Group Case Interviews 2026: Which Firms Publish One

Eleven sources on the group case interview, and one silence

The table below is every group exercise in the corpus with whatever specification came attached to it.

McKinsey, BCG and Bain are absent from that table. We read their interviewing and hiring pages for the sibling process work, and none of them names a group exercise. That is a statement about what those three publish. Whether such an exercise exists anywhere inside them is a separate question, and no source in our corpus answers it.

Two of the nine firms treat the whole day as optional. IBM writes that you "may be invited to schedule an interview (phone, video or in person) or an assessment center visit (group exercise and in-person interview)." WTW makes it depend on geography: "Depending on your location and the program you're applying for, you'll be invited to an assessment center or an interview."

Eight attributes, four behaviors, three verbs

Three sources publish what they are looking at, and the three lists barely overlap.

KPMG India publishes the longest: its group discussion "enables evaluation of soft skills like oral communication, articulation, structured thinking, attitude and attributes like proactiveness, presence of mind, teaming and leadership." Counting those by hand gives eight named attributes. Five of the eight describe how you speak or carry yourself, and three describe reasoning or role.

Charles River Associates publishes four observed behaviors, and frames the exercise as an addition: "In addition to evaluating your individual performance, we sometimes want to see how you interact with others. We will be looking at how you speak to and listen to others, your level of professionalism and your comfort working with new people."

SHL, the publisher, reduces it to three verbs: "Candidates work in teams to discuss and solve a real or hypothetical issue. Assessors observe how participants communicate, collaborate, and persuade others."

Persuasion appears only in the publisher's list. Listening appears only in the firm's. And Accenture Australia names a third set: "teamwork, enthusiasm, and communication skills." Nothing in these lists is a scoring scheme, and no source publishes a weight.

Ninety minutes of talk for ten of output

Accenture Australia's post, written by one of its own early careers recruiters, is the only source that times the group work.

The sequence it gives: a 30-minute welcome presentation, then breakout groups with "half an hour or so to get to know each other a bit", then after a break "the case study for around 1.5 hours to discuss and form your ideas", then "a presentation of around 5-10 minutes."

Dividing 90 minutes of discussion by 10 minutes of output gives 9, so the group spends nine minutes working for every minute it speaks. That is the number that should govern how a group behaves, and the recruiter draws the same conclusion in her own words: "make sure you allow enough time to prepare the final presentation and, ideally, have a practice run through."

She also publishes the clock discipline: "The group case study assessments are run to a strict time schedule. It's important to keep track of time to ensure your group completes the exercise." And she suggests who should hold it: make sure there is a timekeeper, and volunteering is allowed.

Are you competing with the room

Two published sentences give opposite answers, and both are worth carrying into the room.

Capgemini Invent says it outright: "You're not competing against anyone else, in fact you'll need to work together to be successful on the day!" The same page offers "an opportunity to have a graduate Mentor before your Assessment Centre to help you prepare."

SHL sells the same room to employers on the opposite logic. Its business case for assessment centres lists as an advantage that they "provide the added advantage of evaluating multiple candidates simultaneously, observing how individuals interact in team settings."

Both are true from where they stand. The employer is comparing people; the day's design still rewards a group that finishes. Accenture Australia lands on the practical instruction: "Don't treat your team members like competitors, instead focus on working together to come up with the best solution to the case study you've been given."

What nobody publishes is whether a group that fails as a group can still contain someone who passes. SHL comes closest with a statement about the whole day rather than the exercise: "If there is more than one exercise, you will have other opportunities to show what you can do. If you feel you have performed poorly on one exercise don't give up. Your performance across all of them will be taken into account."

Thirty minutes, from someone who sat a group case

One first-hand account exists and it needs its label before its content. Caitlin Aykroyd, a financial mathematics graduate, described her day on her university careers blog in October 2022. It was EY actuarial consulting rather than strategy consulting, it is one person, and she describes herself as a mathematics student, which colors her judgment of the numerical parts.

Her day "consisted of an individual written test, numerical test and a group exercise." On the group part she gives the only duration we have from a participant: "Since we had 30 minutes to complete our team exercise, I decided that I was going to keep track of the time, and as a group we decided how the time was going to be divided for the different tasks. I gave frequent reminders of how long we had left which ensured that we stayed on track."

She took the timekeeper role that Accenture Australia recommends, at a different employer, without either source knowing about the other. Two independent sources converging on one behavior is the closest thing to a pattern in this section, and it is still two.

One more detail deserves repeating because no firm page says it: "A year prior to this, I had reached the assessment centre for a summer internship role with EY, but unfortunately didn't get past this stage." She failed the same stage at the same employer and passed a year later.

What to do when you cannot get a word in

Five moves below, each from a published instruction, with the source named. Nothing here is our invention.

1. Volunteer for the clock. Accenture Australia suggests a timekeeper and says you can take the job. The 9 to 1 ratio of discussion to readout means the role has real leverage.
2. Pull in the person who has gone quiet. The same post lists "Asking team members for input, especially if you notice someone hasn't spoken for a while" as a way to show teamwork. The same list includes "Bouncing off the ideas of others" and warns against turning the exercise into an interrogation.
3. Ask a question when you are unsure. SHL's candidate page is direct: "Don't be afraid to ask if you are unsure about what you have to do." Accenture Australia adds that "Assessors love to get questions because it shows that you are engaged."
4. Request your turn out loud. This is the only published script for being talked over, and it comes from a participant rather than a firm: "If you are struggling to be heard, make a request that you be the next person to share your thoughts. If this is denied, this will not look bad on you."
5. Play to the strength you have. Accenture Australia's advice for quieter candidates is to ask questions of the group "to show you're thinking about the problem and adding value to the discussion."

What a group exercise rewards on the reasoning side is the same thing a solo case does: covering the ground before diving into one branch. Our structure rubric weights Coverage 45 out of 100 for that reason, and Separation 25 for keeping the branches from overlapping, which is exactly what a group that talks over itself destroys. Those weights are ours, read from our own code, and no firm or publisher in this corpus publishes anything comparable. The reasoning sits on how an answer is assessed and the structure drill runs free and without an account from the tools page. It cannot simulate five other people.

What no source publishes about the group case interview

Six gaps, and they are consistent across all ten sources.

Nobody publishes the size of a group. No firm does, and neither publisher does. Accenture Australia describes breakout groups without a number.

Nobody publishes the duration of the group exercise itself. The only firm number in the corpus is 1.5 hours of case discussion at Accenture Australia, and the only exercise timing, 30 minutes, comes from a participant at a different employer in a different practice.

Nobody publishes the number of observers, or whether the same people watch every exercise.

Nobody publishes a scoring scheme. SHL says assessors rate "against the competencies required for the job," which puts the framework with the employer, and no employer here publishes it.

Nobody publishes whether roles are assigned. Accenture Australia's suggestion that you can volunteer implies nothing is handed out, without saying so.

And nobody publishes what an unfinished exercise costs. The instruction is to finish on a strict schedule, with no statement of the penalty. For the readout that usually follows see consulting case presentations, for the inbox exercise the same publisher sells alongside this one see work simulation assessments, and for the full assessed day see the assessment center written case.