Every year after CAT results come out, I get some version of the same question. Someone scored 95 or 96 percentile, worked hard for months, and wants to know what the 99 percentile crowd did differently. The honest answer isn't satisfying at first. It's rarely one big secret. It's five or six things done consistently, while most people do three of them occasionally.

It starts with accuracy, not attempts

This is the mistake I see most often. Students stuck in the 90 to 95 range tend to chase attempt count instead of accuracy. CAT's negative marking is unforgiving, and 90 percent accuracy on 18 questions almost always beats 60 percent accuracy on 26. The students who cross 99 decide in advance how selective they're going to be, and they stick to that plan even when a paper feels easy and the temptation to attempt more creeps in.

The section-wise gap is smaller than you'd think

In Quant, the students who do well go deep on Arithmetic and Algebra before spreading effort across the whole syllabus, since these topics show up constantly and reward speed more than Geometry or Modern Math. In VARC, the real bottleneck is usually reading speed, not comprehension, and something as simple as twenty minutes of editorials a day compounds faster than most people expect. In DILR, the skill that actually separates scores isn't solving hard sets, it's knowing within the first minute which ones to abandon. Losing eight minutes on an unsolvable set costs far more than walking away early.

Mock analysis is the quiet difference maker

Taking mocks without reviewing them properly is one of the most common ways students plateau despite putting in real hours. After every mock, the students who improve fastest split their mistakes into careless errors, genuine concept gaps, and questions they wrongly skipped, because each of those needs a completely different fix. Lumping them together into a vague sense of what went wrong is exactly why the same mistakes keep repeating.

Why feedback quality matters more as you get closer

Once you're already around 95 percentile, a plain scorecard stops being useful. The mistakes at that level are subtle, a slightly slower approach on an easy question, a set abandoned two minutes too late, and they don't show up clearly in a percentile number. This is the specific gap CATKing has built its higher-level mock series around, mock reviews walked through by IIM alumni instead of an automated report, plus a separate track for students specifically targeting 99 plus, since that feedback needs to be sharper than what a student still building fundamentals requires.

The takeaway

Most people in the 95 to 99 percentile range already know roughly the same syllabus. What separates them isn't intelligence applied occasionally, it's discipline applied consistently, fewer careless mistakes, sharper selection of what to attempt, and an honest, structured review after every single mock. That's really the whole strategy, and it holds up whether you have ninety days left or nine months.