Bain · Numerical reasoning · Updated 28 September 2026

Bain numerical reasoning: worked examples

Numerical reasoning is reported in both versions of Bain's online test, TestGorilla and SOVA. The math is school-level. What costs people marks is small stuff, like dividing by the wrong base or using an exchange rate backwards. Below are eleven original questions worked in full, each with the trap it's built around, then the shortcuts and pacing that matter on the day.

What to expect

Bain's careers page says its digital assessment runs on TestGorilla or SOVA, and the email after you apply tells you which. Prep sites and candidate reports put a numerical section in both. On TestGorilla it's the Numerical Reasoning test, which TestGorilla lists at 12 questions in 10 minutes. On SOVA it's one section of a longer assessment.

Expect a short table or chart with business figures, then a multiple-choice question that needs one to three steps. The wrong options are rarely random. Most are the answer you'd get from one specific mistake, which is why each example below names the trap.

Worked examples

All eleven are written by us. None is a real test item, and every company is invented.

1. Change in a total

Revenue by product line at Fernhollow Supplies, in $ millions:

Line20242025
Home120138
Garden6054
Tools2028

By what percentage did total revenue change from 2024 to 2025?

  1. 9.1%
  2. 10%
  3. 15%
  4. 20%

Total 2024: 120 + 60 + 20 = 200. Total 2025: 138 + 54 + 28 = 220. Change: 20 ÷ 200 = 10%. Answer: B.

The trap: the lines changed by +15%, −10% and +40%, and their average is 15% (C). Averaging percentages ignores that Home is six times the size of Tools. Option A divides by the 2025 total instead of 2024. Option D mistakes the $20 million change for a percentage.

2. A line chart: fastest month

A line chart shows monthly active users of the Tallow app, in thousands:

JanFebMarAprMayJun
8092104120126144

Between which two consecutive months was percentage growth highest?

  1. Jan to Feb
  2. Mar to Apr
  3. Apr to May
  4. May to Jun

Jan to Feb: 12 ÷ 80 = 15.0%. Mar to Apr: 16 ÷ 104 = 15.4%. Apr to May: 6 ÷ 120 = 5.0%. May to Jun: 18 ÷ 126 = 14.3%. Answer: B.

The trap: May to Jun is the steepest segment on the chart (+18) but starts from a bigger base. And Jan to Feb is close enough that a rough estimate can't separate them. When two options are within a point, calculate both exactly.

3. A bar chart plus a price

A bar chart shows units sold last quarter, in thousands. The prices are given below the chart.

ProductUnits (000s)Price per unit
Alpha45$12
Beta30$20
Gamma18$32

Which product earned the most revenue?

  1. Alpha
  2. Beta
  3. Gamma
  4. Alpha and Beta equally

Alpha: 45,000 × $12 = $540,000. Beta: 30,000 × $20 = $600,000. Gamma: 18,000 × $32 = $576,000. Answer: B.

The trap: the tallest bar is Alpha, and the chart only shows units. Any question about revenue needs the price, which is often in a note under the chart where it's easy to miss.

4. Percent versus percentage points

A lender's default rate fell from 2.5% to 2.0%. By what percentage did the default rate fall?

  1. 0.5%
  2. 20%
  3. 25%
  4. 80%

The fall is 0.5 percentage points. As a share of the starting rate: 0.5 ÷ 2.5 = 20%. Answer: B.

The trap: A gives the change in percentage points, not percent. C divides by the new rate (0.5 ÷ 2.0). D is the new rate as a share of the old one (2.0 ÷ 2.5), which is where the rate ended up, not how far it fell.

5. Ratios

Orvane Labs has 96 staff, all engineers or sales. The ratio of engineers to sales staff is 5:3. It hires 12 more engineers and no sales staff. What's the new ratio of engineers to sales staff?

  1. 5:3
  2. 17:3
  3. 2:1
  4. 3:1

5 + 3 = 8 parts, so one part is 96 ÷ 8 = 12. Engineers: 5 × 12 = 60. Sales: 3 × 12 = 36. After hiring: 72 engineers to 36 sales, which is 2:1. Answer: C.

The trap: B adds the 12 people to the ratio's 5, but ratio parts aren't head counts. Convert a ratio to real numbers before changing anything.

6. Compound growth

Brenlow Foods has revenue of $40 million and a forecast of 10% growth a year. What's the forecast revenue in three years?

  1. $44.0m
  2. $52.0m
  3. $53.2m
  4. $58.6m

Growth compounds: 40 × 1.1 × 1.1 × 1.1 = 40 × 1.331 = 53.24. Answer: C.

The trap: B adds 10% of the original $40m three times (simple growth), missing the growth on growth. A is one year. D is four years. "A year" in a forecast means compounding unless the question says otherwise.

7. Average annual growth

Kessel Rail's ticket sales grew from 50,000 to 72,000 over two years at a constant annual rate. What was the annual growth rate?

  1. 11%
  2. 20%
  3. 22%
  4. 44%

Total growth is 72 ÷ 50 = 1.44. Over two years at rate r, (1 + r)² = 1.44, so 1 + r = 1.2. Answer: B, 20%. Check: 50,000 × 1.2 = 60,000, then × 1.2 = 72,000.

The trap: C halves the total growth of 44%, which ignores compounding. D is the total, not the annual rate. With four options, you can often test each one forward instead of taking a root.

8. Currency conversion

The exchange rate is €1 = $1.08. A client's budget for a project is $54,000. How much is that in euros?

  1. €49,680
  2. €50,000
  3. €54,000
  4. €58,320

Each euro costs $1.08, so dollars ÷ 1.08 = euros: 54,000 ÷ 1.08 = 50,000. Answer: B.

The trap: D multiplies instead of dividing. A subtracts 8%, a shortcut that's close but not equal (it gives 54,000 × 0.92). The sense check: a euro is worth more than a dollar here, so you should get fewer euros than dollars. That rules out C and D straight away.

9. Weighted average

Average spend per transaction at three Quillon Bakery shops last week:

ShopTransactionsAverage spend
Quay Street1,200$30
Station300$50
Hilltop500$42

What was the average spend per transaction across all three shops?

  1. $36.00
  2. $38.50
  3. $40.67
  4. $42.00

Total spend: 1,200 × 30 + 300 × 50 + 500 × 42 = 36,000 + 15,000 + 21,000 = $72,000. Total transactions: 2,000. Average: 72,000 ÷ 2,000 = $36.00. Answer: A.

The trap: C is the plain average of the three shop averages. It treats Station's 300 transactions as equal to Quay Street's 1,200. The answer has to sit nearer the biggest shop's figure, which is a fast way to spot it.

10. Margins

Quarterly results for Pellwood Studio, in $ thousands:

ItemAmount
Revenue800
Cost of goods sold500
Operating expenses180

What was the operating margin?

  1. 15%
  2. 22.5%
  3. 24%
  4. 37.5%

Operating profit: 800 − 500 − 180 = 120. Operating margin: 120 ÷ 800 = 15%. Answer: A.

The trap: D is the gross margin (300 ÷ 800), which stops before operating expenses. C divides operating profit by cost instead of revenue. B is operating expenses as a share of revenue. Margins always divide by revenue. Markup divides by cost: the gross markup here is 300 ÷ 500 = 60%, which is a different number for the same business.

11. Break-even

A meal-kit service has fixed costs of $180,000 a month. Each box sells for $45 and costs $30 in food, packing and delivery. How many boxes a month does it need to sell to break even?

  1. 2,400
  2. 4,000
  3. 6,000
  4. 12,000

Each box contributes $45 − $30 = $15 toward fixed costs. Break-even: 180,000 ÷ 15 = 12,000 boxes. Answer: D. Check: 12,000 × $15 = $180,000.

The trap: B divides fixed costs by the price, as if every dollar of sales covered fixed costs. C divides by the variable cost. A adds price and cost together. Break-even always uses the contribution per unit.

The traps, in one list

  • Wrong base. Percentage change divides by the starting value. Questions 1 and 4 both offer the answer you get from dividing by the end value.
  • Averaging percentages. You can't average growth rates or averages across groups of different sizes. Go back to the totals.
  • Points versus percent. A rate moving from 2.5% to 2.0% is 0.5 points and a 20% fall.
  • Absolute versus relative. The biggest jump on a chart is not always the biggest percentage change.
  • Rate direction. Read "€1 = $1.08" as a price: one euro costs $1.08. Dollars to euros, divide. Euros to dollars, multiply.
  • Simple versus compound. Multi-year growth compounds. Halving a two-year total overstates the annual rate.
  • Margin versus markup. Margin is over revenue, markup over cost.
  • Missing information in a note. Prices, units ("in thousands") and time periods often sit under the chart.

Speed shortcuts

Estimate before you calculate. Spend three seconds deciding roughly where the answer lands. In question 9 the answer has to be between $30 and $50 and closer to $30, which already rules out C and D. In question 8 you need fewer euros than dollars. An estimate also catches calculator slips, which are the most common error once you know the method.

Know the common fractions. These come up constantly as percentages:

FractionPercentFractionPercent
1/333.3%1/812.5%
2/366.7%3/837.5%
1/616.7%5/862.5%
1/714.3%1/911.1%
1/128.3%1/205%

So 300 ÷ 800 is 3/8, which is 37.5%, with no calculator needed.

Round late, unless the options are far apart. If the options are 10%, 20%, 30% and 40%, round early and freely. If two options are within a point of each other, as in question 2, keep full figures until the end. Look at the options before deciding how precise to be.

Work backwards from the options. For growth rates and break-even, testing an option forward is often faster than solving algebra. 50,000 × 1.2 × 1.2 takes five seconds.

Calculator use

Bain's careers page says you can use pen, paper and a calculator on its digital assessment, and SOVA's candidate guidance also allows a calculator. Older guides that say Bain bans calculators are out of date.

Use a physical calculator you already know. Switching to a calculator app on the same screen costs time on every question and risks clicking away from the test. Practice with the same calculator you'll use on the day. Keep pen and paper for writing down the two or three numbers you need before you type anything; most mistakes come from typing a number from the wrong row.

Pacing

TestGorilla. TestGorilla's own page lists its numerical reasoning test at 12 questions in 10 minutes, which is 50 seconds a question. You can't go back to a question once you've moved on (Bain's Seoul FAQ and TestGorilla's candidate help both say so), and TestGorilla's help says wrong answers aren't penalized. If a question is taking more than about 80 seconds, guess and move on.

SOVA. Prep sites and candidate posts consistently report that SOVA shows no visible countdown but records how long you take, and that the time counts. SOVA hasn't confirmed this publicly. Treat it as timed: aim for about a minute a question and don't spend three minutes on one because no clock is showing.

Neither Bain nor the platforms publish a pass mark. A good practice score tells you your method works; it doesn't predict your result.

FAQ

Is there a numerical section in both versions of the Bain test?

Yes, as reported. Candidates and prep sites describe a numerical section in both the TestGorilla and SOVA versions. Bain doesn't publish the exact sections.

How hard is the math in Bain's numerical reasoning test?

School-level: percentages, ratios, averages, growth and unit conversions. The difficulty comes from time pressure and from wrong options designed around common mistakes.

Can I use a calculator on the Bain numerical test?

Yes. Bain's careers page says pen, paper and a calculator are allowed on its digital assessment, on both platforms.

How many questions are in the Bain TestGorilla numerical test?

TestGorilla lists its Numerical Reasoning test at 12 questions in 10 minutes. Candidates report that it appears in both reported Bain TestGorilla mixes.

What score do I need to pass?

Nobody outside Bain knows. Bain doesn't publish a pass mark, and its Seoul FAQ says results aren't shared with candidates. Aim to answer every question and get as many right as you can.

Keep going

Try the numerical section in Bain practice: the free set is 10 fixed questions, timed, with explanations. Then read the platform guide for your email: the Bain TestGorilla guide or the Bain SOVA guide. The next section to prepare is logical reasoning. Back to the Bain online assessment overview.

Sources we checked: Bain's digital assessment careers page, a Bain Seoul recruiting FAQ, TestGorilla's numerical reasoning test page and candidate help, SOVA's candidate preparation hub, and prep-site reports from IGotAnOffer, MConsultingPrep, MyConsultingCoach and PrepLounge candidate posts on SOVA timing.

SolveForge is an independent practice tool and is not affiliated with, endorsed by, or connected to Bain & Company, TestGorilla or SOVA. Every example on this page was written by us; none is a real test item. "Bain", "TestGorilla" and "SOVA" are trademarks of their owners. Practicing here does not guarantee any assessment result.