Assessments / MBB Comparison

Three firms, four formats, one confusing week

Bain Assessment vs McKinsey Solve vs BCG Casey

Q: Can you retake the MBB online assessments?

Not within the same application cycle at any of the three. Reapplication waiting periods are set locally and vary by office.

All three MBB firms screen with an online assessment before interviews, and they are not the same kind of test. One is a psychometric battery, one is a game, one is a case in a chat window. What that means for how you prepare.

Updated September 2026 13 min read Candidate reported formats, sourced below

4 formatsAcross the three firms

25-85 minRange of sitting lengths

1 attemptPer application, all three

0 publishedPassing scores

Overview

They are not the same kind of test

Candidates applying to all three firms usually assume the online assessments are variations on a theme, and prepare for them as one block of work. They are not variations on a theme. The three firms have made genuinely different bets about what predicts a good consultant, and the tests reflect that.

Bain runs a psychometric battery. The Bain SOVA test and the Bain TestGorilla assessment are multiple choice reasoning blocks wrapped around a behavioral questionnaire. No case, no business narrative to carry, nothing to structure.
McKinsey runs a game. McKinsey Solve puts you in an ecosystem or a research scenario and scores both your answers and the way you worked to reach them.
BCG runs a case. Casey is a business case delivered through a chat interface, ending with a recorded video recommendation. It is the closest of the three to an actual interview.

The practical consequence: case practice moves your BCG Casey result, does very little for Bain, and helps McKinsey Solve only indirectly. Treating the three as one prep block is the single most common way candidates waste the fortnight they have.

Which assessment you receive also varies by office and role within each firm, and none of the three publishes its specifications. Your invitation email names the test. That one line is worth more than any guide, including this one.

Comparison

Bain, McKinsey and BCG side by side

	Bain	McKinsey	BCG
Name	SOVA test, or TestGorilla assessment	Solve	Casey, also called the Online Case Experience
Format	Multiple choice reasoning blocks plus a work style questionnaire	Scenario based games with data collection and analysis	A single business case through a chat interface
Length	About 60 to 75 minutes on SOVA, 40 to 50 on TestGorilla	65 minutes for two games, 85 minutes if a third is included	About 25 to 35 minutes, plus a 60 second video
Clock	Hidden on SOVA but recorded, hard countdown per block on TestGorilla	One shared clock per game, no pauses	Timed throughout, with periodic reminders, and no pausing
Going back	No, once a section or item is submitted	Within a game until you confirm, then locked	No, and you must answer to proceed
Case skills needed	None	Indirectly, through data handling	Directly, this is a case
Video component	Only where a separate HireVue is assigned	No	Yes, a recorded recommendation at the end
Scored on	Accuracy and speed, percentile ranked on SOVA	Answers and process, scored separately	Case competencies, the way an interviewer would score them
Published cutoff	None	None	None

Formats are candidate reported and change between cycles and offices. None of the three firms publishes official specifications, and BCG in particular runs several assessments that candidates routinely confuse with each other.

Interactive

What each one is actually like

Switch between the three to see the structure, the timing and the thing candidates most often get wrong about each.

MBB assessment explorer2026 reported formats

Bain: SOVA or TestGorilla

A psychometric battery rather than a case. Bain runs two platforms in parallel, assigned by office and sometimes by candidate, and you find out which when the link arrives.

SOVAFive sections, roughly 60 to 75 minutes, no visible clock

TestGorillaThree or four blocks, 40 to 50 minutes, hard countdowns

Common groundNumerical reasoning appears in every reported version

What it measures: numerical, verbal and logical reasoning, situational judgment, and behavioral fit through a forced choice work style questionnaire. Nothing asks you to structure a problem or carry a business narrative.

What candidates get wrong: assuming case preparation covers it. It does not. The reasoning sections are closer to the GMAT than to a case, and the work style block runs a consistency check that punishes an invented profile.

Detail in the SOVA guide, the TestGorilla guide, and ten worked numerical questions.

McKinsey: Solve

A game based assessment. You are dropped into a scenario, handed data and a clock, and scored on both what you conclude and how you worked to get there.

Redrock Study35 minutes of data collection, analysis and reporting, then six mini cases

Sea WolfA 30 minute optimization game built on averaging

Two versions65 minutes for two games, 85 if the Sustainable Futures Lab is included

What it measures: critical thinking, data handling under time pressure, and decision making. Uniquely among the three, McKinsey records a process score from how you work, not only a product score from your answers.

What candidates get wrong: collecting every available data point rather than only what the research objective asks for. It costs minutes and it leaves a trace in the process score.

Detail in the Solve overview, the Redrock guide and the Sea Wolf guide.

BCG: Casey

A business case delivered by a chatbot. One scenario, a fixed sequence of questions, and a recorded video recommendation to close. The closest of the three to a live interview.

Chat interfaceRoughly 6 to 10 sequential questions on one case

25 to 35 minTimed throughout, no pausing, no skipping

Video closeAbout 60 seconds of recorded recommendation

What it measures: the same competencies a BCG interviewer would score. Structuring, exhibit reading, calculation, business judgment, idea generation and communication, delivered through a screen instead of a person.

What candidates get wrong: preparing for the wrong BCG test. BCG runs several assessments, including the Consulting Career Assessment, a cognitive test and, in some offices, Pymetrics games. If the invitation mentions a case, it is Casey. If it does not, it is one of the others.

Case Prep does not currently publish a BCG Casey simulator. The case library and math drills are the closest preparation we offer for it.

From Case Prep · Assessment Simulators

Two of the three, rebuilt to the real rules

Both Bain formats and all three McKinsey Solve games, with the constraints that actually make them hard: the hidden clock, the hard countdowns, the one way doors. Free sections in each.

assessment-simulators Elapsed 12:04

SituationalDone

VerbalDone

NumericalLive

LogicalLocked

SOVAFive section mock

TestGorillaFour timed blocks

FreeFirst section

Redrock · Investigation. Collect only what the research objective needs. Everything you take is recorded.

1Y1 population, 1,240 recorded in the first survey year

2Decline rate, 18 percent between Year 1 and Year 5

3Range area, 3.2 square kilometers

Research journal Process score recorded Calculator locked until Analysis

76thPercentile

Numerical84%

Logical61%

Data handling79%

Feedback no firm gives you Every item explained

Bain simulatorsSOVA full mock and the TestGorilla sprint, both under the real constraints.

McKinsey SolveRedrock, Sea Wolf and the Sustainable Futures Lab, with the journal and the calculator.

Scored either wayAccuracy, pace and full reasoning on every item, which no firm reports back to you.

Open the Bain simulators McKinsey Solve simulators →

Scoring

What each firm is actually measuring

None of the three publishes a passing score, and none reports your result back to you. What differs is the underlying model, and that changes what good preparation looks like.

Bain: relative ranking on reasoning and fit

On SOVA the cognitive sections are percentile ranked against a norm group of other applicants, so the effective bar moves with your cohort rather than sitting at a fixed mark. Accuracy and recorded time both feed it. The work style block is scored on alignment against a trait profile, with a consistency check running underneath. On TestGorilla there is no percentile reported at all, and section results are read alongside your CV.

McKinsey: product and process, scored separately

Solve is the only one of the three that explicitly scores how you worked. The product score comes from your answers. The process score comes from what you collected, how you navigated, and how you spent time. This is why interface familiarity is a genuine scoring advantage on Solve in a way it is not elsewhere, and why practicing inside a faithful recreation matters more for McKinsey than for the other two.

BCG: case competencies, through a screen

Casey scores what a live BCG interviewer would score, which is why it is the one assessment where case preparation transfers directly. Structuring, exhibit reading, math, judgment, idea generation and communication. The video close adds a communication test the other two do not have.

The one thing all three share: a single attempt per application cycle, no score released, and no retake. Whatever you plan, plan for one sitting.

Difficulty

Which one is hardest?

The honest answer is that it depends on what you are already good at, which is unsatisfying but true. A more useful framing is which one is hardest to prepare for from a standing start.

difficulty-by-dimension.fig

BCG CaseyHardest ceiling, but case prep transfers directly

McKinsey SolveUnfamiliar interface, and process is scored too

Bain TestGorillaEasy content, punishing clock

Bain SOVALongest sitting, gentlest per item

A judgment rather than a measurement. Bar widths reflect how much unfamiliar ground a typical candidate has to cover, not a difficulty score any firm publishes.

Casey has the highest ceiling because it asks you to actually do consulting work, including generating ideas and delivering a recommendation on camera. But it is also the most transferable: every hour of case practice you were going to do anyway feeds it.

Solve is the hardest to walk into cold because nothing in normal life resembles it, and because the process score means the interface itself is part of the test. Candidates who have never seen the research journal lose time learning it while the clock runs.

The Bain tests are the most straightforward in content and the most unforgiving in pacing. The math is simpler than anything Casey will ask, but TestGorilla gives you about 50 seconds per item with no way back, and SOVA quietly records your time while showing you no clock at all.

Strategy

Preparing for all three at once

Most MBB candidates apply to all three firms in the same cycle, and assessments can land within days of each other. The work does overlap, just not evenly.

What you practice	Bain	McKinsey	BCG
Fast chart and table math	Core of the numerical section	Most of Redrock and the mini cases	Several Casey questions
Reading an exhibit against an objective	Useful	The whole Investigation phase	Directly tested
Case structuring	Not tested	Indirect	Directly tested
Abstract logical reasoning	Tested on SOVA, the least practiced section	Not tested	Not tested
Situational judgment	Tested on both Bain formats	Tested in the Sustainable Futures Lab	Embedded in case judgment
Speaking a recommendation	Only if a video is assigned	Not tested	The final 60 seconds

Read down the columns and the sequence writes itself.

Start with fast math
It is the only thing that pays off in all three, and it is the most mechanical to improve. Twenty minutes a day with a stopwatch does more than any other single activity.
Then interface familiarity for whichever lands first
Solve and the Bain tests both punish a cold start, and the fix is sitting a faithful mock rather than reading about the format.
Keep case practice on its own track
It serves Casey and every live interview at all three firms. It does almost nothing for Bain, so do not let it crowd out the reasoning drills in the week before a Bain sitting.
Give logical reasoning one dedicated block
It only matters for Bain SOVA, but it is the section fewest candidates have ever practiced, which makes it the cheapest percentile points available anywhere in this article.

Pitfalls

Five mistakes when you are applying to all three

Preparing for the wrong test at the wrong firm
Bain runs two platforms, BCG runs several, and McKinsey runs two versions of Solve. Every invitation names the assessment. Read it before you plan anything, and confirm with the recruiter if it is ambiguous.
Assuming case practice covers everything
It covers Casey well, Solve partially and Bain barely at all. Candidates who prepare exclusively by doing cases are usually strongest at BCG and weakest at Bain.
Leaving the unfamiliar format last
Solve and the Bain tests both reward interface familiarity, and interface familiarity takes one sitting to acquire. Do it early rather than in the final 48 hours.
Practicing untimed
All four formats are time constrained, two of them brutally. Untimed practice builds accuracy you will not be able to deploy.
Treating a bad feeling as a result
None of the three shows you a score, and most candidates finish all of them convinced they did badly. It is not evidence, and acting on it by disengaging from the next stage is how a passable result gets wasted.

FAQ

MBB online assessments: frequently asked questions

Which MBB online assessment is hardest?

BCG Casey has the highest ceiling because it asks you to do actual consulting work, including a recorded recommendation. McKinsey Solve is the hardest to walk into cold because the interface is unfamiliar and your process is scored as well as your answers. The Bain tests have the simplest content and the least forgiving pacing.

Is McKinsey Solve harder than BCG Casey?

They are hard in different ways. Solve is a game with a data collection phase and a process score, so unfamiliarity costs you. Casey is a case in a chat window, so it is harder in content but far more transferable from ordinary case practice.

Does case practice help with the Bain assessment?

Barely. The Bain SOVA test and the Bain TestGorilla assessment are reasoning batteries with a behavioral questionnaire, not cases. Nothing asks you to structure a problem. Practicing cases is valuable for the interviews that follow, but it will not move your assessment result.

Do all three firms use an online assessment?

All three use one for most candidates, but which assessment you receive varies by office and role, and some candidates skip it entirely. BCG in particular runs several distinct assessments that candidates confuse with each other, so the invitation email is the only reliable source.

What score do you need to pass?

None of the three publishes a cutoff or reports your result. Bain's SOVA percentiles are commonly described as needing the 70th to 80th or better. McKinsey and BCG disclose nothing comparable. The target you can act on is finishing every section with your accuracy intact.

Can you retake any of them?

Not within the same application cycle, at any of the three. Reapplication waiting periods are set locally and vary, so confirm with the recruiter for the office you applied to.

If I only have one week, which should I prepare for?

Whichever assessment arrives first, then fast chart and table math, which is the only skill that pays off across all three firms. Leave case structuring on its own track, since it serves BCG Casey and every live interview but does little for Bain.

Sources

Where these formats come from

No MBB firm publishes assessment specifications. Everything here is candidate reported or drawn from coaching providers, and formats change between cycles and offices.

MyConsultingCoach, BCG Casey guide, 2026. Sequential question structure, the 60 second video recommendation, no returning to previous questions, and the distinction from the CCA and Pymetrics. myconsultingcoach.com
MConsultingPrep, BCG Online Case chatbot guide, 2026. The 25 to 30 minute range, 6 to 8 questions, no pausing, and five minute reminders. mconsultingprep.com
StrategyCase, BCG Online Case, 2026. The competencies Casey scores, and the warning that candidates routinely prepare for the wrong BCG assessment. strategycase.com
CasingLab, BCG interview process 2026. Reported 8 to 10 questions over 30 to 35 minutes, and the note that Pymetrics and the Potential Test are legacy formats in many offices. casinglab.com
IGotAnOffer, Bain Online Assessment guide. Bain's two platforms, section structures and behavioral scoring. igotanoffer.com
StrategyCase, Bain TestGorilla, 2026. Office split between SOVA and TestGorilla, and the completion window. strategycase.com

Keep going

Go deeper on whichever one you have

Bain SOVA Test Guide Five sections, the hidden clock, and how percentile scoring works. Read the guide → Bain TestGorilla Guide Four hard timed blocks, both reported versions, and the pacing math. Read the guide → McKinsey Solve overview How the games fit together, the two versions, and the dual scoring system. Read the guide → Redrock Study guide The 35 minute data module that opens Solve, with a free practice case. Read the guide → What score do you need? Why the published pass rates disagree, and the percentile worth targeting. Read the guide → Numerical practice questions Ten worked items in the format all three firms draw on. Start practicing →