Financial Analysis interview preparation

Questions

100

Traced to a firm

42

Firms

28

Updated

September 2026

Asked at

All firmsMoody's7Bain Capital3SSState Street3AMAres Management2BLBlackRock2DED.E. Shaw2MSMorgan Stanley2Oaktree Capital Management2S&P Global2Bridgewater Associates1Citadel1FTFranklin Templeton1Golub Capital1HWHarris Williams1Houlihan Lokey1J.P. Morgan1Jane Street1MWMarshall Wace1Millennium Management1Morningstar1PIMCO1Sycamore Partners1TSTruist Securities1Two Sigma1Vanguard1WMWellington Management1Wells Fargo Securities1Wolverine Trading1

Topic

All topicsThree statements9Accounting policy and standards5Working capital and cash7Ratio analysis8Forecasting and budgeting9Variance and management reporting7Unit economics and costing8Capital budgeting7Cost of capital and valuation7Markets and rates5Modelling, Excel and data8Business partnering6Brainteasers and estimation4Fit and career10

Level

AnyCoreIntermediateHard

Type

AnyTechnicalCaseBrainteaserMarket viewFit

1. 028What is return on invested capital, and why would you prefer it to return on equity?Ratio analysisIntermediatetechnicalCorporate financeKPO research support

   Say this

   ROIC is after-tax operating profit divided by the capital actually invested in operations, so debt plus equity less cash and non-operating assets. I prefer it to ROE because it measures the business rather than the financing decision.

   Then walk it

   1. The numerator is NOPAT, which is EBIT times one minus the tax rate. That removes interest, so the capital structure does not contaminate the return.
   2. The denominator is the part people get wrong. Strip out surplus cash, investments in unrelated assets and anything not generating the EBIT you put on top, or you understate the return of a cash-rich company.
   3. Then the only comparison that matters: ROIC against WACC. A 19 percent ROIC on a 12 percent WACC means reinvestment compounds value. The spread, not the level, is the signal.
   4. ROE fails this test because it can be engineered. Borrow, buy back stock, and ROE rises with no operating change. ROIC will not move.
   5. Practical cautions: goodwill from acquisitions sits in invested capital, so a serial acquirer looks worse than an organic builder even when it is not; and a business with fully depreciated assets looks better than it deserves. I would show ROIC both with and without goodwill.
   6. For a company with several businesses, group ROIC is close to useless. Segment ROIC is where the real story sits, and the segment note usually gives you enough to build it.

   Where candidates lose it

   Defining the denominator as total assets or total capital without removing cash and non-operating items. That single shortcut makes a cash-rich Indian IT company look mediocre when its operating return is excellent.

   Expect next

   • How would you treat goodwill in invested capital?
   • What is the ROIC minus WACC spread telling you?
   • How would you compute segment ROIC from published accounts?
2. 034Top-down or bottom-up budget: which do you trust?Forecasting and budgetingIntermediatetechnicalCorporate FP&ABusiness finance

   Say this

   Neither on its own. Bottom-up gives you ownership and detail but is systematically sandbagged on revenue and padded on cost. Top-down gives you ambition and consistency but nobody in the business feels accountable for it. Run both and negotiate the gap.

   Then walk it

   1. The bias direction is predictable. If bonuses depend on beating the budget, business units submit conservative revenue and generous cost. I have seen a 6 to 10 percent gap between bottom-up submissions and what the same teams then delivered.
   2. Top-down has the opposite problem: it is arithmetically consistent with the board's growth target and operationally unowned, so nobody defends it in month four.
   3. So the process I would run is a top-down target as a frame, bottom-up build as the content, then a reconciliation that names each gap with an owner. The reconciliation table is the whole value of the exercise.
   4. What breaks the deadlock is drivers, not negotiation. If the top-down needs 14 percent growth and the bottom-up gives 9, the conversation is about headcount, capacity, price and pipeline, not about whose number is braver.
   5. I would also separate the plan from the commitment. A stretch target that drives behaviour and a realistic forecast the CFO uses for cash and covenants can be different numbers, and pretending they must be the same is how forecasts become unusable.
   6. The limitation: both approaches assume last year is a sensible base. Where the business model is changing, neither works and you have to rebuild from unit economics.

   Where candidates lose it

   Picking one and defending it as a matter of principle. The answer interviewers want is that you know the bias in each and run a reconciliation. Quantify the sandbagging gap if you can, because that proves you have seen a budget cycle.

   Expect next

   • How do you deal with a business head who always sandbags?
   • Should the target and the forecast be the same number?
   • How long should a budget cycle take?
3. 035What is a rolling forecast, and why would a company move off an annual budget?Forecasting and budgetingIntermediatetechnicalCorporate FP&AGCC finance centres

   Say this

   A rolling forecast always looks the same distance forward, usually four to six quarters, and gets re-cut every month or quarter by adding a new period as the old one drops off. You move to it because an annual budget is most wrong exactly when you need it most, in the second half of the year.

   Then walk it

   1. The structural flaw in annual budgets: by October you are steering with assumptions set the previous November. Decisions get anchored to a number that everyone privately knows is stale.
   2. A rolling forecast fixes the horizon rather than the calendar, so the planning quality does not decay through the year and you are never looking only three months ahead in Q4.
   3. It only works at a coarser grain. You cannot re-forecast 400 cost centres monthly and survive, so you re-forecast the 20 drivers that move the outcome and let the rest flow on ratios. That is the single most common implementation failure.
   4. Keep the budget for what it is good at: annual resource allocation, incentive targets and the board commitment. The rolling forecast is for steering. Companies that delete the budget entirely usually lose cost discipline.
   5. The cadence I would propose: monthly re-forecast of the current quarter, quarterly re-cut of the full rolling horizon, with a bridge each time from the previous version showing what changed and why. The bridge is what builds trust in the number.
   6. The honest cost is effort and forecast churn. If the number moves every month without explanation, business heads stop believing any of them, so version control and a change bridge are not optional.

   Where candidates lose it

   Describing the mechanics and skipping the granularity problem. Rolling forecasts fail because teams try to re-forecast at budget-level detail. Say that you re-forecast drivers, not line items, and keep the annual budget for targets.

   Expect next

   • What granularity would you re-forecast at?
   • How do you stop the number churning every month?
   • Would you keep the annual budget at all?
4. 036What is zero-based budgeting, and when does it actually work?Forecasting and budgetingIntermediatetechnicalCorporate FP&ABig Four

   Say this

   You build every cost line from zero and justify it against an activity, instead of taking last year plus inflation. It works on discretionary overhead where nobody owns the historical number, and it fails when applied to everything at once.

   Then walk it

   1. The mechanism: define cost packages, attach each to an activity and a service level, then fund them by priority against the available envelope. The question changes from 'why do you need more' to 'what would we stop doing'.
   2. Where it delivers: marketing, travel, consultancy, software licences, facilities, and the long tail of overheads that grew by accretion. Ten to 20 percent savings on those categories is a realistic outcome in the first pass.
   3. Where it does not: direct cost driven by volume, regulated or contracted spend, and anything where the answer is fixed by physics or law. Applying ZBB there wastes months to confirm the existing number.
   4. The real cost is organisational. It is slow, it is political, and it requires cost owners with the authority to say no. Done annually it collapses into incremental budgeting with extra paperwork.
   5. So my recommendation would be a rotating approach: full ZBB on a third of the overhead base each year, so every category is properly rebuilt every three years without the whole organisation stopping.
   6. And the failure mode to name: cutting cost without removing the activity. The cost comes back within two quarters, usually as contractors or as a service failure somewhere else.

   Where candidates lose it

   Selling it as a universal answer. Interviewers listen for whether you know it fits discretionary overhead, not volume-driven direct cost, and that cutting cost without removing activity just delays the spend.

   Expect next

   • Which cost lines would you exclude from ZBB?
   • How would you stop the cost coming back?
   • How does ZBB interact with a rolling forecast?
5. 038What is the difference between scenario analysis and sensitivity analysis, and when do you use each?Forecasting and budgetingIntermediatetechnicalCorporate FP&AFinancial modelling

   Say this

   Sensitivity moves one variable at a time to see which assumptions the answer is most exposed to. Scenario moves a coherent set of variables together to describe a state of the world. Sensitivity finds the levers; scenarios tell the story.

   Then walk it

   1. Sensitivity is mechanical: flex volume by plus or minus 10 percent, hold everything else, record the effect on EBITDA and on cash. Do that across ten assumptions and you get a tornado chart that shows you the three that matter.
   2. The point of sensitivity is triage. If a 10 percent move in raw material price changes EBITDA by 18 percent and a 10 percent move in overheads changes it by 2 percent, you know where to spend management attention and where to hedge.
   3. Scenario analysis is internally consistent by design. A recession case is not just lower volume, it is lower volume with discounting, worse mix, longer receivables, higher inventory and a tighter borrowing cost, all together, because that is how a recession actually arrives.
   4. So scenarios are for decisions and communication, sensitivities are for diagnosis. The board wants three scenarios with a probability view; the modeller needs the sensitivity table to know which three to build.
   5. Practically I run sensitivity first to find the drivers, then build two or three scenarios using only those drivers, and always include a breakeven case: how far can volume fall before we breach the covenant. That last one is the number treasury actually uses.
   6. The limitation of both is that they are still your assumptions. Neither tells you the probability of anything, and a Monte Carlo on made-up distributions just adds decimal places to the same guess.

   Where candidates lose it

   Using the two words interchangeably. The distinguishing feature is that scenarios move variables together and consistently. Also name the breakeven or covenant-breach case, because that is the version a CFO asks for first.

   Expect next

   • How many scenarios would you take to a board?
   • Would you assign probabilities to them?
   • Where does Monte Carlo add value and where does it not?
6. 039Walk me through a price, volume and mix variance.Variance and management reportingIntermediatetechnicalCorporate FP&ABusiness finance

   Say this

   You are splitting a revenue or margin gap into three causes: you sold at a different price, you sold a different quantity, or you sold a different blend of products. Price times budget volume, volume times budget price, and mix as the residual from the change in the product blend.

   Then walk it

   1. Price variance: actual price less budget price, times actual volume. Volume variance: actual volume less budget volume, times budget price. Mix variance: actual total volume times the difference between the actual and budget average price arising purely from the blend.
   2. The order matters because the terms are not independent. State your convention up front, price first at actual volume, and be consistent. Interviewers care far more about consistency than about which convention you pick.
   3. A worked case. Budget was 100 units at 1,000 rupees, so 1 lakh of revenue. Actual is 110 units at 950. Revenue is 1.045 lakh, so 4,500 favourable. Price variance is minus 50 times 110, so 5,500 adverse. Volume variance is 10 times 1,000, so 10,000 favourable. Net 4,500. The headline was positive and the price story was negative.
   4. Mix appears the moment you have more than one product. If you sold more of the low-margin SKU, total volume is up and average realisation is down, and mix separates that from a genuine discount.
   5. The reason FP&A lives on this: each variance has a different owner. Price belongs to commercial, volume belongs to sales and demand planning, mix belongs to the product and channel strategy. One number, three conversations.
   6. I would always run the same decomposition on gross margin, not just revenue, because a favourable volume variance on a low-margin product can be worse than no growth at all.

   Where candidates lose it

   Mixing conventions mid-answer so the three variances do not add back to the total. Say your convention first, then reconcile at the end. And do the margin version, not just the revenue version, because that is where the real conclusion sits.

   Expect next

   • Now do it for gross margin instead of revenue.
   • Who owns the mix variance in your organisation?
   • How would you present this to a sales director in one slide?
7. 042Walk me through a month-end close. What are you actually doing and where does it go wrong?Variance and management reportingIntermediatetechnicalCorporate FP&AGCC finance centres

   Say this

   Close is a sequence: cut off the subledgers, post accruals and provisions, reconcile, review flux, then report. A clean close is about five working days, and the two things that break it are late accruals and reconciliations left to the end.

   Then walk it

   1. Days one and two: cut off sales and purchases, post goods-received-not-invoiced accruals, run depreciation, accrue payroll and bonus, revalue FX balances, and book revenue cut-off entries.
   2. Days two and three: reconciliations. Bank, intercompany, GST recoverable against the portal, fixed asset register to the ledger, and inventory to the physical or cycle count. Intercompany is the usual culprit in a group with a shared service centre, because both sides must agree in the same period.
   3. Day three or four: flux review. Compare actual to budget, to forecast and to prior period for every material line, chase anything above a set threshold, and get an explanation with a name attached before anything is published.
   4. Day four or five: management reporting. The pack, the commentary, the variance explanations, the cash position, and the re-forecast if the cadence requires it.
   5. Where it goes wrong: accruals arriving after the flux review, so numbers move after commentary is written; a manual journal with no support; and the classic, an unreconciled intercompany difference parked in a suspense account for three months.
   6. The two controls I would insist on are a close calendar with named owners and cut-off times, and a rule that no journal is posted after the flux review without the controller's approval. That single rule takes a day out of most closes.

   Where candidates lose it

   Describing close as bookkeeping. The analytical part, the flux review before publication, is what an FP&A interviewer is listening for. Also name a concrete failure mode; 'sometimes things are late' is not an answer.

   Expect next

   • How would you take two days out of a five-day close?
   • What is the most common reconciliation problem in a shared service centre?
   • Who signs off the pack and what do they check?
8. 044How would you design the KPI set for a business unit that has never had one?Variance and management reportingIntermediatetechnicalCorporate FP&ABusiness finance

   Say this

   Start from the decisions the unit head makes weekly, work back to the drivers behind them, and pick the fewest metrics that cover outcome, driver and risk. Five to eight, each with an owner, a definition and a target. Any more and none of them get acted on.

   Then walk it

   1. One or two outcome metrics that the unit is judged on, for example gross margin in rupees and cash conversion. These are lagging and that is fine, they are the scoreboard.
   2. Three or four leading driver metrics that move those outcomes and can be influenced this week: pipeline coverage, on-time delivery, utilisation, realisation per unit, receivable days. Leading metrics are the only ones that let you intervene in time.
   3. One risk or quality metric to stop the drivers being gamed. Push utilisation without tracking attrition or rework and you get a short-term win and a longer-term problem.
   4. Every metric needs four things written down: the definition including the exact data source, the owner by name, the frequency, and the target with a threshold for action. A metric without a defined denominator will be argued about instead of acted on.
   5. Test each candidate against two questions: can the owner actually influence it, and would a red reading change a decision? Anything that fails both comes out. That usually removes half the first draft.
   6. The failure mode I would guard against is Goodhart's law. Whatever you measure gets optimised, including in ways you did not intend, so I would review the set after two quarters and check what behaviour it produced, not just whether the numbers improved.

   Where candidates lose it

   Producing a long list of financial metrics. The interviewer wants the outcome, driver and risk structure, the fact that drivers are the actionable ones, and an owner and definition for each. Mentioning gaming risk sets a good answer apart.

   Expect next

   • Give me a KPI set for a warehouse operation.
   • How do you stop a KPI being gamed?
   • How often should the set change?
9. 049How are margins and operating leverage at the company, and what does high operating leverage mean for you as an analyst?Unit economics and costingIntermediatetechnicalMoody'sCorporate Finance · New York · 2018

   Say this

   Operating leverage is the share of the cost base that is fixed, and it decides how violently profit moves when revenue moves. High operating leverage means a small revenue change becomes a large EBIT change, in both directions.

   Then walk it

   1. The measure is the degree of operating leverage: percentage change in EBIT over percentage change in revenue. It equals contribution divided by EBIT, so a business with 40 crore of contribution and 10 crore of EBIT has a DOL of four.
   2. So at a DOL of four, revenue up 10 percent gives EBIT up 40 percent. Revenue down 10 percent gives EBIT down 40 percent. That symmetry is the whole point, and analysts routinely model the upside and forget the downside.
   3. High leverage sits with cement, steel, hotels, telecom, airlines, exhibition and any asset-heavy business. Low leverage sits with distribution, trading and staffing, where cost follows revenue almost one for one.
   4. For a credit view, operating leverage and financial leverage compound. A cement company at four times operating leverage and three times net debt to EBITDA converts a mild demand slowdown into a covenant breach. I would never assess one without the other.
   5. What I actually do with it: build the cost base into fixed and variable, compute the revenue decline that takes EBIT to zero, and compare it against the worst historical peak-to-trough volume decline in that industry. That is a far better risk statement than a margin forecast.
   6. The caveat is that fixed costs are only fixed for a while. Management cuts discretionary spend in a downturn, so realised downside leverage is usually a bit better than the arithmetic, and realised upside leverage a bit worse because of wage and maintenance catch-up.

   Where candidates lose it

   Defining operating leverage as 'high fixed costs' and stopping. Give the contribution-over-EBIT measure and a number, then insist on the downside case. Interviewers at rating agencies are specifically testing whether you compound operating with financial leverage.

   Expect next

   • What revenue decline takes this company's EBIT to zero?
   • How does operating leverage interact with financial leverage?
   • Which sector in India has the highest operating leverage?

   Reported by candidates at Moody's (Corporate Finance, New York, 2018). Source: Wall Street Oasis.

10. 054NPV and IRR disagree on which project to pick. Which do you follow, and why does IRR mislead?Capital budgetingIntermediatetechnicalCorporate financeCorporate FP&A

    Say this

    Always NPV. It measures value created in rupees at your actual cost of capital. IRR is a rate, and a rate tells you nothing about size, assumes you can reinvest every interim cash flow at that same rate, and can have more than one solution or none.

    Then walk it

    1. The scale problem: a project returning 60 percent on 10 lakh creates 6 lakh of value. One returning 18 percent on 5 crore creates far more. IRR ranks the first higher and shareholders prefer the second.
    2. The reinvestment assumption: IRR implicitly assumes every interim cash flow is reinvested at the IRR itself. If the project shows 35 percent and your genuine reinvestment opportunity is 12, the 35 percent is fiction.
    3. The timing problem: IRR favours projects that return cash early, which systematically disadvantages long-dated infrastructure or R&D projects that create more total value.
    4. The mathematical problem: a project with alternating cash flow signs, such as a mine with a large end-of-life restoration cost, can have two IRRs or none. NPV always has exactly one answer.
    5. And it ignores the cost of capital except as a hurdle. Two projects at 20 percent IRR are not equivalent if one is a domestic capacity expansion and the other is a greenfield in a new country with a different risk profile.
    6. Where IRR is genuinely useful is communication. Business heads and boards think in rates, and comparing a rate against the cost of capital is intuitive. So I would present both, decide on NPV, and use MIRR if the reinvestment distortion is large.

    Where candidates lose it

    Saying IRR is wrong without naming the mechanism. You need the reinvestment assumption and the scale problem specifically. Also do not dismiss IRR entirely, because every real investment committee uses it and saying so shows commercial sense.

    Expect next

    • How does MIRR fix the reinvestment problem?
    • When would you get two IRRs?
    • How would you rank projects under a fixed capital budget?

Firm tags come from public, anonymous candidate reports on Wall Street Oasis: strong signal, not sworn testimony. Firms are named as the places a question was reported, not as partners of Fin Maverick. Answers are written for this page to show how to think out loud; they are not scripts to recite.