health struggle that motivated it. The project became a personal commitment, not a credential. Same move on the renewable-energy siting research with a university professor abroad: originally 'I worked with a professor.' We brought out the family visits where she saw power outages and pollution firsthand. Both projects ended up reading as evidence of who she is, not what she can do.
One pedagogical identity across multiple essays.
Her TA essay, her community tutoring essay, and her Stanford distinctive-contribution essay all converged on the same teaching philosophy: guiding students to their own aha moments rather than giving them answers. The specific anecdote (helping a student debug a program by suggesting they test edge cases, leading to a moment of self-discovery) appeared across multiple essays as the same throughline. This isn't accidental. We established aha-moments-as-identity as the consistent frame so every reader sees the same pedagogical mind from different angles.
The family + chess anchor appears in five different essays.
Cross-essay consistency builds pattern recognition for the AO. The family-teaching-chess moment appears in the Common App personal statement, the universal writing supplement, the Stanford distinctive-contribution essay, the personal experiences short essay, and indirectly in the 'five things important to you' list, where family and time with elders leads. The repetition isn't lazy. It's deliberate. The first reader sees the chess anchor and remembers it. The next reader sees the same anchor refracted through a different lens.
Activity ordering matters.
Research projects listed first (the real spike). Chess positioned as activity #7 (the personality anchor, surfaced in the Common App). Cross-country and community service as supporting texture below. This is how an AO reading at speed builds a mental model: top three activities tell you what kind of applicant, the rest fills in. Putting chess at the bottom would have made it invisible. Putting it as #1 would have made her look like a chess specialist, not a CS applicant.
Right-sized the school list.
She applied to Berkeley CS directly (no major pivot) because her profile cleared the bar at competitive public flagship CS programs. We didn't pivot her to Linguistics or Data Science as backdoor options because she didn't need to. At top-tier privates (Stanford, Cornell, Rice, Northwestern, GT), her SAT was below the CS-specific median and the lack of USACO-tier hooks meant the rejections were predictable. Accurate self-assessment of the list is part of the strategy: she applied to schools where she was competitive at CS, not schools where she was reaching at CS by an SAT gap and a research output gap.
Why the denials make sense
Stanford, Georgia Tech, Cornell, Northwestern, and Rice are top-tier private CS programs where her SAT was below the CS-specific median and the elite-CS-hook signals she didn't have are common. Even strong profiles get rejected at these schools without those hooks. UT Austin and Purdue CS have become nearly as competitive as top privates for out-of-state and non-Texas applicants, respectively, and cleared her at RD. Accurate self-assessment of the list is part of the strategy: she applied to schools where she was competitive at CS, not schools where she was reaching at CS by an SAT gap and a research output gap.
The takeaway
For top public CS programs without a USACO-tier hook, the bar is: extraordinary academic floor, multi-domain research, and a memorable non-CS Common App. The first two without the third reads as well-rounded but generic. The third without the first two doesn't have enough substance. All three together is what worked here.
30-minute call. No obligation. Trusted by 100+ Bay Area families.