Imperial MSc Finance Supplemental Essay Example: Avoiding Vague Answers (Score 93)

During my third year at the Universidad Autónoma Benito Juárez de Oaxaca, my four-person research group was tasked with producing a policy memo on algorithmic content moderation affecting small-scale creators in southern Mexico. Three weeks before submission, our lead analyst discovered that the engagement-rate dataset we had used to benchmark platform behaviour contained a systematic collection bias: rural creators using low-bandwidth connections were underrepresented by roughly 40 percent, skewing every recommendation we had drafted. Disclosing the flaw meant restarting the quantitative section entirely; staying silent would have produced a polished document built on a misleading foundation. The tension was real — our group had already invested forty hours, and the deadline was fixed. I called an emergency session and argued that we had to rebuild the dataset from scratch using creator-submitted analytics exported directly from the platforms, supplemented by structured interviews with eight independent creators in the Oaxacan Valles Centrales region. To manage the time constraint, I designed sensitivity tables showing how our policy conclusions shifted under three audience-reach scenarios — low, mid, and high rural penetration — so that even with incomplete replacement data, reviewers could see the directional robustness of our recommendations. I also restructured our scenario framing to acknowledge the data limitation explicitly in the memo's methodology section rather than burying it in an appendix, which required persuading two teammates who feared it would undermine our credibility. The revised memo received the highest mark in our cohort. More concretely, our faculty supervisor shared it with a digital rights NGO in Mexico City, Artículo 19's technology programme, which cited our rural-creator framing in a subsequent submission to the Instituto Federal de Telecomunicaciones. For our team, the episode demonstrated that Integrity — refusing to advance a flawed argument — and Collaboration — redistributing the analytical load transparently rather than quietly correcting the error alone — produced a stronger outcome than either value would have in isolation. Innovation entered through the sensitivity-table design, which converted a methodological weakness into a visible, honest feature of the analysis. The procedural lesson I carry into cohort work is straightforward: disclose data limitations early and convert them into structured scenarios rather than hoping they go unnoticed. In a programme like Imperial's MSc Finance, where quantitative rigour underpins every group deliverable, the instinct to surface a flaw immediately — and to reframe it as a bounded uncertainty rather than a fatal error — is more valuable than any single technical skill. I intend to bring that norm into case competitions and group projects at Imperial: when a model assumption looks shaky, name it in the room, run the alternative, and let the analysis speak honestly.