Numerical reasoning
15 questions · 15 minutes total, with a hard stop every 3 minutes per data set
What it tests. Financial numeracy and data interpretation: multi-step percentages, compound growth, ratios, margins and rapid extraction under working-memory load.
Worked example. Given revenue and gross-profit-margin and operating-expense data for three years, calculate the percentage growth in net operating income (gross profit minus operating expenses) from the base year to the final year. The worked answer is roughly a 78% increase, and the option array rewards the closest approximation.
Common traps. Using the final year as the denominator instead of the base year; mixing thousands and millions or currencies; sinking 2.5 minutes into one question and leaving the rest of a set blank when it auto-advances.
How to handle it. Round aggressively, pre-format a gridded scratchpad, and abandon any question you cannot map a path to within 60 seconds.
Verbal reasoning
20 questions · 15 minutes for the whole block
What it tests. Comprehension and logical deduction: distinguishing a stated fact from a probabilistic inference from an unsupported generalization.
Worked example. A passage reports a body's empirical findings on dual-class share structures without stating a policy recommendation; a statement claiming the body recommends avoiding dual-class IPOs is therefore Cannot Say, because the text reports findings, not advocacy.
Common traps. Importing outside market knowledge; treating qualifiers like frequently or tend to as absolutes; assuming adjacency implies causation.
How to handle it. Read the statements before the passage, interpret literally, and default to Cannot Say when the text does not explicitly map the connection.
Logical / abstract reasoning
Roughly 12-15 matrices · 10-12 minutes
What it tests. Fluid intelligence, pattern recognition and spatial reasoning: isolating rules (rotation, progression, inversion, spatial overlap) and predicting the answer.
Common traps. Satisfying one rule (rotation) while ignoring a second (color inversion or line count); panicking at visual complexity that is just noise over basic operations.
How to handle it. Isolate one variable at a time and eliminate options that violate it; if no rule works across rows, test down columns.
Situational judgment
10-12 scenarios · Generous overall, but requires deliberation
What it tests. Professional judgment, ethics, prioritization and hierarchical awareness on a lean deal team.
Worked example. Late on a Friday before a Monday client pitch, you find a structural EBITDA discrepancy and the Associate is unreachable. The most effective action is to build a corrected shadow model alongside the baseline and email both to the Associate and VP, flagging the accounting risk and asking for guidance; the least effective is to halt work entirely until they respond.
Common traps. The unilateral maverick (fixing a major issue without telling the team); the passive passenger (escalating everything without a preliminary solution); the socially pleasant but low-impact choice.
How to handle it. Favor extreme ownership plus absolute transparency: build a solution, then escalate it for approval before any client exposure, and never bury an error in a footnote.
Behavioral & personality questionnaire
60-80 items · Untimed
What it tests. Conscientiousness, emotional stability, openness and agreeableness mapped against the firm's top performers.
Common traps. Selecting Strongly Agree or Strongly Disagree on everything (flags distortion); contradicting yourself across items; projecting a flawless robot the algorithm detects as inauthentic.
How to handle it. Anchor authentically toward conscientiousness, resilience and team-first competitiveness; keep internal consistency and reserve the extremes for core values.
Game-based modules (occasional)
~12 games · 2-3 minutes each
What it tests. Risk appetite, reward sensitivity, impulsivity, cognitive flexibility, working memory and sustained attention. The core US undergraduate track is anchored on Suited, so games appear mainly in specific global or offshoot pipelines.
Common traps. Over-correcting into paralysis after a failure; assuming the game rewards extreme risk aversion when a calculated appetite is valued.
How to handle it. Keep a steady operational rhythm, plan planning-game moves for ~5 seconds before acting, and use a real mouse with no background processes.