SHL Numerical Reasoning Test: Format, Practice, and Cutoffs
SHL Numerical Reasoning in 2026 ships in three versions: Verify Interactive (10 in 18 min), multiple-choice (16 in 20 min), and Verify Numerical Ability (18 in 25 min). Format, scoring, cutoffs, and a prep plan that matc
SHL Numerical Reasoning Test: Format, Practice, and Cutoffs
"SHL Numerical Reasoning" is not one test. In 2026 SHL still ships at least three live versions of the numerical battery, and an employer that says they are sending you "the SHL Numerical" might mean any of them. The versions share the question style, charts and tables with a calculation, but they differ on count, clock, and whether the test adapts under you. The cutoff you need also varies more by employer than by version. This page lays out the three formats, the topic mix you will actually see, the percentile bands employers tend to cut at, and a prep plan that matches the real pacing.
Quick takeaways
- Three live versions: Verify Interactive (10 questions in 18 minutes), multiple-choice (16 questions in 20 minutes), and Verify Numerical Ability (18 questions in 25 minutes).
- Verify Interactive is adaptive and mobile-enabled. The other two are fixed-form and almost always desktop.
- Every version is built around charts and tables. Percentage change, ratio, and rate questions are the bulk of the mix.
- A calculator is allowed on every version. Mental estimation still beats it on roughly 70 percent of items.
- Cutoffs run from the 55th percentile for tech graduate schemes to the 90th for bulge-bracket investment banks. Big Four graduate usually sits at the 60th to 70th.
- The pacing target is roughly 60 to 90 seconds per question. Burning two minutes on one chart is the single most common reason candidates miss the cutoff.
- The SHL Numerical score reports as a percentile or sten against a norm group, not as a raw mark.
What "SHL Numerical Reasoning" actually refers to
When SHL itself talks about numerical reasoning, it means one of two things. The first is the standalone numerical battery, which an employer adds when the role leans on numbers, finance and commercial roles being the obvious cases. The second is the numerical section embedded inside Verify G+, the adaptive General Ability test that blends numerical, inductive, and deductive reasoning into one combined score. For the full breakdown of how G+ works and how the numerical section sits inside it, the SHL Verify G+ General Ability overview is the pillar page; this article focuses on the standalone numerical assessments.
The standalone numerical comes in three live versions today: Verify Interactive Numerical Reasoning, the original multiple-choice Numerical Reasoning, and the Verify Numerical Ability test. The format you face depends on which version the employer licensed, not on the role title. A Big Four graduate applicant in 2026 is most often given the multiple-choice version; a mid-level commercial role at the same firm might get Verify Interactive instead. The high-level picture is in SHL Assessment: All the Tests Explained, and the worked sample lives in SHL Practice Tests: Free Walkthroughs by Section.
The three versions, side by side
The differences between the three versions are mainly about count and clock. They are all chart-and-table tests with calculator allowed. Verify Interactive adds drag-and-drop and on-screen entry; the other two are pure multiple choice.
The chart below lays out each version's question count, total time, average pacing budget, and where you are most likely to meet it.
Verify Interactive Numerical Reasoning
The newest of the three. Ten questions, 18 minutes, roughly 108 seconds per question. The test is adaptive and the response formats are interactive, meaning you drag data points, type values, or adapt a chart rather than picking a, b, c, d. SHL designed it to be harder to coach with a generic question bank. It is also mobile-enabled and used most often for mid-level commercial roles in finance, IT, professional services, production, and sales.
Multiple-choice Numerical Reasoning
The classic SHL numerical that most candidates have seen. Sixteen questions in 20 minutes, around 75 seconds per question. Five answer options per item, single correct answer. This version has been the dominant graduate-screen format through 2024 and 2025, and remains the most common version a Big Four or banking graduate applicant will face in 2026.
Verify Numerical Ability
A longer fixed-form battery. Eighteen questions in 25 minutes, around 83 seconds per question. Employers use it when they want a standalone numerical score that does not get blended into Verify G+'s combined cognitive score, typically for roles where numerical ability is the headline competency.
The numerical topic mix you will actually see
Across all three versions, the question types repeat with predictable frequency. Practising the right topic mix matters more than practising more questions.
Percentage change is the largest single category, somewhere around 28 percent of items. You get a table or chart with two values from different periods and an answer choice set that includes plausible distractors at plus or minus 1 percentage point. Ratio and proportion questions, including currency conversion, are another 22 percent. Rates and growth, including compound growth and per-unit rates, are around 18 percent. Chart-reading items, where the calculation is trivial but the chart selection is fiddly, are 18 percent. Multi-step word problems with a setup paragraph are the remaining 14 percent.
The implication for prep: drill percentage change and ratio first. They are 50 percent of every version, and they are the categories where a single misread cell costs you the full point.
| Topic family | Share of items | Common employers it shows up at |
|---|---|---|
| Percentage change | About 28 percent | Big Four, FMCG graduate schemes |
| Ratios and proportions | About 22 percent | Banks, consulting graduate streams |
| Rates, growth, currency | About 18 percent | Commercial banking, retail FMCG |
| Chart reading (line, bar, pie) | About 18 percent | All employers, every version |
| Multi-step word problems | About 14 percent | Consulting, strategy roles |
How the test is scored
SHL Numerical scores as a percentile or sten (a 1-to-10 standardised score) against a comparison group, usually a Fortune Global 500 or UK graduate norm. You do not get a raw count out of 10, 16, or 18. A percentile of 70 means you out-performed 70 percent of the norm group. A sten of 8 corresponds to roughly the 89th to 96th percentile. For cross-test benchmarking, what is a good cognitive test score lays the bands out alongside CCAT, Wonderlic, and Watson-Glaser.
On the Verify Interactive version, which is adaptive, the score also reflects the difficulty of the items you handled, not just the count correct. Two candidates who both answered eight of ten items correctly can finish at different percentiles because one of them was being shown harder items by the end. The practical consequence is that rushing the first few items, which sets your trajectory, is the worst pacing mistake on Interactive.
Cutoffs by employer
The single biggest variable is the employer. SHL does not publish "pass marks." Employers do, internally. The bands below are what we see in 2026 across graduate-scheme and commercial intake.
| Employer type | Typical numerical cutoff | Example employers |
|---|---|---|
| Big Four graduate | 60th to 70th percentile | Deloitte, PwC, EY, KPMG |
| Big Four consulting stream | 75th to 85th percentile | Deloitte Strategy & Operations, PwC Strategy& |
| Bulge-bracket investment bank | 80th to 90th percentile | JPMorgan, Citi, Barclays, Deutsche |
| FMCG graduate scheme | 60th to 70th percentile | Unilever, P&G, Mars |
| Magic Circle paralegal and trainee | 70th to 80th percentile | Linklaters, Clifford Chance |
| Tech graduate scheme | 55th to 70th percentile | IBM, Capgemini, Wipro |
| Boutique consulting | 70th to 80th percentile | LEK, OC&C, Oliver Wyman commercial |
| Public-sector graduate scheme | 50th to 60th percentile | NHS Graduate, Civil Service Fast Stream |
A useful rule: if you are aiming for a competitive graduate stream at any of these employers, plan around the 75th percentile, not the 60th. The 60th gets you past the basic filter but not into the shortlist for the strongest cohorts.
Two SHL-style sample items, walked through
These are not SHL's published items. They are written to match the format, difficulty, and pacing of what you will see in either the multiple-choice or Verify Interactive numerical sections, so you can rehearse the method.
Sample 1: Percentage change with a distractor band
A regional logistics firm reports parcel volumes for two consecutive quarters. Q3 2025: 1,420,000 parcels. Q4 2025: 1,632,300 parcels. By what percentage did volumes increase from Q3 to Q4?
Options: A) 13.5%, B) 14.0%, C) 14.9%, D) 15.0%, E) 15.3%
Method: change is 1,632,300 minus 1,420,000, which is 212,300. As a fraction of Q3, 212,300 divided by 1,420,000. Estimate first: 212 over 1420 is roughly 0.149. The calculator confirms 14.95 percent. The answer is C, 14.9 percent. The trap is that B (14.0) and D (15.0) are both within the rounding window of a sloppy mental estimate; C is the only option that survives a careful calculation. Time target: under 60 seconds.
Sample 2: Rate with a unit conversion
A printing press runs at 7,200 impressions per hour on a 22-minute setup cycle. Setup is unpaid output. The press runs three full hours of paid output after each setup. What is the effective paid impression rate per hour over a full setup-plus-run cycle?
Options: A) 5,860, B) 6,180, C) 6,275, D) 6,330, E) 6,420
Method: a full cycle is 22 minutes setup plus 180 minutes run, total 202 minutes. Paid impressions are 3 hours times 7,200, which is 21,600. Effective rate per hour is 21,600 divided by 202, times 60. That is 21,600 divided by 202 equals 106.93 impressions per minute; times 60 gives 6,416 impressions per hour. Round to the nearest answer, which is E, 6,420. The trap is that B (6,180) corresponds to ignoring the setup minutes entirely, and C (6,275) corresponds to spreading setup across only two run hours. Time target: under 90 seconds.
If you want a longer walked-through set of SHL items across all four reasoning sections, the SHL Practice Tests: Free Walkthroughs by Section page is the next stop.
A 10-day prep plan that matches SHL pacing
Most prep advice tells you to "do as many practice questions as possible." That is the wrong shape. SHL Numerical rewards a narrow skill set practised under realistic timing, and over-practising the wrong items burns time without moving your percentile.
Days 1 to 2: Diagnose
Take one untimed set of 16 mixed numerical items. Mark each one not just right or wrong, but by topic and by failure mode. The four failure modes are misread the chart, set up the calculation wrong, arithmetic slip, ran out of time. Most candidates' weak spot is one or two of those, not all four.
Days 3 to 5: Drill the weakest topic, untimed
Whatever was weakest in the diagnostic, drill it for 30 minutes a day untimed. The goal is to make the setup automatic before you add a clock. If percentage change keeps biting you, drill 15 percentage-change items a day until you can read the question, identify the two values, and write the formula without thinking.
Days 6 to 7: Add pacing
Switch to timed drills at 75 seconds per item for the multiple-choice format, or 108 seconds for Verify Interactive. Mark every question that takes you over the target. After each set, review only the items you missed or ran long on. Skip the ones you nailed.
Day 8: Full-length timed mock
Sit one full-length mock with the right question count and clock for your version. Do not pause it. Score it against the percentile table to see where you are sitting.
Day 9: Targeted clean-up
Revisit only the question types that hurt you on day 8. One more 30-minute drill on each, timed.
Day 10: Light review
Re-read your notes on calculation setups. Sleep early. Hydrate. Normal caffeine on test day, not extra.
Test-day execution rules
A handful of behaviours separate candidates who hit their cutoff from candidates who are five points short.
Read the question stem before you look at the chart. Charts are noisy. The question tells you which rows or bars matter. Reading the chart first wastes the first 10 to 15 seconds of every item.
Estimate before you calculate. A rough mental estimate within 5 percent of the answer narrows the option set to one or two, and lets you verify the calculator result instead of trusting it blindly. Calculator transposition errors are the second-most-common reason candidates fail SHL Numerical.
Move on at 90 seconds. If a question is not yielding at 90 seconds, take your best estimate from the remaining options and move on. The unread questions later in the test cost you more than a half-finished hard question costs you now.
On Verify Interactive specifically, slow down on the first three items even if they look easy. Wrong early answers cap your ceiling for the rest of the test.
Where SHL Numerical sits in the hiring funnel
For graduate schemes, SHL Numerical almost always sits at the second filter, after the application form and before any interview. It is sometimes the first filter for high-volume employers like Unilever or PwC's main graduate streams. For mid-level commercial roles, it usually appears after the initial screen call and before the technical interview.
The OPQ personality questionnaire often arrives in the same email link as the numerical. They are different tools and SHL treats them very differently; the numerical is a hard percentile gate, the OPQ is a fit profile read alongside the interview. For how the OPQ works and how it is read, the SHL OPQ Explained overview is the best starting point.
FAQ
Is the SHL Numerical Reasoning test the same as Verify G+?
No. Verify G+ is the adaptive General Ability test that blends numerical, inductive, and deductive reasoning into one combined score. SHL Numerical is the standalone numerical assessment. Some employers send both. The General Ability format is covered in the SHL Verify G+ overview.
Can I use a calculator on the SHL Numerical test?
Yes, on every live version. SHL allows a basic calculator, either on-screen or your own physical one. The skill is in setting up the right calculation and reading the right cells, not in mental arithmetic for its own sake.
How long is the SHL Numerical test?
Eighteen minutes for the Verify Interactive version (10 questions), 20 minutes for the multiple-choice version (16 questions), or 25 minutes for the Verify Numerical Ability version (18 questions). Pacing budget is roughly 60 to 90 seconds per question.
What is a good SHL Numerical score?
There is no universal pass mark. Big Four graduate intake typically cuts at the 60th to 70th percentile; bulge-bracket investment banks at the 80th to 90th. If you are targeting any competitive graduate stream, plan for the 75th percentile rather than the 60th.
How is the SHL Numerical test scored?
As a percentile or sten against a comparison group, not as a raw count out of the total. On Verify Interactive, your score also reflects the difficulty of the items you handled, because that version is adaptive.
Can you fail the SHL Numerical test?
You cannot fail it in the test-publisher sense; SHL just reports a percentile. You can fail the employer's cutoff for the role you applied to. That is the practical pass-or-fail, and it varies by employer.
Is the SHL Numerical test hard?
The difficulty is in pacing, not in the maths. The arithmetic itself rarely goes beyond percentages, ratios, and basic algebra. The hard part is doing it reliably in 75 to 90 seconds while reading the right cells from a noisy chart.
Can I retake the SHL Numerical if I miss the cutoff?
Some employers allow a retake after 6 to 12 months; many do not. SHL holds your previous results across employers that use the same norm group, so a retake at a different employer often pulls your prior score. Read the employer's policy before retaking blindly.
Does the SHL Numerical penalise wrong answers?
Not on the fixed-form versions. On Verify Interactive, wrong answers push the next item's difficulty down and cap the ceiling on your final score, which is a different kind of penalty but a real one.
Related on PrepClubs
- Pillar. SHL Verify G+ General Ability overview. The full pillar page with format, scoring, and prep paths for the adaptive cognitive test.
- Suite. SHL Assessment: All the Tests Explained. What every test in the SHL suite is and how each one is scored.
- Overview. The SHL Test in 2026: Verify G+, OPQ, Numerical, and Verbal. What "the SHL test" usually means and how each component is read.
- Practice. SHL Practice Tests: Free Walkthroughs by Section. One worked sample for each Verify reasoning section.
- Compare. CCAT vs SHL General Ability. Adaptive versus speed: how the two cognitive tests differ and which is harder for whom.
- Benchmarks. What is a good cognitive test score. Cross-test percentile benchmarks including SHL.
Practice on PrepClubs
Closest-match practice for SHL Numerical at real pacing.
PrepClubs does not run an SHL question bank, but the cognitive skills you need for SHL Numerical, fast chart reading, percentage and ratio under a clock, mental estimation against multiple-choice distractors, are exactly what the CCAT cluster drills. 1,350+ CCAT items, 14 timed mocks, 12 topical drills, with worked walkthroughs on every question. $39 one time. Pass Guarantee.
FAQ
