Boston Consulting Group · Psychometric Tests
Boston Consulting Group sifts candidates through BCG Online Case (Casey chatbot) before any interview. Below: the exact format, what each section tests, and how to practise it in identical conditions until you walk in confident.
Boston Consulting Group Pack
One-time payment. Yours for the whole application season.
Applying to several firms? The Season Pass is a yearly subscription covering every firm, £139 a year, cancel anytime.
The format
The BCG Online Case Assessment (the Casey chatbot) sits immediately after the CV and academic filter, a mandatory pre-interview gate before the one-way video or live rounds.
Most psychometric tests split into 3-5 sections, each with its own clock. Speed and accuracy both count.
Modern formats get harder if you answer correctly, easier if you struggle. Your final score reflects what you can actually do under time pressure.
Boston Consulting Group sets a pass mark per test type. Below it, you don't progress regardless of how strong the rest of your application is.
Completion window. Triggered automatically when the CV clears (a 2:1 prediction, strong A-Levels, target benchmarks); a strict 72-hour to 7-day completion window, enforced by the London team.
By division. A unified interactive business simulation across consulting tracks; BCG X technical candidates face additional automated programming or data-engineering assessments alongside the core case.
Recent changes. BCG replaced the old multiple-choice Potential Test with the conversational chatbot and added an integrated proctoring and anti-fraud suite (live webcam feed, tab-lockout, copy-paste and second-screen detection) to combat external AI tools.
The provider
Boston Consulting Group configures its own selection of BCG Online Case (Casey chatbot) modules. Below: the exact products in the suite, why they were chosen, and what the provider's reputation is across the candidate pool.
History at Boston Consulting Group. BCG's proprietary Casey online case replaced the legacy BCG Potential Test.
Candidate reputation. Built to measure practical, applied problem-solving in a realistic virtual simulation rather than abstract logic. The interface mimics modern workplace messaging and progresses through a linear, locked question tree: once you submit, you cannot return or revise, so adaptive time management is critical.
Section breakdown
Each section has its own format, timer and trap pattern. Worked examples below show exactly what you will see and where candidates drop points.
Opening questions of the case · About 2 minutes target per question
What it tests. MECE structuring: breaking an ambiguous problem into component parts without overlap or gaps.
Worked example. For a premium EV maker with a 12% profit decline despite 8% higher unit sales, the right framework dissects profit into Revenues (P x V) and Costs (FC + VC), not Porter's Five Forces or the 4Ps, because volume is up so the issue is price or costs.
Common traps. The over-engineering trap: choosing complex qualitative frameworks when the prompt needs a direct quantitative diagnostic.
How to handle it. Choose the option mapping closest to the fundamental economic drivers: revenues, costs, volume, capacity constraints or value-chain stages.
Embedded in the case · About 1.5 minutes target per question
What it tests. Analytical data prioritisation: filtering out background noise to focus on the core metrics.
Worked example. To verify whether a margin drop is driven by production inefficiency, select the Bill of Materials (raw battery costs) and the labour-hours and plant-overhead logs, not the customer demographics or competitor pricing tables.
Common traps. The data-collection trap: selecting every table for fear of missing information, which the assessment explicitly penalises as disorganised.
How to handle it. Before selecting, ask: 'does this spreadsheet directly help calculate a variable in my core diagnostic equation?' If not, leave it.
Several multi-step calculations · About 4 minutes target per question
What it tests. Multi-step business math and mental-math speed under time pressure.
Worked example. If battery packs are 40% of a £30,000 variable cost per unit and rise 15%, the new unit cost is £13,800 + £18,000 = £31,800, so a 30,000-unit run totals £954,000,000.
Common traps. The zero-check error: mismanaging zeros or decimals when scaling from a unit to a full production run (confusing millions with billions).
How to handle it. Write calculations cleanly on a physical scratchpad with units, and match the platform's explicit rounding and unit instructions before submitting.
Embedded in the case · About 2.5 minutes target per question
What it tests. Graphical synthesis and business acumen: extracting a clear narrative from multi-variable visual data.
Worked example. If market share falls 15% to 10% while unit discounts climb 2% to 12%, the conclusion is that discounting is failing to halt a structural share decline, pointing to deeper product competitiveness issues.
Common traps. The dual-axis trap: reading the wrong axis and drawing false conclusions about the trend.
How to handle it. Spend 10 seconds reading the legend, both axis units and any footnotes before analysing the trend.
Embedded in the case · About 2 minutes target per question
What it tests. Written synthesis and concise, executive-level communication.
Worked example. 'Consolidating regional distribution centres addresses the root cause by eliminating redundant lease costs, a guaranteed £14m overhead reduction, whereas marketing introduces demand uncertainty.'
Common traps. The rambling trap: a long paragraph repeating case facts instead of a direct, data-backed conclusion.
How to handle it. State the recommendation immediately, give one supporting metric and contrast it with the alternative.
1 pitch · 60 seconds prep, 60 seconds record, no pauses or re-records
What it tests. Executive presence, verbal structuring and communication synthesis under pressure.
Worked example. State the recommendation in the first 10 seconds, give two supporting data points, outline the key execution risks and close with next steps.
Common traps. Countdown panic: stumbling or losing structure while watching the 60-second timer.
How to handle it. Use the pyramid principle and rehearse a clean three-point structure on the notepad during the prep window rather than scripting it.
Pass mark
The assessment uses an algorithmic comparative percentile model against a target-cohort benchmark (Oxbridge, LSE, Imperial, top European schools), not a flat percentage pass mark.
Methodology. Performance combines across sections into a candidate profile. The case runs linearly with no skipping or revision, so an incorrect early calculation cascades into later related answers. Quantitative accuracy and logical structuring are weighted most heavily, so a major early error can drop the aggregate below threshold even with a polished video.
Response time. 7-14 business days as recruiters audit proctoring flags and verify video clips.
Score visibility. No raw scores or feedback reports are shared; a standardised system update confirms whether you advance.
How to practise
Same provider, same section structure, same time pressure. With a scored report after every test so you can fix the weakest section before the real one.
Free practice section, scored. Full report unlocks with the Pack.
Why candidates fail
Specific failure patterns documented across past applicants. Most are avoidable with structured preparation.
Mismanaging time on early sub-questions
Over-analysing the opening framework leaves under 60 seconds for the multi-step math at the end.
Cascading math errors
Because the case is linear and builds on prior results, one early error makes subsequent answers wrong too.
Missing format and unit instructions
Typing 14,500,000 or adding currency symbols when the prompt wants 14.5 (millions) flags as incorrect.
Overcomplicating rationale boxes
Long jargon-filled paragraphs lower the communication score and waste time.
Script-reading on the video pitch
Shifting eye-lines and robotic pacing flag the presentation as unnatural and unviable.
Proctoring flags
Second screens, open messaging apps or looking away trigger the anti-fraud system and a manual review.
What works
Concrete habits drilled by candidates who clear the cut-off, drawn from applicant accounts and practice patterns.
Fluent interface management
Complete the practice prompts to learn the text-entry boxes, pop-out exhibits and timers before the real test.
Methodical time budgeting
About 2 min structuring, 1.5 min data selection, 4 min math, 2.5 min graphs, 2 min rationale.
Clean, linear scratchpad layouts
Step-by-step formulas with units, easy to double-check before entering the final number.
Focus on core profit and operational drivers
Ground answers in volume, pricing, variable costs, capacity and capital efficiency.
Strong video synthesis pacing
A recommendation upfront, two validated data points, key risks and next steps, not every metric.
Avoid blind guessing
With a -1 penalty for wrong answers, eliminate options or leave blank rather than guess.
From past applicants
Anonymised candidate accounts of how recent applicants approached the Boston Consulting Group assessment. Each covers the prep, the experience, and the outcome.
Full-time Associate (London, advanced)
Prep. About 25 hours of practice.
Experience. The linear-progression rule meant certainty before submitting was essential. Spent a little too long on an early data-selection question and had to rush a percentage-change calculation later, but a clean scratchpad prevented a zero-counting error. Looked at the lens during the video pitch rather than the countdown, and read the rounding instructions carefully.
Outcome. Progression notification nine days later.
Summer Associate (London, advanced)
Prep. About 40 hours under strict countdown conditions.
Experience. The system tests data filtering by presenting charts with irrelevant axes. Used the 60-second video prep to outline a clean three-point structure rather than a full script, led with a clear recommendation backed by two data points, and hit stop with five seconds left to keep it crisp.
Outcome. Advanced to live interview rounds.
Practice strategy
Generic psychometric material is insufficient; practise interactive, chatbot-style cases tailored to the Casey chatbot interface.
Interactive chatbot simulations
Use CaseBasix and MConsultingPrep mock suites that replicate the Casey interface, the single-shot rule and the video window.
GMAT Data Sufficiency and Integrated Reasoning
Train fast data-sufficiency judgement and extracting insight from multi-axis charts.
Free practice on Intervyo
Run numerical and case practice in the real formats to build pacing under time pressure.
Time investment. Successful applicants dedicate 20-40 hours over 2-3 weeks, sharpening mental-math speed, graph interpretation and recorded mock video pitches.
Time management
FAQ
A desktop or laptop with a webcam and mic on Chrome or Safari; mobile is not advised as the data charts and pop-out exhibits are not optimised for small screens.
The other rounds
Psychometric Tests is one of four rounds. The Pack covers all four end to end.
Boston Consulting Group Pack
Practise the exact BCG Online Case (Casey chatbot) format ahead of time, scored against the Boston Consulting Group pass mark. One Pack covers HireVue, psychometric tests, live interviews and the assessment centre.
Get the Boston Consulting Group Pack · £69Intervyo is not affiliated with or endorsed by Boston Consulting Group, SHL, Pymetrics, Cubiks, AON or any other assessment provider. Test details are sourced from past applicants and published guidance; verify on the firm's careers site before applying. Sector: Strategy Consulting.
Boston Consulting Group Pack
£69 one-time