How to read a GMAT Focus Table Analysis prompt in under 45…

GMAT Focus Table Analysis is the item family inside the Data Insights section that rewards a particular kind of literacy: the ability to look at a structured grid of numbers or labels, decide what the question is really asking, and operate on the table without rebuilding it from memory. The format is unfamiliar to many candidates because it sits between the verbal and quantitative halves of the test, but the underlying skill is older than the GMAT itself: disciplined table reading. The Data Insights section contains roughly twenty scored items drawn from five official families, and Table Analysis is one of the two pure-data-literacy families, the other being Graphic Interpretation. A candidate who treats the table as scenery will lose time; a candidate who treats it as a workspace will pick up the points that other test-takers leave behind.

The mechanics are short. A prompt introduces a table with three to five columns and a comparable number of rows, then asks a single multi-part question, usually a two-part or three-part multiple-choice stem, occasionally a single-select statement. The table is sortable, filterable, and clickable on the official platform, which is why preparation strategy for this family is largely about decisions, not about arithmetic. Most candidates reading this will already have the arithmetic; the gap is the decision layer above it. The article that follows breaks that layer into five archetypes, explains the pacing envelope, names the prompt words that change everything, and shows how scoring on Table Analysis interacts with the rest of the Data Insights section.

The five table archetypes you will actually see on test day

Most Table Analysis prompts are not as varied as they feel. In a working set of practice items, the same five shapes keep reappearing, and once you can name the shape you are halfway through the item. Naming is not a gimmick. It tells you which columns matter, which rows to ignore, and whether the answer lives in a single cell or in a comparison between rows. I would group the prompts into five working archetypes, and I would drill each one until the recognition is automatic.

Archetype 1: the maximum-or-minimum hunt

The prompt names a category, asks which row wins, and the table is small enough that you can read it whole. These items look easy and they are easy, but they are the ones that test discipline. Candidates who skim lose them to misaligned columns. The right move is to identify the column, scan the values, and lock the answer before clicking. Do not over-sort; the table is already short.

Archetype 2: the threshold filter

The prompt gives a numeric cut-off and asks how many rows meet it, or which row meets it, or what is true of the rows that meet it. The right tool here is the filter function, not the sort function. Sorting by the relevant column puts the rows in order but does not tell you which cross a threshold. Filtering does. Candidates who sort when they should filter waste ten to fifteen seconds per item, and over a stack of Table Analysis items the cost compounds.

Archetype 3: the rank-order question

The prompt asks for first, second, or third place under some constraint, and the trick is that the constraint forces a secondary sort. A candidate looking only at the primary column will pick the right row only by accident. The working method is to identify the secondary key, sort on it, and then read the ranking on the primary key. Rank-order questions are where a two-step mental model pays off, and they are the items where sorting earns its keep.

Archetype 4: the conditional change

The prompt says, in effect: if a value in the table were changed, what would happen to a downstream answer. These are reasoning items, not retrieval items. The right approach is to identify the dependency chain, simulate the change mentally on the row in question, and answer the conditional without rebuilding the table. Conditional-change items reward candidates who read the table as a model rather than as a picture.

Archetype 5: the comparison pair

The prompt names two rows and asks which is larger, smaller, or different, often on a derived metric that the table does not display. The right move is to find both rows, isolate the relevant columns, and do the small piece of arithmetic the question implies. The arithmetic is rarely harder than subtraction or a ratio, but it must be done on the two rows in question, not on the table as a whole.

Sorting versus filtering: the decision that controls your pacing

The single biggest decision in Table Analysis is whether to sort the table or to filter it, and the wrong choice is the most common reason candidates run out of time on Data Insights. The two tools look interchangeable, and on a small table they behave similarly, but on a busy twelve-row table they are very different operations. Sorting reorganises the table; filtering hides the rows that do not match. The prompt word decides which is correct, and the rule is simple enough to memorise: if the question asks about a single row or the order of rows, sort. If the question asks about which rows meet a condition, filter.

For most candidates, the temptation is to sort first and ask questions later. Sort feels productive. The rows are in a new order, the screen has changed, and the brain registers motion as progress. But sorting costs roughly four to six seconds, and if the prompt is a threshold question, the sort has not removed the rows that fail the threshold. The candidate then has to scan every row again. That is twenty seconds wasted on a fifty-second item, and the GMAT Focus scoring algorithm does not forgive the leak.

Filtering, by contrast, hides the rows that are not relevant, which means the surviving rows are exactly the rows the question is about. On a threshold item, a single filter leaves the candidate with a table that has shrunk from twelve rows to four. Reading the four is faster than reading the twelve, and the chance of selecting the wrong row drops to near zero. The cost of a filter is roughly three seconds, and it pays for itself the moment the prompt mentions a value, a range, or a condition.

There is a third option that the better candidates learn to use: no tool at all. On prompts that name a specific row by index or by label, the candidate should read the table once, identify the row, and answer. Sorting or filtering such a prompt is a form of procrastination. The table already has the answer; the tool only adds noise. Knowing when to do nothing is part of the preparation strategy for this family, and it is the move that separates the 645 scorer from the 755 scorer.

Practical pacing target: aim to spend no more than fifty seconds on a typical Table Analysis item, and no more than seventy on the hardest third. If you find yourself on an item past seventy seconds with no clear path, the right move is to mark the most defensible answer, flag the item, and return after the section. The Data Insights section does not penalise an unanswered item, but it does penalise candidates who let one prompt eat the time that two later prompts needed.

Reading the prompt word: the small vocabulary that changes the answer

Table Analysis prompts are short, usually one or two sentences, and almost every important word is load-bearing. The candidate who treats the prompt as prose loses items to a single word. The candidate who treats the prompt as a list of operations wins them. The vocabulary below is the working set I drill with students, and I would argue that mastering these ten words is more useful than any amount of practice with the table itself.

Most: a maximum question. Sort descending on the relevant column, read the top row, confirm the answer.
Least: a minimum question. Sort ascending, read the top row of the new order.
At least: a threshold filter. The answer is a count of rows that meet or exceed the value, not a single row.
At most: a threshold filter in the opposite direction. Again, a count, not a single row.
Which of the following is true: an inference question. The answer is a statement that holds for the table as a whole, not a value from a row.
If ... were ...: a conditional change. The candidate simulates the change, does not re-sort the table.
Greater than / less than: a comparison pair. Identify both rows, isolate the columns, do the arithmetic.
Approximately: a permission slip. The candidate does not need an exact value, only a defensible estimate, which means rough arithmetic is acceptable.
How many: a count. Filter rather than sort, and read the surviving row count, not the value of any single row.
Which row: a single-row retrieval. Read the table once, find the row, confirm against the prompt, answer.

The list is short on purpose. The Data Insights section is built on operational vocabulary, and Table Analysis is the most operational of the five families. A candidate who can name the operation in the first three seconds of the prompt is ahead of the candidate who is still parsing the sentence. The parsing happens anyway, but naming the operation makes it visible, and visible processes are faster processes.

Two-part and three-part prompts: how the stem changes the work

Most Table Analysis items are two-part or three-part multiple-choice stems, which means the candidate answers two or three sub-questions about the same table. The structure is efficient: one table, one read, two or three answers. The mistake candidates make is treating each sub-prompt as a separate item, which doubles or triples the table-reading cost. The right approach is to read the table once with the full prompt in mind, identify the columns and rows that the sub-prompts share, and only drill down on the parts of the table that differ between the sub-prompts.

Two-part prompts almost always share a row or a column. The first sub-prompt might ask for the row with the maximum value of column A, and the second might ask for the value of column B in the same row. The candidate who finds the row for sub-prompt one already has the answer to sub-prompt two. The table does not need to be re-read. The eye moves from the row in column A to the same row in column B, and the second answer is a single glance.

Three-part prompts add a third dimension. A common shape is: identify the row that meets condition A, confirm it meets condition B, and select the value of condition C in that row. The candidate should write the row identifier down, mentally, before answering. Writing the identifier prevents the third sub-prompt from drifting to a different row, which is the most common error on three-part items. The error rate on three-part prompts is higher than on two-part prompts not because the items are harder but because candidates lose the row identifier between sub-prompts.

For most candidates, the preparation strategy for multi-part prompts is to over-practice two-part items and under-practice three-part items. The inverse is more efficient. Two-part items are forgiving; the table is small and the answer keys are short. Three-part items are where the row-identifier discipline pays off, and the candidate who trains on three-part items first will find two-part items easy. This is one of the few places where the standard advice, drill the easier items, gets the ordering wrong.

The Data Insights section in context: how Table Analysis fits the scoring

The GMAT Focus Data Insights section draws items from five official families: Data Sufficiency, Multi-Source Reasoning, Table Analysis, Graphic Interpretation, and Two-Part Analysis. The section is unscored by family; what matters is the composite score on a 60-to-90 scale, and the candidate's percentile within the testing population. Table Analysis typically contributes a small number of items to the section, but the family is over-weighted in the preparation strategy of most candidates because it is the family with the most predictable shape. The items that look unfamiliar on test day are usually Table Analysis items, and the items that feel hardest on test day are usually Data Sufficiency.

Family	Typical item count	Skill emphasis	Preparation priority
Data Sufficiency	High	Reasoning about sufficiency	High
Multi-Source Reasoning	Medium	Cross-referencing two or three tabs	Medium
Table Analysis	Medium	Sorting, filtering, inference	High for discipline, low for content
Graphic Interpretation	Medium	Reading a chart with a variable toggle	Medium
Two-Part Analysis	Low	Two coordinated answers in one stem	Medium

The table is a useful snapshot. Notice that Table Analysis carries a high priority for discipline and a low priority for content. The reason is that the underlying arithmetic is rarely harder than subtraction or a ratio, and the underlying content is the kind of business data that a college-educated candidate has seen before. The discipline is the gap, and discipline is what the section is actually measuring. Candidates who prepare by reviewing statistics and probability for Table Analysis are preparing for the wrong family. The preparation is in the table-reading habits, not in the math.

Common pitfalls and how to avoid them on Table Analysis items

The same handful of errors shows up in almost every diagnostic I run. Naming them out loud is half the fix, and the other half is the small habit that prevents each one. The list below is ordered by frequency, not by importance, and every entry is something I have personally seen cost a candidate five to ten points on a real section.

Sorting when you should filter

The classic threshold error. The prompt says at least, the candidate sorts, the table reorders, the candidate scans every row again. The fix is one rule: if the answer is a count of rows, filter first, count second. Filtering on a numeric condition shrinks the table in three seconds, and the count is the number of surviving rows.

Losing the row identifier on a three-part prompt

The candidate finds the correct row for sub-prompt one, answers it, then answers sub-prompt two from a different row because the eye drifted. The fix is to hold the row identifier in working memory as a name, not a position. Use the row label, not its order in the table, and the drift disappears.

Reading the wrong column

The candidate finds the right row and reads a column adjacent to the one the prompt named. The fix is to underline the column mentally before reading the value. A candidate who does not name the column out loud, or in their head, is reading the table as a picture, and pictures are easy to misread.

Re-sorting after a conditional change

The prompt says if value X were 50 instead of 30, and the candidate resorts the table to find the new ranking. The fix is to recognise that the conditional is local. Change the value in the row in question, simulate the new ranking, and answer. The table is still the table; the candidate is testing a counterfactual.

Trusting the sort before confirming the prompt

The candidate sorts descending on a column, sees a row at the top, and selects it without re-reading the prompt. The row at the top is the answer to most, not the answer to the prompt. The fix is to read the prompt twice: once for the question, once for the operation. Two reads take four seconds; the saved time is on items the candidate would otherwise have flagged.

A working preparation strategy for the next four weeks

A four-week plan is enough to lift a candidate's Table Analysis score by ten to twenty percentile points, which is roughly the difference between a 645 and a 715 on the composite. The plan is short because the content is short, and the goal is to build the decision habits, not the content. Below is the working plan I assign, with week-by-week goals and the practice items that should sit behind each goal. The plan is calibrated for a candidate who is also preparing for the other Data Insights families, which is the usual case.

Week one is diagnostic. Take a full-length Data Insights section under timed conditions, score it by family, and identify which Table Analysis archetype produced the most lost points. The diagnostic is the only way to know where the gap is, and a candidate who skips the diagnostic is preparing generically, which is the single most common cause of plateaued scores. After the diagnostic, the candidate should be able to name, for every Table Analysis item in the section, which archetype it belonged to and which of the pitfalls above caused the error.

Week two is single-archetype drilling. Pick the archetype that produced the most errors in week one, and complete twenty items of that archetype under timed conditions. The candidate should track, for every item, whether the error was a recognition error (the wrong archetype was named) or an execution error (the right archetype was named but the work was wrong). Recognition errors are the ones that cost the most points, and they are the ones that the rest of the plan targets.

Week three is mixed-archetype drilling. Complete thirty Table Analysis items drawn from at least three archetypes, in random order, under timed conditions. The goal of week three is to keep the recognition habit sharp under switching cost. Most candidates find that their accuracy drops five to ten points when the archetypes are mixed, and the drop is the signal that the recognition habit is not yet automatic. Week three is where the recognition habit is converted from a deliberate process to a fast one.

Week four is full-section simulation. Take a full-length Data Insights section, score it, and compare the Table Analysis sub-score to the week-one diagnostic. The expected lift is ten to twenty percentile points, and the lift is the validation of the plan. Candidates who do not see the lift should return to week two on the weakest archetype and repeat. A four-week plan that does not produce a lift is not a four-week plan; it is a habit that needs another pass.

Putting the pieces together: a single item, walked through

The walkthrough below is a typical Table Analysis item from the working set I use with students, anonymised and simplified to make the decision steps visible. The table has five columns: Product, Region, Units Sold, Average Price, and Return Rate. The rows are ten products, each assigned to a region. The prompt is a two-part stem: which product has the highest revenue in the West region, and what is the return rate of that product. The first move is to name the operation. The prompt asks for the product with the highest revenue, which is a maximum question, but revenue is not a column. The candidate must compute revenue as Units Sold times Average Price, which means the prompt is actually a comparison pair dressed as a maximum. The candidate should filter on Region = West, then sort the surviving rows by Units Sold, then read Average Price for the top row, and finally multiply. The second sub-prompt is then a single read on the Return Rate column of the same row. Total time: under fifty seconds for a candidate who has drilled the archetype. The same item, attempted as a sort-then-scan, runs seventy to ninety seconds and produces a higher error rate on the second sub-prompt because the row identifier has been lost in the noise. The point of the walkthrough is not the arithmetic, which is trivial, but the recognition that the prompt word "highest revenue" hides a derived column. The candidate who sees the derived column is the candidate who scores well on Table Analysis; the candidate who sees only the visible columns is the candidate who scores 645.

This is the kind of item I would put in front of a student on day one of the plan, not because it is the hardest item in the family but because it is the most representative. It combines a threshold filter, a sort, a derived metric, and a multi-part stem, and the recognition of those four shapes is the recognition that the family is built on. A candidate who can see all four shapes in a single item is ready for the section; a candidate who sees one is not.

Conclusion and next steps

GMAT Focus Table Analysis is a family that rewards decision habits more than content, and the habits are trainable in four weeks with a focused plan. The five archetypes, the sort-versus-filter rule, the prompt-word vocabulary, and the multi-part discipline together form a working framework that holds across the items the section will throw at you. The composite scoring on Data Insights is unforgiving, but it is also predictable, and a candidate who lifts the Table Analysis sub-score by ten to twenty percentile points is also lifting the section score by a comparable amount. TestPrep Europe's diagnostic assessment is a natural starting point for candidates who want their Table Analysis baseline scored by family and by archetype, and the four-week plan above fits cleanly behind it.

Frequently asked questions

How many Table Analysis items are on the GMAT Focus Data Insights section?

The Data Insights section draws items from five official families, and Table Analysis typically contributes a small but consistent share of the total. The exact count varies by adaptive form, but a candidate should expect Table Analysis items to appear in roughly the first half of the section, alongside Data Sufficiency, and to require the same per-item pacing budget as the other families.

Should I sort or filter first on a Table Analysis prompt?

Filter first when the prompt asks for a count of rows that meet a condition, including any prompt using at least, at most, greater than, or less than as a cut-off. Sort first when the prompt asks for the row that wins on a column, including prompts using most, least, first, or last. When the prompt names a specific row, use no tool at all and read the table once.

What is the biggest error to avoid on multi-part Table Analysis items?

The most common error is losing the row identifier between the first and second sub-prompt. A candidate who finds the right row for the first sub-prompt and then answers the second sub-prompt from a different row is the textbook example of a multi-part error. The fix is to hold the row identifier as a label, not a position, and to read the prompt twice before selecting any answer.

How long should I spend on a single Table Analysis item?

Aim for fifty seconds on a typical item and no more than seventy seconds on the hardest third. Items past seventy seconds with no clear path should be flagged and returned to after the section, because the Data Insights section does not penalise an unanswered item but does penalise a candidate who lets one prompt eat the time that two later prompts needed.

Is Table Analysis a quantitative skill or a reasoning skill?

It is a reasoning skill built on quantitative scaffolding. The arithmetic is rarely harder than subtraction or a ratio, and the content is the kind of business data a college-educated candidate has seen before. The preparation is in the table-reading habits, not in the math, which is why the family is over-weighted in most preparation plans relative to its content load.

How to read a GMAT Focus Table Analysis prompt in under 45 seconds