How to find and narrow a topic for your thesis

The hardest hurdle in a thesis isn't writing it — it's deciding what it's about. Here is a method for going from an idea to a question you can answer.

Start from genuine curiosity

A topic you don't care about gets heavier every week. Begin with questions you actually wonder about in your field — something you read, heard in a lecture, or saw in practice that didn't give you a complete answer. Write down five such questions before you choose.

Narrow relentlessly

Most topics start far too broad. 'Sustainability in Norwegian business' is not a thesis — it's an entire field. Narrow along three axes: who (which group, sector or context), what (which concrete mechanism or effect), and where/when (time and place). A narrow question done well always beats a grand one done thinly.

Tip: Test the topic as a conversation you can join. If you can't point to what others have already said about it, it's either too new or too broad.

Write a one-paragraph pitch

Once you think you have a topic, force it down to one paragraph: what you're investigating, why it matters, and how you'll go about it. If you can't explain it to a fellow student in two sentences, it isn't narrow enough yet. This paragraph also becomes the core of your proposal.

Match the topic to your constraints

A good topic respects reality: the page count, the time, the skills and the data you can actually access. Do you need interviews you won't have time to run, or register data you can't get access to? Then the topic is wrong — however exciting it is. Choose what you can finish well.

Common pitfalls

• Choosing a topic because it sounds impressive, not because you can answer it.
• Framing the topic as a statement instead of a question.
• Postponing the narrowing until after data collection — by then it's too late.

Ready to move on when…

• You can explain the topic to a friend in two sentences.
• It's narrow enough to have a real answer, yet broad enough that the answer matters.
• It fits the page count, time and data you actually have.