Interviewer-led case interview: format, worked math, and how to practice

An interviewer-led case, which McKinsey calls the Problem-Solving Interview, walks you through a realistic business scenario one question at a time. McKinsey publishes 5 sample practice cases on its interviewing page, and BCG's ADC Recruiting Guide notes that North American generalist candidates complete up to 2 rounds of case interviews with 2 interviews per round. Neither firm looks for a single right answer or standard framework, evaluating instead how you structure your thoughts, talk through paper math without a calculator, and tailor your reasoning to the client.

Updated

How the problem-solving interview format works

When candidates search for an interviewer-led case, they are usually preparing for the format McKinsey calls the Problem-Solving Interview, or comparing it with the case study interview at BCG. In both settings, you are given a realistic business challenge based on client work and asked to work through it step by step. You can practice a case free, no sign-up to hear how an interviewer guides the prompts out loud.

On McKinsey's official practice cases, such as Diconsa, Beautify, and Talbot Trucks, the case opens with a client goal, a situation description, and a description of the study. Instead of leaving you to wander through the case on your own, the interviewer guides the conversation through numbered questions (for example, Question 1 through Question 4 in Diconsa and Beautify, and Question 1 through Question 3 in Talbot Trucks). Each question isolates a distinct task: deciding what to investigate, running a multi-step calculation, interpreting survey or cost data, or grouping solutions around a specific stakeholder concern.

Both McKinsey and BCG emphasize that there is no single right or logical solution. McKinsey advises candidates to listen carefully and share tailored solutions rather than standard frameworks. BCG similarly states that what matters most is how you approach the problem and the quality of your reasoning, defining frameworks in its recruiting guide simply as methods for approaching business scenarios using a defined structure to break a problem into its fundamental pieces.

Where the case fits in your overall process depends on the firm and role. At McKinsey, candidates for most consulting roles first take the Solve gamified assessment, and client-facing candidates participate in Personal Experience Interviews (PEI) alongside Problem-Solving Interviews (and expertise interviews for some roles). At BCG, the ADC Recruiting Guide states that North American generalist consultant candidates participate in up to two rounds of case interviews, with two interviews per round.

  • Initial intake: Write down important information as the scenario is read, and ask the interviewer to repeat or explain anything that is unclear.
  • Structuring question: Take time to organize your thoughts and develop an overall approach before diving into details.
  • Quantitative question: Write your calculations on paper without a calculator and talk the interviewer through every step.
  • Data and chart question: Take time to look at the information, note observations, and identify trends that are not immediately obvious.
  • Recommendation and synthesis question: Consider the specific issues raised in the prompt and group your thoughts around them.

Reasoning and communication steps from the firms' prep guides

Because the interviewer controls which question comes next, your score depends on how you handle each turn of the conversation. Both McKinsey and BCG publish explicit guidance on what to do and what to avoid during each stage of the case.

At the start of any question, pause before speaking. McKinsey's Diconsa, Beautify, and Talbot Trucks pages repeat a core helpful hint: take time to organize your thoughts before answering and develop an overall approach before diving into details. Doing so shows the interviewer that you have a logical, organized approach regardless of the accuracy of the final outcome. When you need to build a tailored structure from scratch, our guide on consulting case study examples shows how to split a problem into branches that fit the exact client.

During quantitative questions, calculators are not allowed. McKinsey instructs you to write out calculations on paper and talk your interviewer through your steps. The firm notes that the more you talk through your logic, the easier it is for your interviewer to help you. BCG likewise notes that cases always include a quantitative portion where you perform quick calculations and clarify assumptions.

When the interviewer shares a table or survey results, do not just read the top numbers back. McKinsey advises taking time to note down observations and challenging yourself to identify trends that are not immediately obvious from the data. When the interviewer asks how to overcome client objections or what to do next, consider the exact issues raised in the question and group your thoughts around them so your answer stays relevant.

BCG's official Do and Don't guidelines for case study interviews (BCG Case Interview Preparation page)

BCG's official Do and Don't guidelines for case study interviews (BCG Case Interview Preparation page)
Do during the caseDon't during the case
Listen ActivelyPanic Under Pressure
Be YourselfRush Your Response
Think StructurallyOvercomplicate Solutions
Communicate ClearlyIgnore Feedback
Clarify AssumptionsStay Silent
Show Your ThinkingInterrupt

Official McKinsey practice cases and what each one tests

McKinsey publishes 5 sample cases on its interviewing page to show the logic, thought processes, and suggested answers used in the Problem-Solving Interview. BCG also provides sample cases on its careers website.

Reviewing the exact prompts in McKinsey's sample cases shows how an interviewer-led case sequences its questions. In Diconsa, Beautify, and Talbot Trucks, every question shifts the angle of analysis while keeping you anchored to the client's overarching goal.

McKinsey's 5 official sample practice cases (McKinsey Interviewing page)

McKinsey's 5 official sample practice cases (McKinsey Interviewing page)
Practice caseClient scenarioPublished questions and tasks
BeautifyGlobal beauty products company evaluating whether to train in-store beauty consultants to use virtual channels to connect with customers.Q1: Factors to consider when shifting staff to virtual social media-beauty advisors. Q2: Features to make a department store customer switch to virtual sales. Q3: Years until the upfront investment turns profitable on €1.3B revenue. Q4: Analyzing competitor AI chatbots and recommending next steps.
DiconsaStudy assessing whether Mexico's government-owned chain of 22,000 rural Diconsa stores can provide basic financial services to the rural poor.Q1: What to investigate to determine if Diconsa should provide basic financial services. Q2: Total annual savings across rural Mexican families collecting state benefits at Diconsa stores. Q3: Observations and trends from rural survey and store data. Q4: How to sell financial products effectively given minimal education, unfamiliarity with banking, and significant crime and fraud.
Talbot TrucksEuropean truck manufacturer assessing the attractiveness and feasibility of producing and selling electric trucks (eTrucks).Q1: Information to collect on the attractiveness of producing and selling eTrucks in Europe. Q2: Inferences from Total Cost of Ownership (TCO) data comparing a €100,000 diesel truck and an eTruck. Q3: Maximum eTruck price so TCO equals a €100,000 diesel truck.
Electro-LightCore sample case listed on McKinsey's interviewing page.Listed on McKinsey's main interviewing page as one of the 5 core sample cases.
Transforming a national education systemCore sample case listed on McKinsey's interviewing page.Listed on McKinsey's main interviewing page as one of the 5 core sample cases.

Worked math examples from Diconsa and Beautify

Every case includes a quantitative portion, and practicing the exact arithmetic from McKinsey's published prompts is one of the best ways to build speed without a calculator. Below are the worked setups and calculations from the numbers McKinsey states in Question 2 of Diconsa and Question 3 of Beautify. You can also compare these with our full McKinsey case interview practice page.

In Diconsa Question 2, McKinsey gives the following inputs: Mexico has a population of 140 million. Twenty percent of the population is rural, and half of that rural population currently receives state benefits. Families have an average of 4 members across all regions. Collecting benefits at local Diconsa stores reduces a family's monthly collection cost in transportation and food by 30% for the 60% of the rural population living furthest from a bank branch, and by 20% for the 40% living relatively close to a bank branch. The question asks for the total saved per year across all rural families receiving state benefits.

In Beautify Question 3, McKinsey gives the following inputs: Beautify's current revenues are €1.3 billion. With virtual advisors, the team expects a 10% overall increase in incremental revenue starting in the first year. Beautify expects a 5% annual depreciation (loss in value) on its upfront IT investment each year, and €10 million in all-in yearly costs starting during the first year. The question asks how many years it takes until the investment turns profitable, framed as when incremental revenues are greater than the cost of the project.

  • Diconsa Step 1 (Rural population receiving benefits): By our calculation from McKinsey's Diconsa page, 20% of 140 million people is 28 million rural residents, and half of that group is 14 million people receiving state benefits.
  • Diconsa Step 2 (Number of benefit-receiving families): Dividing 14 million people by 4 members per family gives 3.5 million rural families receiving benefits.
  • Diconsa Step 3 (Segmenting the families by distance): 60% of 3.5 million families is 2.1 million families in the furthest segment (where monthly collection cost drops by 30%), and 40% of 3.5 million families is 1.4 million families in the closest segment (where monthly collection cost drops by 20%).
  • Diconsa Step 4 (Weighted cost reduction and annual total): Across all 3.5 million benefit-receiving families, the weighted average cost reduction is 26% (60% x 30% + 40% x 20%). Multiplying each segment's monthly pesos saved per family by the number of families and by 12 months gives the total annual pesos saved.
  • Beautify Step 1 (Annual incremental revenue): By our calculation from McKinsey's Beautify page, 10% of €1.3 billion is €130 million in annual incremental revenue starting in Year 1.
  • Beautify Step 2 (Annual ongoing costs): Beautify incurs €10 million in all-in yearly operating costs starting in Year 1, plus 5% annual depreciation on its upfront IT investment.
  • Beautify Step 3 (Payback comparison across years): To find the year the project turns profitable, compare cumulative incremental revenue (€130 million per year) against the sum of Year 1 upfront investments, cumulative €10 million yearly costs, and cumulative 5% annual IT depreciation.

How to practice out loud (and rules for using AI and virtual setups)

Reading sample cases silently does not build the habit of talking an interviewer through your math or grouping thoughts on the spot. BCG explicitly encourages candidates in its recruiting guide to practice case interviews with someone else when possible, and you can also rehearse BCG case interview formats or behavioral prompts out loud.

McKinsey publishes clear rules on how you can and cannot use AI during interview preparation. On its interviewing page, McKinsey states that using AI to polish your resume, practice interview questions, and ask AI to explain concepts or frameworks as part of your prep is encouraged. However, exaggerating or creating achievements that are not yours, generating real-time answers during an interview, or using AI during assessments (unless specifically permitted) is not allowed.

Use the 6-step rehearsal checklist below, written by our team from McKinsey's and BCG's published guidance, every time you practice a question out loud.

  • Set up a real interview desk: Put away your calculator, open blank paper and a pen, and keep water and your laptop charger nearby.
  • Take notes and clarify first: Listen actively as the prompt is read, write down every number and constraint, and ask the interviewer to repeat or explain anything unclear before you start structuring.
  • Pause to build an overall approach: Take a quiet moment to organize your thoughts into fundamental pieces tailored to the client rather than forcing a standard memorized framework.
  • Talk through paper math step by step: State your equation out loud before multiplying, label your units on paper (pesos, euros, families, years), and walk the interviewer through each intermediate number.
  • Look past the obvious in charts and tables: Note down initial observations, then push yourself to explain the underlying cause and identify trends that are not immediately obvious from the data.
  • Group synthesis answers around the exact prompt: When asked how to handle a client risk or what to recommend, group your points directly around the concerns raised in the question.

Interviewer-led vs. candidate-led mechanics: handling the mini-case structure

The biggest adjustment in an interviewer-led case is what happens right after you present your opening framework. In a candidate-led case, you build a structure and then choose which branch to investigate first, asking the interviewer for data as you drive the case forward. In an interviewer-led case, as shown in McKinsey's Diconsa, Beautify, and Talbot Trucks sample cases, the interviewer listens to your Question 1 structure and then introduces Question 2, Question 3, and Question 4 in a set sequence.

Do not assume that your initial framework failed if the interviewer moves to a completely different topic on Question 2. Because the interviewer has a set sequence of prompts to cover, they will redirect you even after a great opening structure. Treat each numbered prompt as a self-contained mini-case: pause to organize your thoughts, build a mini-structure or equation for that specific question, talk through your steps out loud, and tie your conclusion back to the client's main goal.

Following the interviewer's sequence does not mean becoming passive. You still drive the analytical work inside every question by clarifying assumptions, showing your thinking on paper, identifying non-obvious trends in tables, and grouping your final synthesis around the client's concerns. You can also practice speaking your personal stories out loud on our behavioral interview prep page. Once you are ready to test your pacing across both case and fit questions, practice a case free, no sign-up.

How interviewer-led and candidate-led case mechanics compare

How interviewer-led and candidate-led case mechanics compare
Case stageInterviewer-led format (e.g., McKinsey sample cases)Candidate-led format
Question sequenceInterviewer moves through distinct, pre-set questions (Q1, Q2, Q3, Q4).You propose which branch of your structure to test next and ask for the data.
Framework transitionYou present an overall approach for Q1, then the interviewer directs you to a specific Q2 prompt.You use your opening framework as a roadmap and choose the first branch to explore.
Handling each promptTreat every question as a mini-case: pause, organize your approach, talk through the steps, and answer the exact prompt.Connect each data request to your current hypothesis and decide when to switch branches.
Staying proactiveClarify unclear points, talk through paper math step by step, spot deeper data trends, and group synthesis points.Drive the agenda from branch to branch while listening and adapting to interviewer cues.

From McKinsey & Company: Interviewing at McKinsey, McKinsey & Company: Diconsa Practice Case, McKinsey & Company: Beautify Practice Case, McKinsey & Company: Talbot Trucks Practice Case, Boston Consulting Group: Case Interview Preparation and Boston Consulting Group: Recruiting Guide & FAQ for Advanced Degree Candidates.

1 practice question and sample answer

Say your answer out loud first, then read the sample answer.

Practice question

  1. Diconsa Question 4 (McKinsey official sample case): Representatives from the state-owned bank express concern about how challenging it will be to offer basic financial products to the rural population. One representative mentions that these are regions where people have minimal education and are unfamiliar with banking products, and there is also significant crime and fraud. How could the bank sell these services effectively to the population given these challenges?

    Sample answer. Example answer written by our team. The applicant is invented. "To help the state-owned bank sell basic financial services across the 22,000 Diconsa stores, I would group our approach directly around the three challenges the bank representative raised: minimal formal education, unfamiliarity with banking products, and high crime and fraud. First, to address minimal education, we can replace text-heavy forms with visual, icon-based enrollment at the store counter and have trusted local Diconsa store clerks explain how benefit collection works in person during regular grocery trips. Second, to overcome unfamiliarity with banking, we should sequence the rollout just as the study proposes: start strictly with dispensing existing cash benefit payments that families already collect, and only introduce savings accounts or bill payments once families see that their money is safe and saves them the monthly trip. Third, to reduce crime and fraud for both residents and stores, we can stagger benefit payout days so stores do not hold large cash balances at once, use biometric or simple PIN verification to prevent identity fraud, and encourage residents to keep funds in basic Diconsa accounts rather than carrying cash home."

    Follows McKinsey's official helpful hint for Diconsa Question 4 by grouping the response directly around the exact issues raised in the prompt (minimal education, unfamiliarity with banking, and crime/fraud) and tying them back to the case details (22,000 stores and phased rollout from benefits to savings).

Practice the case interview, out loud

Soreno's AI interviewer runs a full case one question at a time, checks your math and scores your structure, any time you have 30 minutes.

Questions people ask

Is there a single right answer or standard framework required in an interviewer-led case?

No. Both McKinsey and BCG state that there is not a single right or logical solution in a case interview. McKinsey advises candidates to listen carefully and share tailored solutions rather than standard frameworks, while BCG emphasizes that the quality of your reasoning and how you approach the problem matter most.

Are calculators allowed during the quantitative portion of a McKinsey case interview?

No. McKinsey explicitly states on its Diconsa, Beautify, and Talbot Trucks practice case pages that calculators are not allowed, and recommends writing your calculations on paper while talking the interviewer through your steps.

Can I use AI tools to prepare for my consulting case interview?

Yes, for preparation only. McKinsey encourages candidates to use AI to practice interview questions, explain concepts or frameworks during prep, and polish resumes, but strictly prohibits using AI to generate real-time answers during an interview or during assessments unless specifically permitted.

What should I do if I do not understand a question during the case?

Speak up and ask the interviewer to repeat or explain anything that is unclear. Both McKinsey's interview tips and BCG's core guidelines emphasize active listening and clarifying assumptions early.

How many case interviews do generalist candidates have at BCG?

According to BCG's ADC Recruiting Guide, generalist consultant candidates across North America participate in up to two rounds of case interviews, with two interviews per round.

Keep going