Mumbai Coffee Cups Guesstimate round·Consulting·Easy·20 min

MBB Business Analyst Interview — Mumbai Coffee Cups Guesstimate

20 min · 1 credit · scorecard at the end

Take this on a laptop or desktop — not your phone. The live interview needs a full screen and keyboard (including a sketch whiteboard on coding rounds). You can buy now, but start it from a computer.

Z
ZeroPitch Research Desk
Updated 2026-05-23 · Newly added · be among the first to take it
Field
Consulting
Company
MBB
Role
Business Analyst
Duration
20 min
Difficulty
Easy
Completions
New
Updated
2026-05-23

How to prepare

What this round tests, what strong and weak answers sound like, and the traps to sidestep.

What this round is about

  • Topic focus. A single market-sizing guesstimate: estimate how many cups of coffee are sold in Mumbai in one day, the canonical warm-up consumption case in Indian MBB first rounds.
  • Conversation dynamic. The interviewer states the prompt and goes quiet; you drive scope, structure, math, and the business implication while he pushes back on assumptions.
  • What gets tested. Whether you clarify before calculating, build a structure out loud, segment the population, narrate the math, sense check, and state a so-what.
  • Round format. One scenario section, roughly 19 minutes, with an interviewer playing an Engagement Manager who would staff you on a client.

What strong answers look like

  • Scope before math. You ask whether it is cups or rupees, takeaway or dine-in or at-home, and which Mumbai, before any number appears.
  • Structure stated out loud. You say the equation you will use, for example population times coffee-drinking share times cups per drinker per day, and check it before computing.
  • Real segmentation. You split Mumbai into groups that drink coffee differently, such as office workers, students, tourists, and lower-income residents who lean to tea or roadside options.
  • Sense check and so-what. You sanity check the total against an anchor and close with what the number means for a client weighing a Mumbai coffee entry.

What weak answers look like (and how to avoid them)

  • Framework recital. Naming a memorised approach instead of structuring this specific problem; build your own structure from the prompt instead.
  • Math before structure. Multiplying numbers before stating an equation; say the equation first and confirm it.
  • One-size population. Treating all 20 million Mumbaikars as identical coffee drinkers; segment into groups with different consumption rates.
  • Silent computation. Calculating in your head with no narration; talk through every multiplication so the logic is visible.

Pre-interview checklist (2 minutes before you start)

  • Recall the scope levers. Cups versus rupees, takeaway versus dine-in versus at-home, city proper versus metropolitan region.
  • Have a population anchor ready. Mumbai is often anchored at roughly 20 to 22 million; be ready to propose and defend an anchor.
  • Think of your segments. Office workers, students, tourists, and lower-income residents consume coffee at very different rates.
  • Identify your cross-check. Be ready to triangulate with a supply-side estimate using cafe, tapri, and office density.
  • Pull up the so-what habit. Have a one-line client implication ready the moment you land a number.

How the AI behaves

  • States the prompt and goes quiet. It will not hand you the structure, the population, or the segmentation.
  • Probes every assumption. It asks for the number behind any input you state and pushes weak assumptions.
  • No mid-round praise. It will not say great answer; it acknowledges what you said and pushes deeper.
  • Interrupts on rambling. If you go past about 90 seconds without structure or compute silently, it asks you to back up and narrate.

Common traps in this type of round

  • Cookie-cutter framework. Reciting a named approach the interviewer immediately recognises as not thinking about this problem.
  • Equation skipped. Calculating before stating the relationship between inputs, then losing track of steps.
  • Undifferentiated population. No segmentation, so a 25-year-old office worker and a low-income tea drinker are averaged together.
  • Order-of-magnitude blindness. Landing a number that is wildly off and not catching it on a sense check.
  • No business implication. Stopping at the number and never stating what it means for a client decision.
  • Passive driving. Waiting to be led instead of owning the next step of the estimate.
Reference

The full breakdown

How you're scored, the questions candidates ask most, and the research this interview is built on. Skim it — or just start the interview.

Interview framework

You will be scored on these 6 dimensions. The full rubric with definitions is below.

Scope Clarification Discipline
Whether you pin down cups versus rupees, out-of-home versus at-home, and which Mumbai before any number appears.
18%
Structure Before Math
Whether you state the equation linking your inputs out loud and confirm it before computing anything.
20%
Population Segmentation Quality
Whether you split Mumbai into groups that drink coffee at genuinely different rates instead of one blended average.
22%
Narrated Estimation Control
Whether you talk through the arithmetic and keep the answer in a defensible order of magnitude.
18%
Sense Check And Recovery
Whether you sanity check the total and fix one wrong assumption rather than freezing or restarting when challenged.
12%
Business So-what Articulation
Whether you state in one line what the number means for a client deciding on a Mumbai coffee entry.
10%

What we evaluate

Your final scorecard breaks down across these dimensions. The full rubric and tier criteria are revealed inside the interview itself.

  • Scope Clarification Grounding18%
  • Estimation Structure Rigor20%
  • Population Segmentation Depth20%
  • Narrated Mental Math Control16%
  • Assumption Stress Recovery14%
  • Client Business Implication Articulation12%

Common questions

What does the MBB market-sizing guesstimate round actually test?
It tests how you structure an estimate before you calculate, not whether you reach the right number. The interviewer is watching for a clarified scope, an explicit equation or tree built out loud, a Mumbai population segmented into groups that drink coffee at different rates, deliberately simple assumptions, narrated arithmetic, a sense check, and a one-line business implication. Reaching a defensible order of magnitude with clean structure beats a precise number reached by silent guessing. This mirrors the canonical warm-up consumption guesstimate that opens most Indian MBB cases.
How should I structure my answer to the coffee cups question?
Start by clarifying scope: cups or rupees, takeaway or dine-in or at-home, city proper or the metropolitan region. State the equation you will use out loud before any math. Break the Mumbai population into segments that consume coffee differently, such as office workers, students, tourists, and lower-income residents who lean to tea or roadside options. Pick round numbers on purpose, narrate every multiplication, then sanity check the total against something you know and state what the number means for a client decision.
What are the most common mistakes in this round?
The repeat killers are reciting a memorised framework instead of structuring this specific problem, jumping into multiplication before stating an equation, treating all of Mumbai as one identical coffee drinker, computing silently so the interviewer cannot follow you, stating assumptions without flagging which are weakest, landing a number off by an order of magnitude without noticing, and stopping at the number with no business so-what. Waiting passively to be led instead of driving the case also reads as a red flag at the MBB bar.
How is this AI interviewer different from a real MBB interviewer?
It behaves like a real first-round Engagement Manager, not a coach. It states the prompt and goes quiet, it never hands you the structure or the segmentation, it probes every assumption you cannot ground in a number, and it will not praise you mid-round. It interrupts if you ramble past about 90 seconds without structure and it asks you to narrate when you go silent. The difference from a human is consistency: it probes every claim the same way and produces a transcript-backed scorecard at the end.
How is scoring done in this practice round?
Your transcript is scored against the dimensions a real MBB case leader grades: scope clarification, structure before math, population segmentation, assumption transparency, narrated mental math and order-of-magnitude control, sense check and recovery, and the business so-what. Each dimension has observable signals tied to what you actually said. The scorecard names the specific moment your structure or estimate broke and whether you stated an implication, so you get a concrete diagnosis rather than a vague grade.
What should I do in the first two minutes of this case?
Do not start multiplying. Spend the opening on clarifying questions: confirm cups versus rupees, takeaway versus dine-in versus at-home, and which Mumbai you are sizing. Then say the equation you intend to use in one or two sentences and check it with the interviewer before you compute. Those first two minutes are where the interviewer decides whether you are a structured thinker or a calculator, so make the structure visible before any number appears.
How do I handle it when my final number looks wrong?
Do not restart from scratch. Say out loud that the number feels off and compare it to something you know, such as a per-person cups-per-day intuition or the Mumbai population anchor. Identify the single assumption most likely to be wrong, for example your out-of-home share or your per-capita frequency, adjust just that one input, and recompute. Recovering by isolating and fixing one assumption is a strong positive signal at MBB; freezing or rebuilding everything is not.
What does a strong answer to this guesstimate sound like?
It opens with two or three crisp scope questions, then a stated equation like population times share who drink coffee times cups per coffee drinker per day plus an out-of-home adjustment. It segments Mumbai into office workers, students, tourists, and lower-income residents with different consumption rates, uses round numbers chosen on purpose, narrates each step, lands a number in a defensible range, sanity checks it, offers a supply-side cross-check using cafe and roadside density, and ends with what this means for a client weighing a Mumbai coffee entry.
Should I use a top-down or bottom-up approach for coffee in Mumbai?
Top-down from the Mumbai population is the cleaner primary path for a consumption question like this: start from population, narrow to people who drink coffee, apply cups per coffee drinker per day, then layer an out-of-home versus at-home split. A bottom-up supply-side estimate using number of cafes, tapris, and offices times cups served per outlet per day is an excellent cross-check. The strongest candidates run one as the primary and offer the other to triangulate and reconcile any gap.
Why does the interviewer keep asking what the number means?
Because a number with no business implication is useless to a client. The interviewer here has a beverage client weighing a Mumbai entry, so the so-what is the point of the estimate, not an add-on. After you land a figure, state what it implies, for example whether the daily volume supports a premium chain rollout or what share a new entrant would need to be viable. Candidates who compute correctly but never state the implication consistently underperform at the MBB bar.
How rigorous should the mental math be for an entry-level MBB candidate?
You are not graded on speed or perfect arithmetic. You are graded on whether the math is narrated, whether you chose round numbers deliberately to reduce error, and whether you control the order of magnitude. A small arithmetic slip you catch on a sense check costs little. A silent computation, an unflagged shaky assumption, or a millions-versus-billions error you do not notice costs a lot. Talk through every step so the interviewer can follow and verify your logic.

Sources this interview is built on

Real candidate-report URLs (Glassdoor / AmbitionBox / PrepInsta / GeeksforGeeks / Medium) reviewed when authoring the questions, persona, and rubric. Verify the realism yourself.