Why your GMAT score is stuck.

My first GMAT, I scored 565. I'm not a native English speaker. I have no engineering background. I'd been studying for three months, doing what every guide tells you to do — work through chapters, do practice sets, take mocks, repeat. I was averaging two hours a day. I felt like I was getting better. I wasn't.

For a while I told myself the score was a fluke. So I doubled down. I bought a more popular course. I did more questions. I took a mock every week. My score moved by 15 points in eight weeks. At that pace I was on track to hit my target sometime around 2032.

The thing nobody told me — the thing I now think is the single most useful sentence in GMAT prep — is this:

You don't get better at the GMAT by doing more questions. You get better by understanding the questions you already got wrong.

The ratio is brutal. A question you missed and analyzed properly is worth ten questions you missed and shrugged at. The reason most prep stalls — and this is true at 540, 620, 690, anywhere below where you want to be — isn't a content problem. It's a review problem. People do hundreds of questions a week and review zero of them with any rigor.

The plateau is a review failure, not a content failure

Here's what my study sessions looked like at 565:

1. Open a problem set.
2. Do 20 questions, untimed.
3. Check the answers.
4. Read the explanations for the ones I missed.
5. Move to the next set.

On paper, that's "reviewing." In reality, it's a content-consumption loop with the word "review" pasted on top. I'd read an explanation, nod, and forget it within a week. I'd see the same trap two months later, fall for it again, read the same explanation, and assume I just "hadn't internalized it yet." The honest read was: I'd never really engaged with it the first time.

The plateau wasn't about content I didn't know. By the time I scored 565 I'd covered every Quant topic in the standard curriculum twice. The plateau was about patterns I kept missing — and I had no system to actually catch them.

What changed at month four

I started keeping an error log. Not a fancy one. A spreadsheet with one row per missed question and the following columns:

• Question source. Where I got it.
• Section + topic. Quant / Algebra. Verbal / Critical Reasoning. Etc.
• Mistake type. One of six tags I made up that week: Conceptual, Careless, Time pressure, Misread, Strategy, Other.
• Trap I fell for. One sentence on what the question was actually testing and why I missed it.
• Fix. What I would do differently next time. Also one sentence.

For the first two weeks I just logged. No drilling, no re-attempts, no plan. Just: every miss goes in the log, with those five fields, no exceptions.

At week three I sorted the log. The pattern was almost embarrassing. Out of 140 missed questions, more than half were three things:

1. Inequality sign flips. I'd forget to flip when multiplying by a negative. I'd done this on every other Algebra section for three months and somehow never noticed because it was spread across topics.
2. CR "assumption" questions where I'd eliminate the right answer because it "felt too broad." The trap was that broadness was the point — assumptions usually are broad.
3. DI table-analysis questions where I'd try to compute exact values when the question only required a comparison. I was burning two minutes per question for no reason.

Three patterns. Across hundreds of questions. The content was never the problem. The same five mental moves, repeated against different surface forms.

The system that came out of the log

I built a study loop around the log. It looked roughly like this:

1. Drill 20 questions. Topic-filtered, untimed the first pass.
2. Log every miss with the five fields above. Five minutes per missed question, no shortcuts.
3. End-of-week review. Read every entry from the week. Look for patterns. Group by mistake type and trap.
4. Weekly drill on the dominant pattern. Whatever pattern was the largest cluster that week got a 30-question targeted set. Same trap, different surface forms.
5. Spaced retrieval on prior misses.Every Sunday I'd redo five questions from two weeks ago without looking at the answer. If I missed them again, they got re-flagged with a different trap tag if applicable.

That's it. That was the loop. No course, no lecturer, no videos. The content I'd already covered was fine. The scaffolding I'd built around it was the difference.

From 565 in October, I hit 645 in late January. 695 in March. 735 in May. The slope wasn't linear — most of the climb happened once the log had eight to ten weeks of data and the patterns were visible. Below that, you don't have enough signal to act on.

Why most prep advice misses this

Two reasons. One commercial, one cognitive.

The commercial reasonis that it's easier to sell content than to sell rigor. A platform that says "700+ questions, full-length mocks, expert video explanations" is an easier purchase than a platform that says "a system that forces you to be honest with yourself about your mistakes." One sounds like a product. The other sounds like work.

The cognitive reasonis more interesting. Reviewing a missed question well is uncomfortable. It requires you to sit with the specific reason your reasoning failed. People don't want to do that. They want to read an explanation that makes the right answer feel obvious in retrospect, nod, and move on. That's consumption. It feels like progress, but it doesn't change behavior — because behavior is changed by the moment of recognition where you say "oh — I do this every time, and here's why." Most students never reach that moment because they don't structure their review to force it.

What this means for you, if you're stuck

If your score has been flat for two months, the diagnosis is almost always one of three things:

1. You don't have an error log.Or you have one and don't use it. This is the common one. Start one. Five fields per row. Nothing fancy.
2. You log but you don't aggregate. The log is useless until you sort it. The patterns only show up at the tag level, not the question level.
3. You aggregate but you don't drill the patterns. Knowing your dominant trap is necessary but not sufficient. You have to do targeted drills on it until the recognition becomes automatic. A pattern you can name but can't catch in real time hasn't actually been fixed yet.

Notice that none of those three problems are about content mastery. None of them are solved by a longer course, a more expensive tutor, or a better book. They're solved by the review layer.

The platform

I built Zakarian GMAT because the spreadsheet I used to climb from 565 to 735 deserved to be a real product. The error log is built in. The patterns are aggregated for you. The spaced retrieval is automatic. The diagnostic finds your top three traps in 30 questions. The chapters teach the content but the point is what the system does with your mistakes between chapters.

The system isn't magic. It's the discipline of doing the review work, made easier by software so you actually do it. If you've been stuck and you're ready to take your mistakes seriously, the diagnostic is free and the rest is one click away.