GMAT 695 (Q89, V85, D80): from a 585 first mock to the 97th percentile with e-GMATHow I pulled my own mistakes apart, refused to run a method the way the textbook drew it, and turned the section I almost gave up on into one of my best.The gap I almost planned aroundI came back to the GMAT after giving the old legacy version 3 years ago, so I had some verbal instinct left and a Quant base I trusted, but no Data Insights at all and a nagging memory that I could never make my verbal consistent. My first mock came back at 585, 62nd percentile, and honestly my plan was to push Quant and DI as high as they would go and let verbal land where it landed. That plan is the one thing I am most glad I dropped.
Why I stayed on this platformI did not shop around for months. What I needed was simple: a way to see exactly where I was leaking points, because a "decent" score hides that better than a bad one does. e-GMAT gave me Scholaranium (a large practice bank with topic-level analytics and custom timed quizzes), so instead of "get better at verbal" I had a ranked list of the question types actually costing me. That fit the way I already think.
Critical Reasoning: building my own pre-thinkingCR was where I actually struggled, and it was not the reading, it was the timing. I kept getting stuck between 2-3 options and could not separate them, and single questions were eating 4-5 mins. On the hard CR cementing quizzes (short, exam-level quizzes with a clear pass threshold) I kept missing the last 2-3 questions, and my hard CR accuracy bounced around 40-50%.
The course teaches a pre-thinking method, which is basically deciding what a strong answer should look like before you read the options, and I was openly skeptical because done in full it was too slow for me. Instead of dropping it, I took it into the Neuron OG questions (official-style questions with worked solutions that you can run untimed) and experimented. Skipping pre-thinking entirely dropped my accuracy, doing every step blew my clock. So I built a hybrid: for a strengthen, weaken or assumption question I would skip straight to picturing 1-2 things the answer had to do, and for inference or bold face questions, where pre-thinking honestly does not help, I would go straight to the options. Then I re-drilled the quizzes with that rule. Hard CR cementing climbed out of that 40-50% band up to 80%, and the time per question finally came down.
Quant: the word-problem wallQuant was my strength and my target was a perfect score, so word problems stung. On the hard word problem cementing quizzes I dropped to 30%, then 40%, and each question was taking me 3-4 mins. It was tempting to write the whole topic off as a weakness.
What my mentor and I ran instead was an error loop inside the e-GMAT course: build custom quizzes out of the exact questions I had marked wrong, then go back and find where the time was actually going. It was not the math. It was the comprehension, translating a wall of words into equations, which ate more than half my clock. So I started reviewing the questions I got right as well, hunting for faster setups, and cross-checking approaches on the discussion forums (where you can post any doubt and the experts reply, usually within a day). Across 3 days those hard quizzes climbed 30%, 40%, 60%, 80%. On test day Quant came in at Q89. Data Insights: a time game, not a knowledge gameDI was new to me, and it is really a mix of Quant and verbal under a brutal clock. e-GMAT's DI course got me familiar with every question type (data sufficiency, table analysis, graphics interpretation, two-part, MSR), and then the DI sectional mocks (full timed sections that build pacing under pressure) built the actual execution. I learned to bank time on the quick question types so I could pour 4-5 mins into an MSR. My rule became simple: the test is for 18 questions, not 20. Maximize accuracy on those 18 and guess the last 2, since a blank costs more than a wrong answer. Across my sectional mocks that pulled DI from a shaky start into the 82-88 range against my target of D81.
Honest note: on the real day DI was my weakest section at D80. A monster MSR and a graphics interpretation question I stubbornly refused to let go ate my time. The strategy worked in 7-8 mocks and misfired once, on the day that counted, but I would still run it again.
The mentorship that changed the planAround my 4th week of verbal I was inconsistent, 78 one sitting, 85 the next, and I was ready to abandon the section. This is where the Last Mile Push mattered. My mentor, Jagadish, would not sign off on giving up. He said trust me, then handed me small, specific builds: these 5 inference questions you got wrong, review them again, then quiz yourself on them. It felt like wasted weeks. Then the full mocks arrived, my verbal settled into 83-86, and I was doubting far fewer questions per section. On test day verbal was V85, one of my best. If you are in the closing phase, look into LMP. The value was not magic content, it was someone reading my data and refusing to let me quit on a false belief.
Mocks bounce, so let themMy full mocks swung between 635 and 695 in the final stretch, and once I accepted the bounce as normal I stopped letting it rattle me. Test day landed at 695, 97th percentile.
The one thing I would tell myselfThe GMAT is an exam about edge, not knowledge. The content is not the hard part, how fast you strategize is. If a question that took you 5 mins earlier now takes 4, that is a real jump worth respecting, and it becomes 3, then 2, if you stay diligent and trust the process. That is the whole game.