Numerical reasoning (SHL Interactive)
18 to 20 questions · 25 minutes (about 75 seconds per question)
What it tests. Advanced data interpretation, percentage shifts, currency conversion, ratio analysis and isolating relevant variables in noisy datasets.
Worked example. Given London Q1 revenue of £12.5m growing 4% quarter-on-quarter, and Paris Q1 of EUR15.0m at an exchange rate of £1 = EUR1.20 growing 6%, calculate the combined Q2 revenue in GBP. London Q2 is £12.5m times 1.04 = £13.0m. Paris Q1 converts to £12.5m (EUR15.0m divided by 1.20), then grows to £13.25m in Q2. The combined Q2 total is £26.25m.
Common traps. Failing to check units or exchange rates on chart axes; the test often shows data in millions but asks for thousands, or mixes EUR, USD and GBP across columns.
How to handle it. Set up a physical scratchpad in grid squares before starting so a single miscalculation does not force a full restart. Use a familiar dual-powered handheld calculator, not the on-screen one.
Verbal reasoning (SHL Verify)
24 to 30 questions · 19 minutes (about 40 seconds per question)
What it tests. Text analysis, deductive inference and separating absolute facts from unjustifiable assumptions based solely on the passage.
Worked example. From a passage stating that critics argue UK public-sector contracts yield lower margins than private FTSE 100 engagements, the statement 'Private sector engagements are always more profitable for Strategy&' is Cannot Say, because the text only reports that critics argue this, not that it is a confirmed fact.
Common traps. Injecting external business knowledge. If a statement is not explicitly proven within the text block, choosing True or False on real-world knowledge is penalised.
How to handle it. Read the statement before scanning the passage so you hunt for specific keywords rather than passively reading first, which wastes precious seconds.
Logical / inductive reasoning
12 to 15 diagrammatic patterns · 25 minutes (roughly 60-100 seconds per question, scaling in difficulty)
What it tests. Pattern recognition, spatial awareness, abstract non-verbal logic and identifying complex rules across grid layouts.
Worked example. Apply an M-N-O-P scan: Movement (are elements shifting clockwise or reflecting), Number (are lines, dots or intersections increasing or decreasing), Orientation (are shapes rotating by 45, 90 or 180 degrees) and Properties (is shading or hatching alternating).
Common traps. Fixating on one variable, such as the position of a triangle, while ignoring secondary shifts in colour shading or line weight.
How to handle it. Decompose each pattern into layers, tracking background grid shifts separately from foreground shape rotations, rather than viewing it as a single unit.
Situational judgement (PwC Professional)
15 to 20 scenarios · No strict limit, but response times are tracked
What it tests. Alignment with the five PwC Professional pillars: Whole Leadership, Business Acumen, Technical and Digital, Global and Inclusive, and Relationships.
Worked example. On a CDD project you find two conflicting market-sizing sources (12% versus 2% CAGR) and your manager is unreachable in a client workshop. The strongest option builds both into a sensitivity analysis, documents the source methodology and presents the structural option to your lead on their return, keeping the project moving while protecting analytical integrity.
Common traps. Choosing the path of least resistance, or overly submissive answers that always escalate to a manager without proposing an active solution.
How to handle it. Project proactivity, data integrity and respect for team dynamics. Do not cut ethical corners for a deadline, and do not make independent assumptions on critical client metrics without flagging and testing them.
Game-based assessments (Arctic Shores)
9 to 11 games · 30 to 45 minutes total
What it tests. Risk tolerance, working memory, processing speed, resilience after negative feedback and cognitive agility.
Worked example. Examples include a balloon-pumping game measuring risk versus reward, a safe-cracker memory game under escalating speed, and a face-based emotion-identification task assessing emotional intelligence.
Common traps. Trying to game the system by acting a persona, such as over-inflating balloons on the assumption a strategy firm wants hyper-aggressive risk-takers, which just bursts the balloon and loses capital.
How to handle it. Be consistent and focused; the backend tracks behavioural shifts as much as raw scores. If you swing from cautious to reckless after a failure, the system reads a lack of emotional resilience. Be well-rested.