45 minutes, 20 questions: building a pacing protocol for…

Time management on the GMAT Focus Data Insights section is not a single skill but a layered protocol: a per-item time box, a per-pass triage plan, and a small set of rescue moves for the prompts that overrun. The section presents 20 questions inside 45 minutes, drawn from five item families — Data Sufficiency, Multi-Source Reasoning, Table Analysis, Graphics Interpretation, and Two-Part Analysis. Because the prompts range from a quick two-statement check to a 350-word integrated scenario, the average of 135 seconds per question is almost useless as a planning number. A working pacing protocol has to start with the section's structural facts and end with the moment a student decides to let an item go.

What the section actually gives you, minute by minute

The Data Insights section of the GMAT Focus is built around 20 items and a 45-minute clock. That produces the widely quoted 2 minutes 15 seconds per item, which I would actively discourage students from treating as a target. Two reasons. First, the five item families do not consume time symmetrically. A two-statement Data Sufficiency item can be dispatched in 60 to 90 seconds by a student who has the stem pattern catalogued, while an integrated Multi-Source Reasoning prompt with three tabs and a layered question routinely eats 3 to 4 minutes. Second, the test is adaptive, so the second batch of ten items is calibrated to the way you handled the first ten. Running out of clock at item 18 costs you the easiest two items on the test, not the hardest.

In practical terms I ask students to plan the section as a budget rather than a pace. The 45 minutes becomes 40 minutes of work plus a 5-minute safety margin, with the safety margin itself split into two reserves: a 2-minute late-section reserve for items 16 through 20, and a 3-minute end-of-section reserve that you only spend when the late reserve has been exhausted. That single reframe — "this is a budget, not a pace" — changes how the section feels, because it gives the student a place to put recovered time when a Data Sufficiency item falls in 70 seconds and a place to draw on when a Two-Part Analysis runs hot.

Why the average is a trap

The 2:15 average is a retrospective number, not a prescriptive one. It is what your watch will read if the section went well. It is also what your watch will read if the section went badly and you skipped the last three items. The two outcomes are not equivalent, but the average hides the difference. A pacing protocol has to behave well in both cases, and that means accepting that the last three items of the section are the most expensive items on the test in terms of scoring efficiency. A student who reaches item 18 with four minutes left and a partially built Table Analysis answer is, in my experience, a student who has spent those four minutes badly. The right move at that point is rarely to push the Table Analysis through. It is to triage item 18 as a guess, finish 19 and 20 with whatever time the rest of the section returns, and live with the cost.

The five item families, ranked by per-item time cost

Before a student can build a per-item time box, they need a realistic mental model of what each family costs. I rank the five families by typical per-item time on a prepared student, not on a novice, because pacing protocols are designed for prepared students, and novices have a different problem (they need to learn the families before they can pace them).

Item family	Typical time, prepared student	Dominant time sink	Pacing implication
Data Sufficiency	70 to 110 seconds	Re-reading the stem after the answer choices	Strict two-pass: classify, then evaluate
Multi-Source Reasoning	150 to 240 seconds	Switching cost between three tabs	Read question first, then build a fact list
Table Analysis	120 to 180 seconds	Off-by-one column reads	Mark column headers, then re-mark
Graphics Interpretation	120 to 180 seconds	Unit or axis confusion	Underline the units before any math
Two-Part Analysis	150 to 220 seconds	Forgetting the parts must be answered jointly	Treat as one question with two boxes

These ranges are not promises. A Two-Part Analysis on a clean algebra system can fall inside 110 seconds; a Data Sufficiency item with three nested conditions can run to 140 seconds. The point of the table is not to lock the student into a number per family but to break the assumption that all 20 items are equivalent units. They are not. A pacing protocol that spends 150 seconds on every Data Sufficiency item and 100 seconds on every Two-Part Analysis item is leaking time on roughly half the section.

A two-pass architecture for the whole section

The single most useful pacing structure I teach is a two-pass architecture. Pass 1 covers items 1 through 14 and runs to a hard clock of 32 minutes. Pass 2 covers items 15 through 20 and runs to the section's natural close at 45 minutes. The boundary at item 14 is not magical, but it is convenient: it leaves the last six items, which include the section's hardest material under the adaptive algorithm, in a second pass that the student enters with a clear picture of remaining time.

How to behave inside Pass 1

Pass 1 is where the budget is earned. Three rules govern it. First, no item in Pass 1 is allowed to consume more than 3 minutes of clock. If a student is still reading the third tab of a Multi-Source Reasoning prompt at the 2:40 mark, the item is marked and skipped. Second, every Data Sufficiency item in Pass 1 is treated as a two-pass item internally: a 30-second classification pass that identifies the stem type, followed by a 60 to 80-second evaluation pass. Third, the student writes down the wall-clock time at the end of item 5, item 10, and item 14. Those three timestamps are the section's pacing dashboard. If the item 10 timestamp is later than 20 minutes, the student is in deficit and needs to start skipping aggressively. If the item 14 timestamp is earlier than 32 minutes, the student has a surplus to deploy in Pass 2.

How to behave inside Pass 2

Pass 2 is where the budget is spent. The student enters Pass 2 with a known clock, a known position, and a known level of difficulty on the test. The two most common failure modes here are: pushing too hard on the first one or two items of Pass 2 and burning the late reserve; or, conversely, treating the last six items as a slow read because the clock feels comfortable. A useful rule of thumb: the items in Pass 2 are worth the same per-item time as Pass 1, but they are also the items most likely to overrun, so the student should pre-commit to a 4-minute per-item ceiling and a guess threshold at 4:30. Items that breach 4:30 in Pass 2 are guessed, marked, and moved past. The late reserve is reserved for genuine emergencies, not for ordinary overruns.

Per-item time boxes that actually fit the family

A per-item time box is only useful if the student has a believable internal model of what the family costs. Below is the working set I give to students at the start of a Data Insights cycle, calibrated for a candidate aiming at DI 78 or higher.

Data Sufficiency: 90 seconds. The 90-second box forces a clean two-pass behaviour: 25 to 30 seconds classifying the stem and skimming the statements, 50 to 60 seconds evaluating, 5 to 10 seconds locking an answer. Students who allow Data Sufficiency to run to 130 seconds are almost always re-reading the stem because the classification step was rushed.
Multi-Source Reasoning: 200 seconds. This is the family's ceiling, not its target. A clean MSR item with one tab of context and one tab of data falls in 130 to 160 seconds. The 200-second box exists so the student has somewhere to put the three-tab prompts that the adaptive section will eventually serve up.
Table Analysis: 150 seconds. The 150-second box buys one full read of the column headers, one column-marking pass, and one answer-evaluation pass. A student who finishes Table Analysis items in under 100 seconds is almost certainly misreading a column header; one who runs past 180 seconds is rewriting a header in their head that they should have marked the first time.
Graphics Interpretation: 150 seconds. Same logic as Table Analysis. The 150-second box pays for one axis-and-unit check, one read of the question stem with the answer choices covered, and one calculation. Graphics Interpretation overruns usually come from a unit confusion that the student discovers on the second pass, which is why the box has to allow a second pass.
Two-Part Analysis: 180 seconds. Two-Part is structurally a single item with two boxes, and the 180-second box is the only place in the section where the student is allowed to write a partial answer. If the joint answer is clear at 150 seconds, the student locks it. If the joint answer is still in contention at 180 seconds, the student is guessing, and a calibrated guess here is usually worth more than a fifth minute spent inside the prompt.

How the adaptive algorithm warps the section's middle

The GMAT Focus is section-adaptive, which means the second batch of ten items is calibrated to the way the first ten went. For the Data Insights section, this has a practical consequence that students regularly miss: the section does not get linearly harder, it gets locally harder. The student will not see a smooth ramp from easy to hard. They will see a run of medium items, a run of hard items, and a long tail at the end that includes both. A pacing protocol that assumes a smooth ramp will overspend on the run of hard items and arrive at item 17 with no late reserve.

Two tactical responses follow. First, the student should treat items 1 through 10 as a controlled build and items 11 through 20 as a controlled harvest. The 32-minute boundary at item 14 is where the build ends and the harvest begins. Second, the student should be willing to spend surplus time from Pass 1 in a targeted way: 30 to 60 extra seconds on a single Two-Part Analysis item in Pass 1 is a fine use of recovered time, but 30 extra seconds on each of the last six items is the pattern that empties the late reserve. The budget is a budget. Surplus from Pass 1 should be banked unless the student has identified a specific high-yield item to spend it on.

Common pitfalls and how to avoid them

The Data Insights section punishes a small set of repeated mistakes. These are the ones I see most often, in roughly the order of cost to the final score.

Reading the answer choices before the stem. The Data Sufficiency and Two-Part Analysis families are both vulnerable to this. Reading the choices first drags the student's attention into the answer space and away from the question being asked. Mitigation: physically cover the choices with a hand or a sticky note during the first read. Uncover only when the stem has been classified.
Re-reading column headers on Table Analysis. Table Analysis is the family where a single missed column costs the most points per second. The mitigation is a forced double-mark: mark the relevant columns once on the first read, then mark them again before reading the answer choices. The second mark is what catches the off-by-one error.
Forgetting the two boxes are linked on Two-Part Analysis. Two-Part Analysis items require the student to select two answers from a single list, and the selected pair is graded as a unit. A common failure mode is answering Part A in isolation and discovering in Part B that the constraint forces a different choice. Mitigation: read the full prompt, including both boxes, before selecting either answer.
Switching tabs without a question on MSR. Multi-Source Reasoning items often have a question that targets a specific tab, and a student who reads the tabs in order will spend 90 seconds on context the question does not need. Mitigation: read the question first, identify the tab or tabs it targets, then read only those tabs. The unrelated tab is read only if the question requires it.
Unit confusion on Graphics Interpretation. Graphics Interpretation prompts frequently use thousand-versus-million scaling, percentage-versus-proportion axes, and currency conversions inside the answer choices. Mitigation: write the units next to the answer choice before selecting it, and refuse to pick an answer whose units do not match the stem.

Building the protocol: a three-week timeline

A pacing protocol is not a list of rules, it is a habit. Habits take roughly three weeks of deliberate practice to install, which is why I structure the protocol as a 21-day build, not as a single reading of a list. The first week is item-family work, not pacing work. The student learns to classify Data Sufficiency stems in 25 seconds, to mark Table Analysis columns in 15 seconds, to identify the two boxes on a Two-Part Analysis in 20 seconds. Pacing is impossible without that fluency, because the per-item time boxes are filled with the time costs of the family's mechanics, and an unfluent student will simply overspend the box on the first move.

The second week is single-set timing. The student takes a 20-item Data Insights set under timed conditions and tracks per-item time by family. The output is a histogram, not a score. The student is looking for the families where the per-item time is more than 30 seconds above the target box, and the families where it is below. The first set rarely matches the boxes. The second set usually does, because the student has now seen the histogram.

The third week is the section simulation. The student takes three full 45-minute Data Insights sections under exam conditions, with the per-item time boxes enforced by an external timer, and a brief post-mortem after each. The post-mortem is the part that does the work: which items breached the box, why they breached it, and whether the breach was a budget failure (the student spent too long) or a mechanics failure (the student did not know the family well enough). The two failures have different remedies. The first is a pacing drill. The second is another week of item-family work, and the timeline has to absorb that.

What a good section actually looks like from the chair

The cleanest data point I can offer is the shape of a well-paced section, because the shape is more useful than any per-item time. A well-paced section, in my experience, has a Pass 1 timestamp at item 14 of 30 to 32 minutes, a Pass 2 timestamp at item 18 of 39 to 41 minutes, and a final timestamp at item 20 of 43 to 45 minutes. The student feels slightly pressed at item 12, comfortable at item 16, and slightly tight at item 20. That pressure pattern is the goal. A student who feels comfortable throughout is moving too slowly; one who feels pressed from item 6 onward is moving too fast and is almost certainly leaving points on the table from a mechanics failure.

The other useful data point is what the section looks like at the item level. A well-paced section has roughly 12 to 14 items inside their per-item time box, 4 to 6 items in the 30-second overrun range, and 1 to 2 items guessed because they breached the ceiling. The 1 to 2 guessed items are not a sign of failure. They are the section's natural skip population, and the protocol exists to keep them from becoming 4 to 5 items. The 12 to 14 inside-the-box items are where the section score is actually built. The overrun items are the ones where the student's mechanics and the time box are competing, and the protocol has to be ruthless about not letting those overruns compound.

Conclusion and next steps

A pacing protocol for the GMAT Focus Data Insights section is a budget architecture, a per-item time box, a per-pass triage rule, and a small set of mechanical habits for the five item families. The protocol is built over roughly three weeks of deliberate practice, and the only honest test of it is a section simulation that produces a 30-to-32 minute timestamp at item 14 and a tight but not panicked finish at item 20. The students who reach DI 78 or higher are, almost without exception, the students whose protocols survived a full section simulation without breaking. TestPrep Europe's diagnostic assessment is a natural starting point for candidates building a sharper per-item time box for the Data Insights section.

Frequently asked questions

How many minutes should I plan to spend on each GMAT Data Insights question?

Treat the section as a 45-minute budget with a 5-minute safety margin, not a 2:15 pace. Use a per-family time box — about 90 seconds for Data Sufficiency, 150 seconds for Table Analysis and Graphics Interpretation, 180 to 200 seconds for Multi-Source Reasoning and Two-Part Analysis — and enforce a hard 3-minute ceiling in Pass 1 and a 4-minute ceiling in Pass 2.

What is the best pacing strategy for the adaptive Data Insights section?

Use a two-pass architecture: items 1 to 14 inside 32 minutes, items 15 to 20 inside the remaining 13 minutes. Record the wall-clock time at items 5, 10, and 14 to monitor the budget, and pre-commit to guessing any item that breaches the per-family time box in Pass 2.

Which Data Insights item family takes the most time on test day?

Multi-Source Reasoning and Two-Part Analysis are the two most expensive families. A clean MSR item with three tabs typically takes 150 to 240 seconds, and a Two-Part Analysis with linked answer boxes typically takes 150 to 220 seconds. Data Sufficiency is the fastest family for a prepared student, often falling inside 70 to 110 seconds.

Should I skip a Data Insights question if I am running out of time?

Yes, but skip strategically. Mark the item, guess, and move on once the per-family time box is breached by more than 30 seconds in Pass 1 or by more than 60 seconds in Pass 2. The last three items of the section are the most expensive in scoring terms, so protecting clock for items 18 to 20 is more valuable than finishing a single overrun item.

How long does it take to build a working Data Insights pacing protocol?

Plan on three weeks of deliberate practice. Week 1 builds item-family fluency. Week 2 builds single-set timing by family. Week 3 runs full 45-minute section simulations with a forced post-mortem. A protocol is ready for test day when three consecutive simulations show a 30 to 32 minute timestamp at item 14 and a calm finish at item 20.

45 minutes, 20 questions: building a pacing protocol for GMAT Data Insights