How to Write UPSC Mains Answers That Actually Score Well

I must have read 200+ answer openings by now. And honestly? Most of them are terrible — including ones that scored fine.

The difference between a 120 and a 140 often comes down to the first sentence.

Here's what I mean. A question like "Examine the impact of urbanization on Indian cities" — most people start with:

Urbanization refers to the process of people moving from rural to urban areas.

That's not an answer. That's a dictionary entry. The examiner knows what urbanization means. They asked about IMPACT.

Now look at how Garima Lohia (AIR 2) starts similar questions — she opens with a claim. Something like:

Urbanization in India has created cities that drive 63% of the economy but are governed by institutions designed for a 1950s population.

She doesn't define anything. She states something specific and a bit provocative. The examiner reads that and thinks — okay, this person has something to say.

I noticed this pattern across almost every topper who scored 140+. They don't open with context. They open with an argument.

Ayan Jain (AIR 16) does this too — his GS3 answers start with a number or a tension. Something like "India's fiscal federalism is stretched between concurrent liabilities and state-specific revenue constraints." He's not explaining what fiscal federalism is. He's telling you his take on it.

The thing is — this is terrifying to do in an exam. What if your claim is wrong? What if the examiner disagrees? I think that fear is exactly why most people don't do it. But the toppers I studied seem comfortable with that risk. And it pays off.

This is where I see the biggest gap between average and top-scoring copies.

Almost everyone knows they should use data. The difference is how they use it.

A typical answer might say: "Urbanization has increased rapidly in India." That's a statement with no weight behind it.

A topper answer says: "34% of Indians live in cities, producing 63% of GDP — yet urban local bodies account for just 1% of GDP."

See the difference? The first is a claim. The second is evidence that makes the claim undeniable.

I noticed something interesting though — toppers don't use that many data points. Maybe 2-3 per answer. But the ones they use are carefully chosen. They're not dumping statistics. They're picking the one number that makes their point stick.

Uma Harathi (AIR 3) is particularly good at this in GS2 answers. She'll reference one specific Supreme Court judgment or one parliamentary committee report per topic. Not five. Just one, used well.

Ishita Kishore (AIR 1) weaves data into her narrative naturally — it never feels like she's forcing a statistic because she prepared it. It feels like she genuinely knows the material.

The practical takeaway: instead of memorizing 20 statistics per subject, pick 3-4 that you can use across multiple questions. The ones that surprise people or challenge assumptions work best.

Everyone says "presentation matters." But what does that actually mean?

I've gone through enough copies now to notice specific things:

Underlining. Ishita Kishore underlines the key term in almost every sentence — not whole sentences, just the word. It's subtle but it works. Your eye goes exactly where she wants it to go. Garima Lohia does the same thing, though she underlines a bit more aggressively.

Paragraph length. This is the one nobody talks about but it's everywhere in top copies. The best answers have paragraphs that are 3-5 sentences max. Sometimes just 2. Long paragraphs get skipped. Short paragraphs force the examiner to read every line.

Diagrams. Here's what surprised me — the toppers with highest marks DON'T use diagrams in every answer. They use them in maybe 1 out of 4 or 5 answers. But when they do, the diagram carries a huge amount of meaning. A good flowchart showing cause → mechanism → effect is worth more than a paragraph of text. A bad diagram — and I saw plenty of those — just wastes time.

Handwriting. I'll be honest — most topper handwriting is average. It's legible but not beautiful. The exception is people like Divya Tanwar (AIR 105) whose handwriting is genuinely clean and consistent. But even the messier copies scored well because the content was strong.

What I'm trying to say is — don't obsess over making your copies look like calligraphy samples. Focus on clean, readable, and strategic use of space.

I've been studying these copies for months and there are things I still can't figure out.

Some toppers with average content score really well. Some with great content score average. The marking seems inconsistent in ways I can't explain.

I suspect a lot depends on which examiner you get, what time of day they're reading your paper, and what answer they read just before yours. That's not something you can control, and I think pretending otherwise is dishonest.

What I do know is that the toppers who consistently score 140+ across multiple papers — people like Garima Lohia, Ishita Kishore — they're not lucky. They're doing something repeatable. And the closest I've come to identifying it is: they write like they're explaining something to a smart friend, not like they're answering an exam question.

They don't use jargon unless it's necessary. They don't write long, winding sentences. They make one point, back it up, and move on.

If I had to distill everything I've learned into one piece of advice it would be: write like you're the authority on the topic, not like you're trying to prove you studied.

Short answer: yes.

Longer answer: practicing wrong doesn't help.

I see a lot of aspirants writing 3-4 answers daily and not improving. The reason is almost always the same — they're not reviewing what they wrote against what a good answer looks like.

If you write an answer and nobody tells you what's wrong with it, you'll just get better at writing the same mistakes.

The toppers I studied didn't practice more than others. But they practiced deliberately. They wrote an answer, compared it with a real topper copy, and asked specific questions: "Where did I put my data? Did I open with a thesis or a definition? Is my argument clear or am I listing points?"

That kind of practice works in weeks, not months.

I built an AI tutor that does exactly this — you give it your answer, it compares it against the patterns I noticed in topper copies, and tells you what to fix. It's not perfect (it's an AI, it misses nuance), but it's better than writing into the void. You can try it here.

You have about 7 minutes per answer in the Mains. That's not a lot.

Here's what I've noticed about how toppers handle time:

They spend the first 60 seconds just reading the question. Sounds obvious, but most people don't do this. They scan the question and start writing immediately. Toppers read it twice, underline the actual demand, and think for a moment about what angle to take.

Then they write for about 5 minutes — not stopping, not editing, just getting the content down.

The last minute is for underlining key terms, adding a diagram if it fits, and writing a closing sentence.

When I timed Garima Lohia's copies against this breakdown, it matched almost exactly. She's not faster than everyone else. She's just more disciplined about how she uses those 7 minutes.

The people who run out of time aren't slow writers. They're people who started writing before they knew what they wanted to say, then had to rewrite or restructure halfway through.

If I were preparing for Mains right now and wanted to improve my answer writing, here's what I'd actually do — not what sounds good in a blog post.

First, I'd get my hands on real answer copies. Not photocopies from some Telegram group where half the pages are missing. Actual complete answer sheets from verified toppers. I'd read 10 of them cover to cover before writing anything myself. Just to recalibrate what a good answer looks like. (I've put together a compilation of 50+ copies if you want a shortcut — these are the ones I studied for this guide.)

Second, I'd write one answer per day for 2 weeks. Not 3-4. One. But I'd spend 20 minutes reviewing it afterwards — comparing it to how a topper answered a similar question, noting where I fell short, rewriting the opening sentence until it sounded like something a confident person would say.

Third, I'd stop trying to be comprehensive. The best answers don't cover everything. They pick 2-3 strong points and develop them well. A focused 140 is better than a scattered 100.

Honestly, most of what I read about answer writing is overcomplicated. You don't need a "framework" or a "template" or a "proven formula." You need to know your material well enough to have a genuine opinion about it, and you need the confidence to state that opinion clearly.

The copies in my compilation show you what that looks like in practice. I can try to explain it in words, but seeing 50 real answers side by side will teach you more than I ever could in a guide.