What a political science student learned at McKinsey

I wasn't the most obvious candidate for a consulting internship. Political science, no finance background, no case competition track record. What I had was two years of running an advocacy organization — which meant I'd spent a lot of time doing the core thing consulting actually requires: figuring out what someone needs to decide, then building the shortest path to that decision.

The McKinsey internship ran from May 2025 through June 2026, part-time alongside my final year at NTU. It was the first — and so far only — internship I've done. And it was a significant recalibration of how I think about analytical work.

The automation project

One of my first real contributions was a problem nobody had asked me to solve. The team regularly needed to process historical data from Taiwan's top-2000 companies ranking — a dataset spanning 2007 to 2024. The existing workflow took roughly four hours: two hours of manual work, two hours of computation. Not terrible, but the kind of friction that quietly costs teams across a week.

I spent an afternoon learning enough about web scraping to build a tool that reduced the process to two clicks and fifteen minutes of automated runtime. The underlying insight wasn't technical — it was organizational. The bottleneck wasn't the computer's speed. It was the manual steps that nobody had bothered to eliminate because they'd always been done that way.

That project taught me something I carry everywhere now: the first question isn't how do we do this faster. It's does this step need to exist at all.

Supply chain analysis

I received more than 50 raw, unstructured spreadsheets from a consumer electronics client dealing with rising component costs. My job was to turn that into something a non-technical executive could use to make decisions in under five minutes.

I built a dashboard that surfaced the key cost drivers, modeled different procurement scenarios, and highlighted where the largest savings opportunity sat. The client could see the answer in ten seconds. The analysis behind it took weeks. That gap — between the complexity of the analysis and the simplicity of the output — is exactly where consulting creates value. The output identified over a million USD in potential annual savings.

Geopolitics and the political science advantage

The work I found most natural was the kind that blended political analysis with commercial implications — which turned out to be more common than I expected at a business consulting firm.

One project involved assessing supply chain risk for a client exposed to Taiwan Strait tensions. I was working with cross-strait trade data and political risk indicators, running the kind of analysis I'd done in academic settings — except now it was feeding directly into a client's capital allocation decision. Having a political science background wasn't a gap to compensate for. It was a direct input.

The methodology that stayed

McKinsey has a phrase: "begin with the end in mind." It means before you touch the data, define the decision. Who needs to decide what, and by when? Then work backward to determine what evidence would change that decision, and what analysis generates that evidence.

I'd been doing a version of this at EdYouth — every advocacy campaign started with who needs to believe what for policy to change — but at McKinsey it became systematic. Hypothesis-driven: state your assumption first, then run the fastest test that could falsify it. Communication-first: every piece of analysis should compress into a decision a non-expert can act on in thirty seconds.

Those aren't consulting skills. They're thinking skills. The consulting environment just accelerated how quickly I internalized them.