Performance marketing for AI-era growth – backed by models, not gut feeling.

Who is Andrey

Andrey is 17. He loves mathematics – not as a tool, but as a subject in its own right. He reads about proofs the way other kids read about football. Alongside that, he has spent the past two years genuinely obsessed with longevity research: why we age, whether it can be slowed, and what the science actually says versus what the wellness industry claims. That's what drew him into biology and chemistry – not a general interest in science, but a specific question about the human body and its limits.

The Trigger

Next academic year Andrey starts an A-level programme at a Cambridge-system school. The subject choice had to be made now – and the range was wider than expected. The classical STEM set (Mathematics, Further Mathematics, Physics, Chemistry, Biology) sat alongside options like Business, Computer Science, and others. The Cambridge system caps the choice at 4 subjects, which means every slot carries weight and every inclusion crowds something else out.

One constraint resolves early: Andrey is already significantly ahead of the standard Cambridge Mathematics programme. Further Mathematics is not a burden for him – it is natural territory. The question is not whether he can handle FM, but whether FM is the right use of one of his four slots given everything else the combination needs to achieve.

As a parent, I was also watching what was happening in tech. AI is everywhere. Computer Science specialities are booming. That pulled me in a different direction from Andrey's instincts – and created the real tension.

The Actual Problem

This is a multi-criteria optimisation problem with four constraints that don't naturally resolve together:

University accessPick subjects that keep doors open to decent universities.
Future-proof programmeLand in a field that is still "in demand" in 7–8 years, when Andrey graduates (around 2032–2033).
FlexibilityAvoid locking out alternative programmes if his interests shift in the next 3–4 years (plan B optionality).
CostUniversity fees in the tens of thousands per year is something I would actively prefer to avoid if there's a legitimate alternative.

The question: what is the optimal subject combination, and why is it optimal – not just for getting into a good university, but for the right programme, in the right country, at the right cost, given who Andrey actually is?

Why This Required Research

Every "obvious" answer fell apart under scrutiny:

The default advice wasn't wrong because it was bad advice. It was wrong because it was generic – answering a different question than the one actually on the table.

Problem Statement

Given a student with a strong affinity for mathematics as a discipline and a specific interest in longevity science, who is about to choose A-level subjects in a Cambridge-system school – find the subject combination that maximises expected outcome across four dimensions simultaneously:

The solution must be grounded in actual admissions data across multiple countries, not in received wisdom about "best universities" or "safe STEM subjects."

The Research Scope

To answer the question properly, we needed to map the actual landscape:

The research was conducted in June 2026. Live web verification where possible; training knowledge for JS-rendered sites, labelled accordingly.

What This Research Signals Professionally

This research is not about education. It is about how a certain kind of thinker approaches a complex personal decision – and what that reveals about their professional method.

Most people in this situation Google "best universities for maths" and follow the result. This research instead: defined the decision criteria explicitly, built a methodology, mapped 100 institutions across 19 countries, produced 5 datasets, challenged every obvious answer against actual data, and documented where the received wisdom failed and why.

That is not parenting. That is how a senior strategist scopes a problem.

The four-dimension optimisation framework used here – programme fit, temporal robustness, optionality, cost efficiency – is the same mental architecture applied to growth strategy, channel planning, and media mix decisions. The domain is different. The method is identical.

Also in this series
How I Built It Why ARWU over QS and THE, why 100 universities, and what 1,200+ normalised pairs made possible.
What the World Actually Offers CS at 95 universities. AI at 12. France above Imperial. What the coverage matrix actually showed.
The Unfashionable Four Four A-level subjects. Stable across 31 programmes, 100 universities, 19 countries.
The Floor Keeps Rising Seven scarcity cycles since the 1880s. The compression curve that makes a four-year degree longer than a full cycle.