Consulting Interview Process 2026: Eight Paths, Named

Why the singular consulting interview process is the mistake

Ask how many rounds consulting interviews have and the honest answer is that the office decides, and the firm name barely narrows it. OC&C publishes eight office-level sequences. Adding the interviews across those eight gives 34, an average of a little over four, and the average describes none of them.

Yale's career office says as much from the school side. Its 2026 to 2027 note reads: "The specific processes vary by firm although common to all is the inflexibility of their processes." Variation between firms, rigidity inside each one. That combination is why the invitation email outranks every guide, including this one.

Three concrete consequences follow, and each is documented in a section below. A case interview is not universal, because Deloitte's published UK route has none. A round count is not stable inside a single brand, because BCG publishes an unnumbered global sequence and a fully timed Swiss one. And the vendor behind a screen is not fixed, because Bain uses two for the same step.

Eight OC&C offices, eight interview processes, one brochure

The firm's 2024 global graduate recruitment brochure prints a stage list per office. These are the firm's own words, condensed to the stages and counts.

Add the counts: 4 plus 4 plus 4 plus 6 plus 5 plus 3 plus 4 plus 4 is 34.

Now compare interview time rather than interview count. Rotterdam's published minutes are 45 plus 60 plus 90 in the middle stage, which is 195, and 3 times 60 in the final stage, which is 180. Those total 375 minutes, or six hours and fifteen minutes. Australia's are 45 plus 45 plus 45, which is 135 minutes, or two hours and fifteen. The gap between two offices of the same firm is 375 minus 135, which is 240 minutes of interviewing, four hours.

Two offices have no online test at all, Milan and the United States. Rotterdam is the only one with a written case, and Warsaw is the only one that puts an HR interview between the two case rounds. The firm publishes this and never claims the sequences are equivalent.

Deloitte's two interview processes, one with a case and one without

This is the cleanest contradiction in the set, and it sits inside one brand. The Deloitte interview process is published twice, once per country, and the two versions disagree about whether a case happens at all.

The UK early careers assessment page names four stages plus the outcome: Apply Online, Immersive Online Assessment, Job Simulation, Final Stage Assessment, Outcome. It also states that the shape does not change by programme: "Regardless of whether you're applying for a Summer Vacation Scheme, Industrial Placement, or one of our full-time Graduate or Brightstart programmes, our assessment process follows the same structure". The word case appears nowhere in those four stages.

Two details from that page are worth carrying. On the online assessment: "There's no overall time limit for completing the assessment. However, we do recommend trying to complete it in one sitting", followed by "Some questions are timed, meaning the speed of your response will contribute to your score." And on the final stage: "Before you attend, we need you to complete some mandatory preparation which involves researching and conducting analysis on a topic." That final stage runs as a discussion of 30 to 40 minutes on the prepared topic plus a competencies and motivation interview of around 50 minutes.

The US side of the same firm publishes "two to three rounds of interviews" and "a mix of behavioral, technical, and case-based interviews, depending on the role". Deloitte also hosts ten playable practice cases with a suggested answer under each question, which is the most substantial case material any large firm we read gives away.

So a candidate preparing cases for a Deloitte UK graduate programme is preparing for an exercise the published process does not contain. A candidate skipping cases for Deloitte US is doing the reverse.

The two BCG interview processes, and only one is timed

Two pages carry the Boston Consulting Group interview process and only one of them has numbers on it. The global interview process page names four stages: Application, Skills Interview, Case Interview, Team Interview. It gives no round count, no case count and no duration, and we are not going to supply any.

The Swiss application page gives all of them for one country and one role type: "Our interview process for full-time consulting roles begins with a virtual cognitive test, followed by one virtual first-round case interview and a final round of three in-person case interviews."

Four steps, each specified. The test is "a 30-minute cognitive test" covering "numerical, verbal, and logical skills", mostly multiple choice, and to be answered "without any tools or resources". The first-round interview is 45 minutes and carries the fit portion inside it: "First, we'll get to know what drives you. Then, you'll crack a client case". The final round is three in-person cases in the Zurich office. The page also warns that the process can end early in your favor: "Based on our overall impression from the first round, a decision may already be made at that stage."

One boundary the page draws itself: "At BCG Switzerland, the interview process differs slightly depending on whether you're applying for a full-time consulting role or a Visiting Associate internship." It says the internship differs and does not say how. Treating the Swiss numbers as BCG's global process would repeat an error this site already made once with a US-only assessment timing.

McKinsey names its interviews, Bain names its conditions

The McKinsey interview process is named rather than counted. The firm's interviewing page describes the shape and leaves the quantities out: "For most client-facing roles, you will participate in a personal experience interview, followed by a problem-solving interview. Depending on the role, you may also have expertise interviews. Your recruiter will confirm your interview details in advance." Two named interviews, one conditional third, and the firm pointing at your recruiter for the rest. No round count, no duration.

The Bain interview process goes further into the conditional. Its hiring process page reads: "our team will review your application, you might have a call with a recruiter, and you'll have several rounds of interviews. If you apply for a role such as consultant, you may also have a case interview where you work through a problem." Two of those four elements are hedged with might and may, and several rounds is the only quantity offered. The firm does publish one thing about feedback that nobody else does: "After each interview, you'll learn exactly how it went."

Alvarez and Marsal publishes the loosest version, and it is loose on purpose: "Depending on the role, the number of interviews may vary, and you could be asked to complete a case study, business plan, or technical assessment." Three possible written exercises separated by or, triggered by the role. The phrase case interview does not appear on that page.

KPMG Canada adds a constraint the others do not: "You can only apply to one position each recruitment cycle." A wrong first choice there costs the cycle, which makes the office and service line decision heavier than the interview preparation.

The screen that opens the interview process, and its vendors

The online assessment is the step where the firm is often the wrong source. Bain publishes the mechanism plainly: "you will receive an email after submitting the application detailing whether you are required to complete Sova or Test Gorilla". Two vendors, one step, and the firm does not publish the rule that decides which one arrives.

The rules then diverge. Test Gorilla is to be completed in one sitting, and for Sova the firm points at the vendor's own preparation space. Advice about the Bain assessment written in the singular is therefore wrong about half the time. Bain also sets a hard consequence: "Applications will not be considered if the digital assessment is not completed by the deadline."

Elsewhere the screen is timed rather than vendor-named. BCG Switzerland publishes 30 minutes and no tooling. Deloitte UK publishes no overall time limit with some individual questions scored on speed. OC&C publishes the presence of a test for six of eight offices and never its length, its vendor or its threshold.

Nobody in this set publishes a passing score. That is the single most requested number on this topic and it belongs to the employer, who does not release it.

The calendar, where somebody publishes one

Firms publish deadlines per posting and rarely publish a cycle. School career offices publish the cycle, which makes them the better source for timing.

Yale's career development office writes for graduating students: "Application deadlines vary by degree, interviews occur about 2-4 weeks after deadline, and offers made shortly thereafter". For first-year MBA students it is tighter and more specific: "Application deadlines fall mid-November through December. Interviews start in early January during Winter Break with offers extended within one day to one week. Firms do not re-recruit for interns later in Spring semester."

That last clause is the one to plan against. A single window, no second pass in the spring, and a decision inside a week of the interview.

The same office reports a format shift: "In the past few years, firms have started reverting back to hosting in-person interviews for all or part of their recruiting process." It also notes that firms are experimenting at the other end, "from simply discussing prompts and approaches to using AI to actually giving students access to a LLM during the interview process." The office names no firms for either observation, so neither do we.

What one candidate wrote about the part no page describes

Firm pages describe stages. They do not describe what the exercise feels like from the chair, and the only source for that is somebody who sat in it.

Zhe Huang, writing on the Tuck School of Business blog in March 2015 after receiving offers from McKinsey and BCG Shanghai, described a BCG case that began with two hand-drawn lines: an upward investment curve and a downward return curve for Chinese petrochemicals over the preceding five years. He was asked to explain the divergence and predict what came next.

His account of what happened: "I tried to do some 'standard stuff' at first, but was cut off by the interviewer. I knew frameworks and calculations wouldn't help me this time. I listed out my hypotheses and tested my thoughts with him. We were literally discussing the case." And then the part that no process page contains: "When I finished the case, he told me that I did a pretty good job. That was NOT AT ALL how I felt during the process."

One account from one office in 2015 is an anecdote rather than a pattern, and the format may have moved since. We found no other first-hand report we could use: search returns prep platforms, Reddit does not answer our requests, and two promising university posts turned out to be republished competitor content. So this is one data point, dated and attributed, and it is the only one on the page.

What none of them publish, and what to do about it

Six things are absent from every page we read.

1. A round count offered as a general rule. Every count we found was attached to a named office.
2. The weight of the case in the decision. No firm publishes a scoring sheet of any kind.
3. The passing score of any online test. It belongs to the employer and stays there.
4. Whether the second-round interviewer sees the first round's notes. Bain publishes what the candidate learns, and nothing about what travels internally.
5. What triggers a written exercise. Alvarez and Marsal says it depends on the role, OC&C Rotterdam has one as standard, and neither explains the rule.
6. The share of candidates who clear each stage. Nobody publishes it, and estimating it from headcount arithmetic would be an invention.

So the order of operations for your own process is short.

Management consulting interview prep starts from a diagnosis rather than a calendar. Run one drill of each type, read the dimension the score names first, and work that one. Our weights are ours and no firm publishes anything comparable: the structuring rubric scores an answer Coverage 45, Separation 25, Depth 20, Concision 10, read from src/lib/grading/rubrics.ts.

You can see those weights on how an answer is assessed, run the drills free and without an account on tools, and work the fit signals on behavioral and fit practice. What we do not have: none of the six case simulators are running, and the assessments page states that on its own face. Nothing here reproduces a firm's assessment, and for the Deloitte UK route there is no case to practice at all.