Most people who fail an AWS exam didn’t lack the intelligence or the hours. They spent the hours badly. Here are the patterns that show up over and over.

1. Watching instead of building

The most common one by a distance.

A video course feels like progress. You finish a module, you understood it while it was happening, you move on. Then you hit an exam question about a private subnet and realize that watching someone configure a route table taught you almost nothing about configuring a route table.

**Instead:** for every hour of video, spend an hour in the console. [Six projects worth building](/blog/aws-hands-on-practice-free-tier/), in order. If a study week goes by without you opening AWS, that week didn’t count.

2. Memorizing services instead of learning categories

People make flashcards of service definitions. “Amazon SQS: managed message queuing service.” Then the exam describes a scenario without naming any service, and the flashcard doesn’t help.

**Instead:** for every service, answer four questions. What problem does it solve? What goes in? What comes out? What’s the closest alternative and when would you pick that instead? The exam tests the fourth question more than the first.

3. Ignoring the official exam guide

An astonishing number of people study from a course without ever reading the actual exam guide. That document lists the domains, their percentage weightings, and the in-scope services. It’s the blueprint the exam is literally built from, and it’s free.

**Instead:** read it first, and organize your study time by domain weight. If a domain is 30% of the exam, it gets roughly 30% of your hours. Don’t spend three weeks on something worth 8%.

4. Question dumps

Beyond the ethics and the fact that it violates the exam agreement, it doesn’t work. AWS writes exams specifically to defeat pattern recognition, and the scenarios are rewritten regularly.

What actually happens to people who dump: they either fail because the questions don’t match, or they pass and then get taken apart in a technical interview by someone who asks a follow-up question. The second outcome is worse and harder to recover from.

**Instead:** use legitimate practice questions as a diagnostic. The score isn’t the point. The pattern in your wrong answers is.

5. Reviewing only the wrong answers

This feels efficient and leaves half the value behind.

If you answered a question correctly by eliminating two obvious wrongs and guessing between the last two, that’s a gap. It happened to break your way this time. It won’t next time.

**Instead:** review every question. For the ones you got right, ask whether you *knew* it or *guessed* it. Track the guesses like they were misses.

6. Not practicing under real conditions

A practice exam taken across three sittings, with notes open and the phone nearby, tells you very little.

The real exam is one continuous stretch of dense reading under time pressure. Accuracy drops when you’re tired in ways that a relaxed untimed session will never show you. People who’ve only done untimed sets get blindsided at question 45.

**Instead:** at least three full-length timed exams, no pauses, no lookups, ideally at the same time of day as your booking. [How many practice exams you actually need](/blog/how-many-practice-exams-before-you-book/) covers the sequencing.

7. Booking too early or too late

Too early: you book as a forcing function, aren’t ready, and pay $150 to confirm it. Then you wait 14 days and pay again.

Too late: you study for seven months, learn things that aren’t on the exam, and lose momentum entirely. Open-ended study stretches to fill whatever time you give it.

**Instead:** book when your full-length practice scores clear the pass mark consistently with margin. Not on a feeling in either direction.

The meta-mistake

Underneath most of these is the same thing: **confusing effort with progress**.

Hours logged, videos watched, pages read, questions answered. None of those are the goal. The goal is being able to read an unfamiliar scenario and reason to the right answer under time pressure.

So every week, ask one question: *can I do something this week that I couldn’t do last week?* Not “did I study” but “can I now design a VPC, or explain when to use SQS over SNS, or read a scenario without getting lost.”

If the answer is no for two weeks running, the approach needs changing, not more hours.

If you already failed

None of this is a verdict. Most people who fail once and fix their approach pass the second time comfortably. [What to do after failing an AWS exam](/blog/failed-aws-exam-what-next/) walks through reading your score report and working out which of these seven actually got you.