Best Platform to Practice Case Interviews for MBB 2026

Five forms to practice case interviews, and what each can do

A missing feature can arrive next month. A property of the form cannot. That distinction is the whole basis for choosing, and it is the one thing product comparisons leave out.

A written case will never mark what you wrote. A peer will never guarantee the level of the person sitting opposite you. An hourly coach will never be awake at two in the morning. A firm's own simulator will only ever measure what that firm decided to measure. None of those sentences is a complaint about anyone's execution. They are descriptions of shape.

BCG's case interview preparation page makes the first row concrete. It names twelve behaviors, six to show and six to avoid. Among the six to avoid are Rush Your Response, Overcomplicate Solutions, Stay Silent and Interrupt. Each of those four describes something that happens between two people. A form of practice with only one person in it has nothing to observe them with, however good its scoring is.

What MBB publishes for case interview practice, and what it hides

The firms build the only simulators that are the genuine format. They also publish less about them than any other form of practice publishes about itself, and the pattern in what they withhold is consistent.

McKinsey's Solve FAQ, a four page PDF dated August 2025 and linked from its assessment pages, states that no preparation is needed and that Solve requires no previous business knowledge or gaming experience. On length it says only that the length of your Solve assessment is outlined in your invitation email. It adds that Solve uses advanced AI methods so that no two individuals have the same parameters and combinations of data, and that a candidate may be asked at random to take an additional task in person or to explain the logic used. On technical faults it is blunt: if you do not contact support, the firm cannot make exceptions to allow for reset.

BCG publishes the most numbers, and they sit on a regional page. Its United States Associate page, re-read on 7 October 2026, says the Consulting Career Assessment is sent 30 to 90 minutes after you submit your application, takes approximately 30 to 35 minutes, and gives you 48 hours to complete it once the invitation arrives. The online case runs about 35 minutes and reaches a select group about two weeks after the application deadline. On repeats the page is explicit: candidates that have previously taken the online case will not be prompted to take it again. Those figures belong to that process in that country, and they do not generalize to BCG elsewhere.

Bain's digital assessment page is the one place where a firm points at a form of practice by name, and it points outside itself: it recommends the candidate preparation area of the assessment vendor's website. On the same page it says no preparation is needed and that a calculator is allowed for that assessment only. One caveat on that row, stated plainly. We could not re-open that page on 7 October 2026. The firm returns 403 to our requests, the text proxy that served it in September is now behind a challenge, and the web archive was offline while we tried. The Bain assessment figures therefore carry their 30 September reading date rather than today's.

Bain does publish one thing on its hiring process page that no other firm here matches, and it bears directly on the comparison: it says any feedback you receive will be based on an objective scoring system, and that after each interview you will learn exactly how it went. That is a firm promising the feedback loop which its own assessment withholds. The scoring system itself is still never described. The vendor's own candidate preparation hub, re-read live on 7 October 2026, adds the detail that matters most about this form. Practice answers do not count toward a score, the employer selects which activities appear, and feedback reaches you only if your prospective employer chooses that option. Whether you learn anything about your own performance is a decision somebody else already made.

BCG's virtual experience page offers 13 free modules hosted on a third party platform, with sample answers supplied by its own leaders and consultants, plus digital badges. A sample answer is the same gift a written case gives. It shows you a good answer and stays silent about yours.

None of the five publishes a mark scheme. That absence is the single most important fact on this page, because it means every weighted score a candidate has ever seen was written by somebody other than the firm doing the hiring.

The best platform to practice case interviews for MBB

The form follows the deficit. Name what actually went wrong in your last attempt, then take the form whose shape can repair that specific thing. The weights quoted below are ours, read from our own grading code, and no consulting firm publishes a scoring scale.

1. A branch was missing from your structure. Work written cases, then score the structure itself. Coverage carries 45 points of 100 on our structuring rubric, the heaviest single weight we use, because a missing branch costs more than an untidy one.
2. The arithmetic was right and you said nothing about it. Answer first carries 35 on our synthesis rubric. A tool catches this in one pass, since the omission sits in the text and needs no judgment to spot.
3. You lost the thread when somebody cut across you. Only a peer or a coach reproduces that. BCG lists Interrupt among the behaviors to avoid, which tells you plainly that somebody will.
4. You do not know which screening format is coming. Take the firm's own material and the vendor's practice test. The invitation email is the only document that names which format is yours, so read it before spending a week on anything.
5. Your structure was tidy and empty. This is the coach's territory. Bain publishes the gap on its FashionCo case: a good answer includes different types of ways to increase revenue, while a better answer includes the pros, cons and practicality of each idea. Telling those apart on your own work is hard, and a person does it faster.
6. It is two in the morning, four days out. Take a tool. It is the only form that is awake, and volume at an odd hour is a real advantage of the shape.

Two of those six answers cost nothing, branch three and branch four, and the material behind branch four is published by the firms themselves. That is the ordering most comparisons invert.

On the one question this ordering does not settle, which of a peer and a scoring tool works better, nobody has measured it and that includes us, so the useful question is what each form is able to observe. That argument has its own page: AI case interview tools compared to peer practice sets the two against what each one can detect, and names what we refuse to claim about our own figures.

Where a structured course beats peer practice

A course sells order. Peer practice sells availability. That trade explains most of the disappointment in both directions.

A course arrives with a sequence somebody designed, which is precisely what a candidate four weeks out with no plan is missing. The same sequence cannot rearrange itself around the one dimension you are short on, so a candidate who already has a diagnosis is paying for a decision they have already made. Peer practice has the opposite profile. It adapts instantly to whatever you ask for and guarantees no coverage at all, so a run of sessions can leave whole dimensions untouched with nothing in the process to flag it.

A free curriculum already exists, and BCG wrote it without labeling it as one. Its preparation page lists the tasks inside a case: structure your approach, ask thoughtful questions, analyze data, perform quick calculations, and identify the most important factors. Five tasks read as five modules, published by a firm that screens with them, at no cost. Our own case interview drills page turns that list into timed blocks with the weights attached.

McKinsey's interviewing page supplies the material for the same curriculum: four full practice cases with suggested answers, which the firm recommends reviewing. Its published guidance on AI marks practice of interview questions as encouraged. Between those two pages the order and the material are both free, which narrows what a paid course is actually selling to accountability and pace.

What is free, and what the free material covers

The free tier in this market is larger than it looks, and almost all of it is published by the firms themselves rather than assembled by anybody selling practice.

McKinsey publishes four complete practice cases with suggested answers and recommends reviewing them. BCG publishes 13 simulation modules at no cost, with sample answers written by its own leaders and consultants, plus the five-task breakdown of what a case asks you to do. Bain publishes a worked case with its good-to-better distinction attached, and points at the assessment vendor's free practice test. The vendor publishes practice questions whose answers do not count toward any score, which makes them genuinely free in the sense that matters.

That inventory supplies two of the five forms outright, the written case and the firm simulator, plus the curriculum a paid course would sell you and a published standard to measure an answer against. What none of it supplies is a second person in the room and a mark on your own work. Those are the two things money buys in this market, and knowing it is most of what a comparison of platforms can honestly tell you.

One caution on the free tier. BCG says its simulations help you stand out and attaches no figure to that, so neither will we. Nobody publishes what a badge is worth to an application, and the firms that would know have not said.

What a coach does that nothing else does

BCG's preparation page says there is not always a single right answer. That one sentence defines the limit of every rubric, including ours. A scoring scheme resolves cases where the answer is checkable. It cannot settle whether your judgment on an ambiguous trade was sound, and a person who has run the interview can.

A coach also owns the one move no other form has. They can change the question in the middle of your sentence, then watch what you do with the wreckage. A written case is fixed before you open it. A tool responds to what you submitted. A peer can try it, without the reference points a coach brings to the interruption.

The limits are equally plain. A coach is hourly, booked in advance, and absent at the hour you happen to be free. Nobody books a coach for the fortieth repetition of the same opening, so the form covers judgment well and repetition badly. Those two limits come from the pricing model and the calendar rather than from any coach's ability. On the rate itself this page gives no figure, and the reason is worth stating: no firm, test vendor or institution publishes coaching prices, and the only published rate cards belong to the companies selling the service. We quote primary sources or we say nothing, so what you get here is the shape of the cost rather than a number we cannot stand behind.

Free human feedback does exist and it is routinely overlooked. University career services publish named slots: Harvard's career services page lists drop-in advice Monday to Friday from 2:00 to 4:00pm, alongside workshops and review clinics. That access is limited to the institution's own students and those sessions are about application documents rather than cases. Consulting clubs are the other route, and our case practice partner page covers how to brief somebody who has never run a case.

What none of the five forms supplies

A mark scheme from a firm. None of the three publishes one, and the six rubrics in our own code carry weights we wrote. Any page showing you a firm's scoring scale invented it.

The length of Solve. McKinsey keeps that in the invitation email, so no form of practice can rehearse you against the real clock.

Material taken from somebody else's sitting. McKinsey's assessment integrity expectations prohibit sharing information about the content of an assessment, including specific questions and problem solving scenarios. The same page allows only pen and paper for notes and calculations in interviews, with no calculator and no prepared notes, and says the firm may ask to review your notes. Any form of practice that runs on an on-screen calculator trains a motion the final room forbids.

And which stage you are actually facing. That arrives by email after you apply, and choosing a form before you know the stage is how a candidate spends a week on the wrong thing.

Our six rubrics return a score on one dimension at a time, with the weights above read from our code rather than borrowed. The structure drill and the math drill are the two that carry a daily hour, and how an answer is assessed sets out each dimension and why it is weighted that way. They teach no business content, because none of the dimensions is a knowledge test. Our assessment simulators cover the firm-built screening formats, and that page states that none of the six is live yet, which is the honest current limit.

Questions about where to practice case interviews for MBB

These are the questions that recur across the results for this query, answered from published sources wherever a published source exists and marked as unmeasured where none does.

Is practice with an AI tool as effective as practice with a human partner? Nobody has measured it, ours included, so the answerable version of the question is what each form can observe, worked through on AI case interview tools compared to peer practice.

Is using AI to prepare allowed? McKinsey answers this on its own page. Its published guidance marks practicing interview questions as encouraged and asking AI to explain concepts or frameworks as part of your preparation as encouraged. Using AI during an assessment, unless specifically permitted, is listed as not allowed. The line sits between preparing and sitting.

Can you prepare using only free material? For the written case and the firm simulator, yes, and the section above lists exactly what each firm publishes. What you cannot get free is a second person and a mark on your own answer.

How many practice cases or partners do you need? No firm publishes a number. McKinsey recommends reviewing its four published cases and gives no target count, and none of the three firms publishes a repetition figure. Any specific number therefore comes from somewhere other than the firm doing the hiring.

Should a beginner start with a tool or a coach? A tool suits a candidate who cannot yet name what went wrong, because a score per dimension produces the diagnosis. A coach suits a candidate who already has the diagnosis and needs judgment on an ambiguous answer. Paying for judgment before you have a diagnosis buys the expensive form of a cheap question.

For what each individual tool trains and the stage it belongs to, see case interview tools. For dividing a preparation across the three firms by what each one publishes, see MBB case interview prep.