Financial analyst interview questions from the hiring side

Explore the Finance interviews hub and the finance careers guide.

Almost every financial analyst interview question is asking the same thing. Can your manager send your number upstairs without redoing it first?

I've sat on the hiring side of these loops, and I've managed the analysts who got through them. The technical questions and the Excel test check the mechanics. The behavioural questions check the part that decides most offers: whether you catch your own mistakes and can say what a number means. (For the rounds every finance role shares, see the finance interviews guide.)

What is a financial analyst interview really testing?

It's testing whether a number can be trusted by someone who will act on it.

Most of the analysts I hired were great at the number. The coaching was almost always about what the number meant, and whether anyone had checked it before it left the desk.

The U.S. Bureau of Labor Statistics says financial analysts research and evaluate financial data, forecast future trends, and prepare reports containing recommendations. The recommendations are the part most candidates skip.

When I've had to pick between two good analysts, it came down to two things. First, someone who thinks along the same lines I do, almost a younger me. Second, one strength I really like.

Past those two, the yes is an attitude. Nothing irritates me more than somebody super smart who assumes a super smart answer means everything goes according to plan. It's naive. This is finance, this is numbers, and nothing goes according to plan. Be ready for fires and pushback.

The best analysts I've hired showed it in the first month. They tried to figure out the nuances of our problems before they started fixing anything. The really good ones would say, in effect, this problem is actually stupid and it's fixable, but I want to understand the history behind it.

"Walk me through the three statements"

Keep it under a minute, then stop talking.

Start with net income, which opens the cash flow statement and rolls into retained earnings. Add back the non-cash items, depreciation being the usual one, then adjust for working capital. Capex goes in investing, and debt and equity moves go in financing. Ending cash lands on the balance sheet, and that's the whole answer.

For a refresher the night before, the SEC's beginners' guide to financial statements covers all three in a few pages, and the FP&A interview guide explains why the lecture version loses points.

"If depreciation goes up by $10, what happens?"

This one separates people who memorized the three statements from people who understand them.

Assume a 25 percent tax rate and say it out loud, in order, one statement at a time.

Then say why cash went up, because that's what they're listening for. Depreciation isn't cash, but it lowers taxes, and taxes are.

Practise two more versions (a $10 inventory write-down and a $10 increase in accounts receivable).

"How strong is your Excel?"

Give the rating if they ask, then show what you've built.

I've been asked this one straight, and I said 9 out of 10. That's fine as a headline, but nobody can check a rating, and they can check what you've built.

Many analyst loops include a short Excel test, usually a raw data file with asks like these.

• Turn a transaction dump into a summary by month and by department.
• Build a pivot, then a lookup that pulls a budget number next to each actual.
• Calculate a variance in dollars and in percent.
• Sometimes, link a small income statement to a balance sheet.

If your lookups are rusty, Microsoft's XLOOKUP guide is the quickest refresher.

The check row is the step that gets dropped under time pressure, and it's the one that saves you when whoever built the test left a duplicate row in the source file.

When I open a test file, I'm asking one thing, whether I could follow it without calling the person over to explain it.

The test also checks whether the build changes when the problem does. Nobody wants a one-trick pony on a modelling team.

"How do you make sure your numbers are right?"

Bring a habit, not a promise.

I'll be honest with you. For a while my checking process at close was staying late and verifying everything, and then the next morning we'd catch our mistakes.

That's a routine, but it isn't a check. It just moves the mistake to the morning.

What changed was the start of the close, not the end. Before heading into it, I check that I have everything I need to finish, and if something's missing, I flag it early, chase it down and plan for it not arriving, usually with an estimate I can accrue. If the close still runs late, it's communicate, communicate, communicate.

Being audited helped too. Every quarter the auditors would look at the journal entries, pick a few and make me back them up with evidence. I had to stand behind everything.

You build a number differently once you assume someone will ask you to prove it.

The other habit is knowing roughly what the answer should be before you look. One forecast night, before I trusted a total, I held it against the number I remembered from before any manual changes. "Are you sure of these numbers?" Then, "just double check." My memory was right.

So name the checks you run, in order. When I'm the one asking, an order tells me there's a habit.

The first step covers definitions as well as totals. I once had an expense slide that included allocations, so the total tied to the model while the definition didn't. My question was "Does it tie to my expense view in consolidated?" The model showed direct expenses, so we took the allocations out of the slide to match it.

A total that ties can still be the wrong number.

When a wrong number gets out anyway, early bad news beats late good news. If I catch it and it's material, I send a follow-up: please ignore that last deck, this is the actual right number, this is what happened, sorry about that, won't happen again. Then I go fix the root cause. Leaving people with the wrong number because you don't want to look bad is worse than that email. Nobody bats an eye at the follow-up unless it's a repeat offence. Mistakes happen here and there.

"Tell me about a time you worked with messy data"

They're checking whether you ask where a number came from before you use it.

My favourite example comes from due diligence on a crypto business. Our team was approaching it like a traditional business, almost taking what the other company said as fact. When they heard the word wallet, they assumed it was secure, like money at a bank.

So I asked the plain questions. What do you mean wallet? Is it hot? Is it cold? Is it multi-sig? Who's the final authority?

We didn't want to buy a company whose balance sheet was on a USB stick in somebody's drawer.

The projections got the same treatment. I backed into them and found they only worked if the bull market continued, and their treasury was almost all Bitcoin and Ethereum, with very little cash. I'm a crypto enthusiast myself, but prices come down, and the valuation comes down with them.

It happens inside your own files too. One forecast week, I noticed a revenue view still on an old constant currency method, the way you strip out exchange-rate moves, that didn't match the method in the EBIT flash, our early read on profit. Same measure, two methods, two answers.

Before comparing two numbers, check they were built the same way. Then tell the story in this shape.

Step 4 is the one people forget. A cleaned-up file is nice, but a decision that went differently because you cleaned it is the answer, and if nothing changed, say what would have gone wrong if nobody had looked.

"Explain a financial concept to someone outside finance"

They want to know if people outside finance will act on your numbers.

The analysts I trusted most did one thing the others didn't. Every number got one line underneath it, saying what it means and what the reader should do about it, and I coached that line more than I ever coached a formula.

What I still haven't figured out is telling which analysts actually want that kind of work. Some analysts want to come in, do their variance analysis and their reports, and go home. That's fine. We also need worker bees. The hard part is spotting who's both good and willing to take it to the next level, and finance is mostly introverted, so you can't tell on face value. Sometimes I force it on the wrong person. Sometimes I miss someone. If you want that work, say so in the interview.

If they ask you to explain working capital or EBITDA, do the same thing. One plain sentence on what it is, one on why a manager would care this month, then stop.

"Tell me about a time you had too many deadlines"

They're checking whether you sort requests or just work faster.

Everything arrives labelled urgent, so the sorting question is when they're presenting it. (Where that question comes from is in the FP&A version of this question.)

For an analyst, say who you told.

The other kind of fire is a missing input. I once had a deck due while the source files for two slides were still with someone else. My call was "let's take what we have right now" and build everything that didn't depend on those files, with the two slides marked as waiting.

Walk me through a simple case

The setup counts for more than the math.

Some analyst loops include a short case, like when a product breaks even.

A product sells for $50, it costs $30 in variable cost to make, and fixed costs are $2 million a year. The ask is how many units it takes to break even.

Contribution per unit is $50 minus $30, which is $20. Divide $2 million by $20 and you get 100,000 units, using the same formula the U.S. Small Business Administration uses: fixed costs divided by price minus variable cost.

Now the part that scores. Before trusting that number, say what you'd check, like whether the $30 is really variable at every volume and whether the $2 million steps up before 100,000 units. Then say what you'd tell the business: we need 100,000 units to break even, we sold 80,000 last year, so the plan needs a 25 percent volume jump or a price change.

I've had a clean number fail that check myself, on a small business idea. I thought offering bookkeeping would be an easy sale. Everything looked great on paper but as soon as I picked up the phone, I could tell after 15 dials that this wasn't going to work.

The math was fine. The assumption under it was wrong, and only a real conversation could show me that.

So in a case, name the assumption you'd test first, and say how you'd test it.

Getting to the answer too fast is the other way to lose a case round. I know because I did it, and what the interviewer told me afterward is worth reading before your loop.

What changes in a senior financial analyst interview?

The technical questions stay about the same, and what changes is ownership.

It shows up as three questions, asked directly or not.

• The process you own end to end, one recurring report or forecast that runs because you run it.
• The people you partner with outside finance, a business leader who calls you before they make a decision, not after.
• The person you've trained, one junior analyst whose work got better because of you.

Show one report you run every month and one thing you changed so it runs better than when you got it, with a count if you have one.

If you're moving toward FP&A from here, the FP&A careers hub covers what that next seat looks like. If the ledger side pulls you more, read how controller interviews test the close.

More financial analyst interview questions

These come up less often, and each one wants one true, specific line.

• "Why do you want to be a financial analyst?" Name the part you enjoy. My own answer has always been the part after the number, understanding what drove it and using that to help the business decide what to do next.
• "What's the difference between a budget and a forecast?" The budget is the number you agree to hit for the year. A forecast is where you think you'll actually land, and it gets updated.
• "What's EBITDA, and why do people use it?" Earnings before interest, taxes, depreciation and amortization, used to compare operating performance before financing and accounting choices get in the way. Mention that it isn't cash.
• "Walk me through a model you built." Open with the decision it was built for, then the drivers, the output, and who used it.
• "Tell me about a mistake you made." Pick a real one you caught yourself, and say what check you added so it doesn't happen again.
• "How do you handle a manager who wants a different number?" They're checking whether you can hold a line politely. I don't just say no and drop a different number. I show them what would need to be true for their view to hold, and then I let the analysis do the arguing for me.
• "What are your salary expectations?" Know the going range before the call. If there's a posted range, say you're okay with it and where in it you'd put yourself. Don't name one number first, because they'll anchor to it. I pulled the pay ranges from 1,262 real US FP&A postings, split by level.

What should you ask at the end of a financial analyst interview?

Ask questions that show you're already picturing the work.

One I've asked in my own interviews is what prompted the hiring, and if it's a replacement, how the mandate has changed. You find out whether you're filling a gap or building something new, and those are very different first 90 days.

• Ask who reads the reports this seat builds, and what they do with them.
• Ask how much of the week is recurring reporting and how much is new analysis.
• Ask what a good first 90 days looks like. (Have your own 30-60-90 day plan ready in case they turn it around.)
• Ask what the last person in this seat moved on to.

That last one tells you whether the seat goes anywhere.

How do you prepare for a financial analyst interview in one evening?

You need 5 true stories, one flow-through question and 20 minutes in Excel.

Keep each story to 5 lines, with at least one number in each.

Then say the depreciation flow-through out loud until you can do it without looking. When I prep an answer, I want the why first, then one good example, and then I play it over, question it and grade myself.

After that, rebuild one pivot and one lookup from a blank file.

Most analyst questions are one of these 5 stories asked a different way, and I'd rather hear one true story with a check in it than five polished ones without.

Check the number, then say what it means.