GMAT® Focus Data Insights — The Complete Guide for 2024 and Beyond

Welcome to the GMAT® Strategy Podcast. You're here because you believe there's a better way to study for the GMAT® and so do we.

We created the GMAT® Strategy to maximize your results and minimize your efforts so you can get to the fun parts about business school and life as quickly as possible.

My name is Isaac Puglia and I've been teaching GMAT® classes and tutoring privately for the GMAT® for almost a decade and I've achieved a 99th percentile score on the GMAT® and helped thousands of students get into the business schools of their choice.

I'm excited to be a part of your MBA journey since we all at TGS think our world can benefit from the best possible business leaders that we can find.

If this show is bringing you value, please share it with your friends and family who are studying so that together we can make this process as easy and as painless as it can possibly be.

Let's go!

Let's talk about the brand new Data Insights section of the brand new GMAT® Focus Edition.

This is something we've been getting a ton of questions about and we want to break everything down that you need to know.

Now, we debated about whether to turn this one episode into three episodes — one based on strategy, one based on integrated reasoning (which is one of the question types in DI), and one based on data sufficiency (which is the other question type in DI).

But we ultimately decided just to wrap it all into one episode, kind of like a mega episode. So there's going to be a lot of line items here.

Let us immediately address the elephant in the room and hopefully decrease your stress level on Data Insights quite a bit by letting you know that yes, it is a brand new section, but don't let that fool you. Because it's just a combination of two old GMAT® Classic problem types with one relatively minor change, and that's it.

So let's say you've studied for the old version of the GMAT® and you understand what integrated reasoning is. Integrated reasoning is comically the same as it was on the old test. In fact, there's still pretty much the same number of integrated reasoning questions.

They've just lengthened the section a little bit and added some data sufficiency in there. They took data sufficiency out of the quant section and just put it into Data Insights.

So if you've been stressed thinking like, "Oh, there's this new thing for me to master and learn," don't worry about that. We promise you it's going to be very familiar. The first Data Insights section you take, you're going to be like, "Oh, it's just this."

And if you're just starting out, there's no need to fear that the Data Insights section is "new." Yes, technically it is new, but it is only new in the sense that it combines two problem types that have been around for a very, very long time.

Section Format

So let's dive in. There's two question types in Data Insights. The first, like we mentioned, is integrated reasoning. The second is data sufficiency. There is no change to integrated reasoning in the new section. And we'll dive deep on everything you need to know about that in a moment.

There is one change to data sufficiency questions, which is that about half of them have to do with mathematics — which has always been the case. And the other half are now basically pure logic questions where there's very little or zero computation to do.

It's more like a real world scenario, usually. Such as: three friends go out for ice cream, Friend A gets chocolate and strawberry, Friend B gets neither chocolate nor strawberry, and Friend C gets chocolate only if the other two people both got chocolate — or something like that, where there are technically numbers in the question, but you don't need to set up an equation to answer the problem.

And all you need for those questions is the same data sufficiency process that you've always had. The only real change to this is that your scratch work habits are just more important than they used to be, because instead of missing a negative sign or something like that, you might miss one of the constraints of the problem.

So that's the only difference — you'll see some data sufficiency questions that are pure logic questions. And then you'll see old familiar data sufficiency questions that have numbers in them, exponent rules, standard similar content to what you would see in the quantitative section.

Quick Basics

What is the DI section? It's 20 questions in 45 minutes. So far, we've seen about 12 to 15 integrated reasoning questions on any given section. The old integrated reasoning section on GMAT® Classic was 12 questions every single time. So there's a little bit more variance here. You might see a little bit more IR on some days and a little bit less on other days, which leaves room for about 5 to 8 data sufficiency questions.

What that means is integrated reasoning is about twice as important as data sufficiency. Now, if you want a top score, you definitely need to be great at both.

Strategic Pointers

What should my goal be on the DI section? For most top schools, a 78 out of 90 or better is going to be acceptable. Below that, there's still a chance, but you should probably work on it. If you're applying outside of the top 20, 76 or higher is probably fine. If you're applying outside of the top 50, probably 73 or higher is fine.

Once you reach that 78 level, you're just looking for the section that you can stack the most points on the fastest. So if you're really great at quant and there's a lot of opportunity in the quant section and you're hitting 78 in DI but it's been a huge struggle to get there, you might want to invest 90% of your effort in moving the quant score because that's going to be the thing that drives the overall three-digit score.

Roughly: one to two sub-score points on each individual section will probably move your overall three-digit score by about 10 points.

Key #1 — The Data Insights section is adaptive. That's a bit of a departure from the old exam. On the old exam, integrated reasoning was not adaptive, but now it is. So it works exactly the same way as quant and verbal.

There was actually, unfortunately, a typo in the 2023 official guide that says it's not adaptive, and that has been corrected by GMAC. So it's definitely adaptive. All sections on the GMAT® are question adaptive now.

If you want more information about what that means and how to strategize on an adaptive exam, head to our website. There's a completely free video — it's 40 minutes long.

Key #2 — There is a calculator. You can show and hide it with a button usually on the upper left portion of the test screen. We recommend that you have it shown the whole section so that you don't forget to use it when it's valuable or when it's needed.

Most of us are conditioning ourselves super heavily on the quant section to not need a calculator, which makes sense because there's no calculator on the quant section. And so we tend to forget that there's a calculator on DI.

Our advice — which goes against a lot of people's advice out there — is to use the calculator as often as possible. It cuts down on computation errors and it also decreases the brain drain across the entire exam.

But many questions that are very computationally intensive are going to be better to just estimate. Here's a good barometer: if you're about to do more than 90 seconds of calculations on the calculator, you should probably just estimate the problem instead. If it's a quick 30 seconds or 10 seconds or 20 seconds of computations on the calculator, then for sure, just use the calculator.

Integrated Reasoning — The Four Question Types

This used to be its own section. GMAC was piloting the section for many years. We all suspected that it would become a lot more important — and lo and behold, it has now become very, very important.

Why does this matter? Integrated reasoning used to be a separate score and it didn't affect your overall three-digit score. But now that's not the case. So integrated reasoning has become just as important as quant and verbal, and that's because data literacy is really, really important these days. That's why business schools wanted the section introduced to the exam.

What is integrated reasoning? Well, let's start with the word "integrated." That means it's going to combine quant and verbal concepts. Now that might sound a little scary, but it's actually good news because all the studying you've done for quant and verbal will help out your integrated reasoning score.

That's why a lot of providers are going to recommend that you start with your quant and verbal studies first, be all in on those, and then transition to integrated reasoning and data insights later — because a lot of times increasing your quant and verbal knowledge and skills will drive the DI score.

There's no new content. But integrated reasoning is unique in that there are four new question formats. Once you master those four question formats and master the content from quant and verbal, you've mastered integrated reasoning. The only thing that could hold you back is timing.

Type 1: Two-Part Analysis

These are basically just algebra problems where you solve for two variables — hence the name. Usually these were problems of some type. It could be statistics like mean, median, mode, range — or it could be something very similar to a critical reasoning question in the verbal section.

The main strategy is just a structured approach to reading carefully. When we see people miss two-part analysis questions, it's usually not because they don't know the math. It's usually because they missed a keyword somewhere or a key concept.

Super simple structure: start by writing down what the question is asking you. Get crystal clear on that. Then go back and write down the information from the prompt that is related to answering that question.

One of the key skills of IR is sifting through lots of information to find the key data. In the quant section, pretty much 100% of the time if you're given something in a problem, it's crucial for solving the question. IR is pretty much the opposite of that. There's tons of extraneous information that you don't need, and so the core skill is finding the data you do need.

One note: be prepared to guess and test with the answer choices. They're giving you two variables and they'll sometimes ask which of the following could be values of X and Y. There isn't always one equation that you can just solve.

Type 2: Graphics Interpretation

If you've ever seen a graph before, you should pretty much be good to go. But one valuable strategic point: read the title and the axes of the graph very carefully. In fact, we recommend that you read those carefully before you even read the question.

And make note on your scratch pad if there are weird units on either of the axes — such as the X axis is in millions or in hundreds of millions. The X axis on the graph might just say 1, 2, 3, 4, 5, 6, 7, but then there's a key that says "X axis is in hundreds of millions." And that could be the difference between getting the exact answer and not.

You might see some weird graphs like bubble charts or strange things that might be a little bit unfamiliar. Try not to let that throw you off too much, but if you're staring at the graph and you really can't figure out what's going on, don't be afraid to just pick a random answer and move on. Remember, you can always come back and change your answer on up to three questions in every section on the GMAT® now.

Type 3: Table Analysis

These are basically just spreadsheets. That should be pretty darn familiar for most of you. You can sort the data based on each column — it's a rudimentary way to interact with the data, but it is helpful.

The key strategy with table analysis is very simple: you always want to sort the data based on what the question is asking you. If the question is asking about common price points, sort based on common price points. If the question is asking about revenue per customer, sort based on revenue per customer. If it's asking about profit per customer, sort based on profit per customer.

You can be very robotic with these questions. And as simple as that strategy is, it really, really helps unlock table analysis for most people — even people who struggle a lot with those.

Type 4: Multi-Source Reasoning

This is pretty much just reading comprehension from the verbal section with numbers. The format looks kind of like a web browser with tabs open — there's usually two or three tabs that you can click and view one at a time.

You'll cycle through usually two or three questions on that data set. So it's very similar to reading comp, where you'll see the passage on the left side of the screen and then you'll see three to four questions that you'll cycle through on that same passage.

The information itself is like tabs on a web browser, so you can only view one tab at a time. You're basically clicking back and forth between these tabs and answering questions based on how the data in individual tabs is related.

For a lot of you, this is going to be fairly similar to your day-to-day jobs — like you got an email in one tab and a spreadsheet in another tab, and you're writing an email to your staff or compiling data for a report.

The main strategy: treat it exactly how you would treat reading comprehension. Quick scan of all the tabs and all the info they give you, make some light notes if that's helpful, and then answer based on the evidence that they give you — not your opinion. Even if they use words like "infer" and "apply" or "suggest," they're just asking you: which one is 100% provable based on the data?

Extra IR Strategic Notes

Most questions have multiple parts — and there's no partial credit. In a two-part analysis, if you get variable A right and variable B wrong, the whole question is wrong. On multi-source reasoning, you'll often see yes/no questions — there'll be three of them within one question. If you got two out of the three right but the third one wrong, the whole question is wrong.

Before you let that stress you out, it actually can be reframed as a good thing. Because if you're a little bit unsure on a question, you should be more willing to give away those multi-part questions — because your odds of guessing correctly are lower than on a single-part question.

All GMAT® Classic resources for IR and DS are very helpful for the Data Insights section. Don't worry if you feel like there's not very many materials out there. Use integrated reasoning questions from GMAT® Classic, use IR sections from past GMAT® exams, use GMAT Club to find old integrated reasoning questions. All of it is 100% relevant.

Many people feel a lot of time pressure in the DI section, and this is usually because of integrated reasoning. Instead of 12 questions in 30 minutes, it feels more often like 30 questions in 30 minutes. Most timing issues in DI are going to be IR-related.

If you need to improve IR and especially your timing, this is actually one question type that does respond well to a lot of repetition. Unlike most GMAT® question types where we recommend fewer questions with deeper review, integrated reasoning is one where doing a lot of problems can be really, really helpful. Don't be afraid to do a lot of IR practice questions if you're struggling with timing or accuracy.

Data Sufficiency — Deep Dive

If you're feeling a bit overwhelmed at this point with all the info, feel free to pause and come back. Data sufficiency is a very technical problem type. It's important to have a deep understanding of DS because it's such an unfamiliar format.

Data sufficiency questions are unique to the GMAT® and the related exams such as the Executive Assessment. Problem solving questions — you just get an answer and you're done. Solved for X, X is seven. Seven is option B, bubble B, move on. Very familiar.

But data sufficiency is not that simple. It's more like a logic question that uses math content to test logical reasoning skills.

The format: There's always a question at the top of the screen. That'll be followed by two facts underneath that question — Fact 1 and Fact 2. The exam calls these "statements," but it's good to think about them as facts because they're always telling you the truth. You never have to question Statement 1 or Statement 2.

Your job is different than on a problem solving question. On data sufficiency, we're not just trying to get a numerical answer most of the time. Instead, we're trying to answer: how much data do I need to theoretically answer this question definitively?

The key here is: we're just asking how much data would I need to answer the question? Not what is the answer.

The Five Answer Choices (They Never Change)

• Option A: Statement 1 by itself is sufficient, but Statement 2 is not.
• Option B: Statement 2 by itself is sufficient, but Statement 1 is not.
• Option C: Statement 1 and 2 independently are not enough, but together they are enough.
• Option D: Statement 1 by itself is sufficient AND Statement 2 by itself is sufficient.
• Option E: Even combining both statements, there's still not enough data.

You're going to want to memorize those as soon as possible. Test yourself on those every single day until you just know it.

Value Questions vs. Yes/No Questions

Value questions ask for a numerical value — "What is the value of X?" or "How many minutes does it take to complete a certain task?" Sufficient data means you always get one singular value. If sometimes X could be six and other times X could be seven, that's not sufficient. It needs to always be six.

Yes/No questions are like "Is X equal to five?" There are only three possible outcomes: yes, no, and maybe (sometimes it's five and other times it's not).

On a yes/no question, you need a definitive answer — you either always get yes or always get no.

Here's where it gets brain-melting: if the question is "Is X equal to five?" and Statement 1 says "X is equal to seven" — that actually IS sufficient data. Because you know for sure the answer is no. It's never five. It's seven. Definitive answer.

But if Statement 1 says "X is larger than four" — that's not sufficient. It could be five, it could be six, could be seven. Sometimes yes, sometimes no.

And here's another brain-melter: if Statement 1 says "X is larger than six" — that IS sufficient. Because you know for sure X is not equal to five. Definitive answer: no.

If you're struggling with value vs. yes/no, just write "VALUE" or "YES/NO" at the beginning of each DS question on your scratch pad. Getting those mixed up is very, very common.

The Six-Step Process

Step 1 — Write what's given and asked. The most important step. Physically write down the given information and the question on your scratch pad 100% of the time. When there's a constraint given such as "X cannot be equal to zero," that constraint is always given for a reason — really want to note that. And if there's no constraint given on X, then mentally note that X could be negative, a fraction, positive, or a decimal.

Step 2 — Can you simplify the question? If it's a complicated question, ask if you can do some algebra to make it simpler. Probably 20–50% of DS questions can be simplified, and it's really helpful when you can.

Step 3 — Evaluate Statement 1 alone, in a vacuum. Do not consider Statement 2. Respect that Answer Choice A could be the answer. Is this enough data by itself to answer the question?

If Statement 1 is not sufficient → eliminate A and D. If Statement 1 is sufficient → eliminate B, C, and E.

Step 4 — Forget Statement 1. Evaluate Statement 2 alone, in a vacuum. This might be the hardest part. Delete Statement 1 from your consciousness. Scratch it off on your scrap paper to help you mentally delete it.

If Statement 2 alone is sufficient and Statement 1 was not → pick B. If Statement 2 alone is sufficient and Statement 1 also was → pick D.

Step 5 — If neither is sufficient alone, combine them. You've already determined that 1 by itself is not enough and 2 by itself is not enough. Now you're allowed to put them together.

If combined they're sufficient → pick C. If combined they're still not enough → pick E.

Step 6 — Pick your answer. That's it. That's your full data sufficiency process.

Write Everything Down

You should always write out all of your work on data sufficiency questions. If it's happening in your head on a DS question, it should be happening on the page.

This is very controversial. Many professionals and other students will tell you that you don't need to do that. Our opinion and our experience is that not writing everything down is like telling someone they don't need to obey the law — it's extremely risky.

The whole point of a data sufficiency question is to try to trick you — to make it seem like something that is sufficient isn't, or something that isn't sufficient is. Writing out your scratch work does two things: it helps you track where you are logically in the problem, and it frees up more brain power to process and problem-solve.

If you're just starting out, let yourself be super slow on data sufficiency questions. Don't feel the pressure to finish them in two minutes or even five minutes. Just focus on great scratch paper techniques and stepping through the process. That's going to solve 90–95% of the issues you're having on data sufficiency. The other 5% is probably just content or doing enough practice questions to build speed.

If you've been struggling with data sufficiency and you feel like it might be because your process or your scratch work is not very good, then take the time pressure out of the situation. Practice as many as you need to untimed to build up your scratch paper technique and thoroughness and build up your systematic approach to DS. And then at that point, you should be good to go.

With enough time and repetition, you can get fast and great at data sufficiency questions. The risk you take if you don't develop these great habits from the beginning is you get fast and bad at data sufficiency questions — which is better than slow and bad, but not by much. Being fast and bad at DS is not going to be great if you want to go to a top MBA program.

As always, our greatest hope is that this material will make your studies as easy and as painless as possible.

If you want more tips and strategies for optimizing your performance on the GMAT®, head to our website — theGMATStrategy.com — and check out our free video presentation on how to achieve your dream GMAT® score in half the normal time.

In the meantime, this is a weekly show, so please subscribe and stay positive and stay consistent with your studies. Reach out to us if there's anything we can do to support. Talk to you all soon.