The first thing most people want after a failed attempt is not study advice. It is the rules: how many more tries, how long they have, and whether they have just wasted the whole application.
So the rules come first, and then the part that actually decides the second attempt.
What PMI Allows
These come from PMI's own certification handbook, so they are the authoritative version rather than forum recollection:
- Once your application is approved, you have one year to take the exam. That window is the eligibility period.
- You may take the exam three times within that one-year period.
- If you fail all three times inside the eligibility period, you must wait one year from the date of your last attempt before you can reapply.
Three things follow from that, and they are worth being deliberate about.
Your application does not die with a failed attempt. The eligibility period is what you are spending, not the application. If you failed in month three, you still have nine months and two attempts on the same approval.
The one-year lockout only triggers after the third failure. A second attempt is not a cliff edge. A third one is, so the gap between the second and third attempt is the one that deserves real planning.
And re-examination is a paid attempt. PMI charges a re-examination fee, lower than the initial exam fee, and PMI's fee schedule varies by membership status and region — check the current figures on pmi.org for your location rather than trusting any number you read in an article, including this one.
How to Read the Score Report
The score report does not give you a percentage. PMI does not publish a numeric passing score, and the report does not show you one. What it gives you is a rating on each of the three domains, on a four-level scale:
| Rating | What it means |
|---|---|
| Above Target | Performance exceeded the requirement for that domain |
| Target | Performance met the requirement |
| Below Target | Slightly under the requirement |
| Needs Improvement | Well under the requirement |
Plus an overall result. Target or Above Target overall means you passed; Below Target or Needs Improvement overall means you did not.
This is genuinely less information than candidates expect, and it leads to a predictable mistake.
The Mistake Almost Everyone Makes Next
The score report says Below Target in Process. So the plan becomes: study Process harder, drill more Process questions, rewatch the Process videos.
That reads like the obvious response, and it is usually the wrong one.
A domain rating tells you where your errors landed. It does not tell you why you made them. Those are different things, and only the second one is fixable by studying.
Consider two candidates who both got Below Target in Process. The first genuinely does not know how the change control path works — an actual knowledge gap, and more Process study will fix it. The second knows change control perfectly well but reflexively picks the documentation option when the exam is asking what to do first, and that instinct misfires across Process questions because Process questions are where sequencing gets tested. More Process content teaches the second candidate nothing. They already know the content. They lose on a reflex that will follow them into the retake unchanged.
In our experience with candidates who fail a first attempt while having done substantial practice, the second pattern is far more common than the first. Someone who has worked through a thousand practice questions rarely has a pure knowledge hole large enough to sink a domain. They have an instinct that is consistently one step off.
Finding the Actual Failure Pattern
The diagnostic is boring and it takes about an hour.
Go back to your practice questions — not new ones, the ones you already did — and pull forty that you got wrong. For each, write one sentence: why the answer you picked looked right at the time. Not why the correct answer is correct. The key already told you that, and it is not the useful information.
Then read all forty sentences in one sitting. The clustering is usually blunt:
- "I escalated when the situation had not actually escalated yet."
- "I picked the process-correct option when the question asked what to do first."
- "I chose the option where the PM decides, because the team's approach looked wrong to me."
- "I answered what the PM should have done when the question asked what to do now."
Most candidates find four to eight of these, and those few reflexes generate the large majority of their errors. That short list is your retake plan. It is worth more than the next thousand questions, and we have written about the mechanism in more detail in how many practice questions you actually need.
A Four-Week Retake Plan
This assumes you already studied properly the first time and failed anyway — which is the situation this article is about. If you went in underprepared and know it, you need the ordinary study plan, not this one.
| Week | Focus | Volume |
|---|---|---|
| 1 | Error autopsy on past misses; write the failure-pattern list. Re-read only the ECO tasks tied to your weakest domain. | 40 review sentences, ~150 questions |
| 2 | Drill specifically against your named patterns, in small filtered sets. Check after each set whether the pattern repeated. | 300-400 questions |
| 3 | Two full-length timed mocks, written review of every miss against the pattern list. | 360 + review |
| 4 | Re-drill whatever survived. One final fresh mock. Taper the last two days. | 200 + 180 |
Two notes on that.
The mocks in week three need to be fresh questions. Redoing a mock you have already seen measures recall of that mock, not readiness, and the scores it produces are cheerfully misleading. If you exhausted your bank on the first attempt, this is the one legitimate reason to buy access to a second set of questions.
And the pattern check in week two is the part that does the work. After each filtered set, look at your misses and ask whether they are the same reflex again. If they are, the drilling has not landed yet and another 300 questions will not change that — go back to the sentences.
When to Rebook
There is a real tension here. Sitting the exam while the experience is fresh preserves your feel for the question style and pacing. Waiting longer gives you time to actually change something.
The practical answer for most people is three to five weeks. Sooner than about three weeks and you are usually re-sitting with the same instincts, which produces the same result at full price. Much beyond six or seven weeks and motivation tends to decay, the material goes stale, and the retake keeps getting pushed.
One exception worth naming: if your first attempt failed on pacing rather than judgment — you ran out of time, or you were exhausted by question 140 — then timed full-length practice is the fix and you can move faster. Stamina responds to training much more quickly than instinct does.
What Not to Do
Do not buy a second full course. If you had the content well enough to reach Target in one domain, you have the content. A second course is expensive procrastination that feels like effort.
Do not redo your existing question bank from the beginning. By the second pass you are recognizing items rather than reasoning about them, and your scores will rise for reasons that have nothing to do with the exam.
Do not treat a Needs Improvement as a verdict on yourself. It is a rating on one domain from one four-hour sitting. The exam is 180 questions in 240 minutes, weighted 33 percent People, 41 percent Process and 26 percent Business Environment, and a bad hour in the middle of it can move a domain rating on its own.
Do not skip the error autopsy because it feels unproductive. It is the only step in the whole plan that produces information you did not already have.
The Short Version
You have a year and three attempts on one approved application, and a one-year wait only if all three fail. The score report tells you where your errors landed, not why you made them. Find the why — forty sentences, one hour — and drill against that instead of against a domain label.
Second attempts have a good record, but not because people study more the second time. It is because failing once finally makes the specific problem visible.
If you need a fresh set of questions to re-drill against, start with 50 free practice questions — no card required — and sort your misses by pattern rather than by domain.
---
Exam rules and attempt limits in this article were verified against PMI's certification handbook on 20 September 2026. Fees change and vary by region, so confirm current figures on pmi.org. Disclosure: we run a PMP question bank, so treat the closing recommendation accordingly — the diagnostic method above works with any bank, including the one you already own.