What changed was not the number of questions I did. It was having a framework to hang them on, and someone telling me to show up again the next day.
Two numbers frame everything I have to say here. One is Q90, a perfect Quant on my most recent attempt. The other is 20 percent, which is what my first two hard CR cementing quizzes scored. The distance between them is the story. Until I found e-GMAT I was working out of the Official Guide on my own, and the trouble was never effort. I was getting through problem after problem without ever grasping the concept underneath any of them, because there was no framework to hang them on. Eventually it became obvious that more questions were not going to fix that, and I went looking for a structured platform instead. A friend of mine, Sarthak, had used e-GMAT himself and pointed me to it for two reasons: how thorough the orientation was, and the pre-thinking approach. The Verbal turnaround
Verbal is where the change was sharpest. My first hard CR cementing quizzes came back at 20 percent. Later on, in a single attempt, I hit 85, and I am at 81 now, with that dip down being a completion penalty rather than anything to do with the concepts. The platform data says the same thing from another angle. Hard CR L20 accuracy, which the system reads off my last twenty hard questions in a subsection, was sitting at 45 and now reads 70. Hard RC went from 70 to 75 over the same period. My verbal sectionals had been stuck in the high seventies for a long stretch; they cleared 82, then ran to 84 and 90.
What moved them was not volume. It was pre-thinking, which gave me something specific to do before I ever looked at the answer choices. That is exactly the step I had been skipping for months in the Official Guide, and the accuracy climbed once it was in place.
The hard part
The worst stretch was about a month of Verbal where I did nothing but attempt questions and nothing moved at all. A plateau like that wears you down more than a bad score does. What broke it was the LMP support and the conversations with my mentor. I had not been consistent before those; afterwards I started showing up every day. That mattered more in my case than it might in most, because I am in a remote part of Odisha with nobody around me preparing for the same exam and no one to talk prep with. Having someone there to encourage me when the results were not moving was a large part of why I kept going at all.
What I leaned on
Scholaranium and Neuron did most of the heavy lifting on accuracy, because between them I could see exactly which concepts and question types were still weak and go straight at them. DI is where that loop is easiest to see. Two Part Analysis went from 40 percent to 90 percent, and not by doing more DI generally: I practised the type on Neuron and then re-tested it with a cementing quiz. Graphics Interpretation took the same route with one addition, which was analysing every question I had got wrong before touching another quiz; the focused Neuron set after that cleared at 80 percent. Multi Source Reasoning needed something different again, mapping the sources up front and noting what each one held rather than re-reading all three every time, and that practice set finished at 80 percent as well. By the end my hard DS accuracy was 85, TPA 80, Table Analysis 75.
DI77 on the day is still where I have the most room left, and my hard GI and TPA numbers agree with that. It is the honest gap in a result I am otherwise happy with.
Quant, and the mocks
Quant closed quietly. Hard Number Properties and hard Algebra both finished at 95, Advanced Topics and Word Problems at 80. There is no dramatic story there, which is rather the point: the modules came in a sequence, I worked them in order, and the accuracy arrived where it needed to. The mocks moved the way people warn you they will. Mine read 595, then 645, 665, 675, and 695 in the run-up to the exam. The dips along the way were never worth reading much into. What counted was the direction, and the fact that the section which used to pull the whole score down had stopped doing it.
Signing off
If you are starting out, start on a structured platform rather than a stack of questions. And if you are somewhere remote like I am, e-GMAT together with the LMP program can genuinely change things, simply because of how extensive and well built the course is.