Running out of time on a GMAT section is the single most common reason a well-prepared candidate walks out of the test centre with a score two or three points below their practice range. The exam does not reward the slowest careful reader or the most meticulous arithmetic checker. It rewards the candidate who, on every section, knows exactly which questions to guess on, which to bisect, and which to solve in a single clean pass. This article lays out a triage protocol for the moment the clock turns against you: a sequence of moves that protects the score you can still earn rather than chasing the score you have already lost.
Three structural facts about the GMAT shape every timing decision. First, each section is one adaptive block that adapts question by question: after every answer the test updates its estimate of your level and picks the next question, so the difficulty you see depends on how you have been answering, not on a fixed position in the section. Second, the test picks each next question from your answers so far, and GMAC advises answering every question, even with an educated guess. Third, you can bookmark questions and, if time remains at the end of the section, review them and change up to three answers. The clock, however, is uniform: the same minutes-per-question average applies whether the section is serving you easier or harder items. That mismatch — variable difficulty, fixed time — is the engine of nearly every timing collapse. Pacing on the GMAT is, in practice, the management of that mismatch.
Diagnose the timing collapse before you triage
Most candidates who feel the clock slipping do not actually have a time problem. They have a question-selection problem that they are misreading as a time problem. The first move is therefore diagnostic, not tactical. Ask which of three failure modes describes your section.
The first failure mode is front-loaded slow-down. You burn eight or nine minutes on the opening two questions because they look hard, you refuse to let go, and by question 6 you are already two minutes behind budget. The correct response is to accept a 90-second cap on the opening two or three questions: the moment you cross it, pick your best answer, bookmark the question and move on. The cost of guessing on question 4 in order to finish question 21 cleanly is almost always lower than the cost of spending four minutes on question 4 and rushing the last seven.
The second failure mode is mid-section friction. The first ten questions went fine, but questions 11 to 15 produced a string of items where you read the stem three times, eliminated two choices, then hesitated between the remaining two for over a minute. This is the classic Data Sufficiency trap in Data Insights or a Critical Reasoning strengthen-versus-assume confusion in Verbal. The fix is a hard stop after 60 seconds of elimination without commitment. Pick your best guess, bookmark the question, and move. The expected value of returning to a bookmarked question at the end of the section is rarely above the expected value of giving the next two questions the time they need.
The third failure mode is end-of-section panic. Eighteen questions are answered, the clock shows 4 minutes left, and there are several questions remaining. Most candidates in this position spend the entire four minutes on the next question and leave the rest unanswered, which is the worst outcome: GMAC counts the number of questions you leave unanswered in the score calculation and advises answering every question, even with an educated guess. The correct response is to switch to triage pace: give each remaining question a quick first look as it appears, answer it properly if it resolves fast, and otherwise eliminate what you can, guess and move on, so that every question gets an answer before the clock hits zero. The remaining questions are not a block to be solved in full. Each one gets a decision.
The 90-second rule and where to apply it
On the GMAT Quant section, candidates have 45 minutes for 21 questions, which works out to just over two minutes per question. On Verbal, the ratio is similar: 45 minutes for 23 questions. On Data Insights, candidates have 45 minutes for 20 questions across five question types, and the per-question budget feels tighter because Two-Part Analysis and Multi-Source Reasoning items carry a heavier reading load.
The 90-second rule is a single tactical constraint: no single question in a section consumes more than 90 seconds of active work before you decide whether to commit or guess. Active work is the time from the moment you finish reading the stem to the moment you click an answer choice. The cap exists for a structural reason: a single wrong answer is recoverable, but a cluster of questions left unanswered at the end of the section, or rushed into random guesses, is not.
Where the 90-second rule bends
Two question families legitimately justify a longer investment. The first is a Data Insights Graphics Interpretation item that requires extracting a value from a multi-axis chart and then performing a ratio calculation. These items can take 110 to 120 seconds without indicating a timing problem, because the alternative — guessing on every chart question — collapses the entire section. The second is a Reading Comprehension inference question near the end of a long passage, where 30 extra seconds of re-reading the relevant paragraph is the difference between a correct inference and a plausible-sounding distractor. Everywhere else, the cap holds.
Where the 90-second rule tightens
On the first three questions of a section, tighten the cap to 75 seconds, so that a nervous start does not eat the buffer you will need later. On the final two questions, if time is short, tighten the cap to 60 seconds so that both get a considered answer before the clock runs out; GMAC advises answering every question, even with an educated guess.
Guess-and-bookmark as a scored tactic, not a failure
Most candidates treat the bookmark button as a confession of weakness. The opposite is true. On the GMAT, guessing and bookmarking is one of the highest-leverage tools for protecting a section score when the clock is moving faster than your problem-solving. A bookmark is a deferred decision: you have answered the question, it is logged, and if time remains at the end of the section you can return to it and, within the limit of three answer changes per section, revise it. Meanwhile you are free to engage with the next item in full.
For this tactic to pay off, the discipline must be mechanical. When a question crosses the 90-second line, the move is identical every time: click the answer choice you would defend if asked right now, click the bookmark, click next. Three clicks, two seconds. GMAC advises answering every question, even with an educated guess, so do not leave a question blank to come back to; the only choice is how good your committed guess is. Make it the best answer you can defend in that moment, then bookmark it for review.
How much a second look is worth
If you finish the section with a few minutes remaining and three bookmarked questions, a second look is worth taking. If you finish with very little time, it rarely is: a meaningful reconsideration needs about 30 seconds per question, and a hurried glance is rarely enough to overturn a considered first answer. You can change at most three answers per section in any case. The pragmatic move with little time left is to spend it on the single bookmarked question whose stem you remember most clearly and leave the others as they are.
Triage ladders for each section when the clock is winning
Generic pacing advice rarely helps because the three sections fail in different ways. Quant collapses when candidates refuse to let go of hard items. Verbal collapses when candidates get stuck in elimination loops. Data Insights collapses when candidates read every chart in full. Each section needs its own triage ladder.
Quant triage ladder
When the clock shows 10 minutes left and 6 Quant questions remain, you have well under the two-minute average for each of them. As each question appears, give it a 20-second first look. If it is a single-step arithmetic or direct proportion problem, solve it. If it is not, eliminate the choices that are clearly impossible, pick the best of the rest, bookmark it and move on. The aim is that all six get an answer before time runs out, with your real effort spent on the ones that resolve quickly.
Verbal triage ladder
Verbal triage covers the only two Verbal question types, Reading Comprehension and Critical Reasoning. Because questions arrive one at a time, you cannot choose which to answer first; you can only choose how much time each gets. RC questions attached to a passage you have already read are cheap, so answer them carefully. A new passage near the end is expensive, so read it for structure only and answer from the lines the question points to. On a long Critical Reasoning stimulus with little time left, find the conclusion, eliminate the answer choices that do not touch it, guess among the rest and bookmark.
Data Insights triage ladder
Data Insights collapses on reading load, not arithmetic. A candidate who is behind on time in DI is almost always behind because they are reading every chart in full, then re-reading the table, then re-reading the paragraph. The triage move is to read the stem first, not the stimulus. The stem tells you which value you need from the chart. Extract that value only. If the chart does not contain the value the stem is asking for, you have answered the question in under 30 seconds without ever reading the chart's title or its secondary axis.
When the questions start to feel harder: how the triage plan changes
Because the GMAT adapts question by question, a run of correct answers usually brings harder questions, and a run of misses brings easier ones. The test does not switch you into a separate easy or hard stage; the difficulty simply follows your answers. That gives you a useful reading of how the section is going, and the pacing plan can respond to it.
When the questions feel manageable, the dominant failure mode is over-checking: you solve in 80 seconds, then re-verify in 40 seconds, then click a slightly different answer because of an uncertainty you cannot name. Here the timing problem is rarely a clock problem and almost always a confidence problem. The fix is to commit to the first answer that survives a 30-second sanity check and move on.
When the questions feel hard, that is often a sign you have been answering well. The dominant failure mode is the opposite: you engage with a hard item, burn three minutes, and either answer it correctly with too little time for the next four, or give up after three minutes and still have too little time for the next four. The triage move is to give hard questions a firm cap, guess and bookmark any item where the cap is reached without a confident answer, and revisit bookmarks only if you reach the end of the section with time left. The score-protecting outcome is a section in which every question has a considered answer, not one in which the last few were left unanswered.
| Situation | Dominant failure | Per-question cap | Guess and bookmark when | Review at the end? |
|---|---|---|---|---|
| Questions feel manageable | Over-checking, second-guessing | 75 seconds | No answer after 2 minutes | Only if several minutes remain |
| Questions feel hard | Refusal to let go, sunk-cost thinking | 90 seconds | No confident answer at the cap | Bookmarked items only, single pass |
| Behind the clock | Panic, unanswered questions at the end | 60 seconds | Any item that does not resolve fast | Rarely; protect the remaining questions first |
Common pitfalls and how to avoid them
The first pitfall is treating the review screen as a planning tool. The GMAT does not show you which questions you got right, so the review screen at the end of a section is simply a list of your bookmarked items, and you can change at most three answers. Candidates who spend the last 30 seconds staring at the screen, debating which bookmark to open, lose the chance to open any of them. The review screen is a launchpad: pick the question, click, reconsider, submit. The debate happens in your head before you reach it, not on the screen.
The second pitfall is panic-substitution. In the last minute of review, candidates often change answers simply because they were made quickly, reasoning that rushed answers must be wrong. A rushed change is no more reliable than the rushed original. The single rule that protects against panic-substitution is: change an answer only when you can name the specific error in your first reading. Leave the other answers alone, especially since only three changes are allowed per section.
The third pitfall is the late-section RC collapse. A Reading Comprehension passage can appear anywhere in the Verbal section, which means a candidate who is on time at question 10 can still be in trouble by question 15 because one passage took three minutes longer than budget. The fix is to read RC passages at a constant pace — about 90 seconds for a 4-paragraph business passage — rather than reading the first paragraph slowly and the rest faster. A consistent reading speed prevents the late-section RC collapse that drives many Verbal timing failures.
The fourth pitfall, in Data Insights, is treating Data Sufficiency as a computation problem. DS items collapse under time pressure not because the math is hard, but because candidates re-derive both statements separately when one statement is sufficient. A 30-second glance at statement 1 is often enough to classify it as sufficient, insufficient, or indeterminate, and the same for statement 2. Most DS items can be classified in under 60 seconds without a single calculation, which is why top scorers routinely finish DS with time to spare.
The fifth pitfall is skipping the on-screen timer. The GMAT interface includes a persistent timer that most candidates ignore until it is in the red. A candidate who glances at the timer after every 3 to 4 questions, even if only for half a second, accumulates an accurate internal clock that prevents the front-loaded slow-down that triggers most timing collapses. The timer is not there to make you anxious. It is there to give you a data point on whether your pacing plan is working.
Rehearsing the bailout: training the triage moves before test day
No triage protocol works on test day if it has not been rehearsed. The reason most candidates fail to apply the moves above is that they encounter them for the first time in a high-stakes setting, and the cognitive load of choosing a tactic overrides the tactic itself. The fix is rehearsal during preparation, not in the final two weeks.
Drill 1: the 75-second cap on the first three questions
Take any set of 10 official-level Quant questions. Solve the first three with a strict 75-second cap on each. The moment the cap is reached, commit your best guess, bookmark the question and move on. Repeat this drill across three or four sittings. The goal is not to get the first three right; the goal is to internalise the feeling of letting go of a question at 75 seconds, which is the only way the move is available to you when the clock is genuinely short.
Drill 2: end-of-section triage
Take any 21-question Quant set from the official question bank and work through it in order, as on the real test. Aim to reach question 15 by minute 28. Whatever time is left for the last questions, switch to triage pace: a 20-second first look at each, solve it if it resolves fast, otherwise eliminate, guess and bookmark, so that every question has an answer when the clock runs out. The drill trains the end-of-section triage that most often rescues a section that has already slipped.
Drill 3: the review protocol
Take any Data Insights set and bookmark the questions you guessed on during the first pass. With the time left at the end, run a single review pass: spend about 30 seconds per bookmarked question, change an answer only if you can name the specific error, and stop at three changes, as on the real test. The drill trains the discipline of single-pass reconsideration, which is the difference between a salvaged section and a wasted last few minutes.
Conclusion and next steps
A timing collapse on the GMAT is recoverable in the moment, but only if the recovery moves have been rehearsed. The protocol above — diagnose the failure mode, apply a 90-second cap, guess and bookmark with a committed answer, switch to triage pace near the end of the section so that nothing is left unanswered, and rehearse the bailout during preparation — protects the score that is still on the table rather than chasing the score that is already lost.
TestPrep Istanbul's timing-diagnostic module is a natural starting point for candidates building a sharper section-by-section recovery plan for the GMAT.