Finance, Accounting & Insurance

How to get hired as a financial analyst in 2026-27

Updated 5 October 2026 · 26 min read · by SkillSculptor

The short answer

To get hired as a financial analyst in 2026 or 2027, first decide which version of the job you are applying for, because "financial analyst" covers corporate FP&A, investment banking, equity research, commercial credit, treasury and government budget analysis, and those are hired by different people against different tests. No license is required for corporate or FP&A analyst work: the practical floor is a bachelor's degree in a quantitative, accounting or business field plus demonstrable modeling ability, and securities licenses such as the FINRA Series 79 come only after a member firm hires and sponsors you. The stage that decides the hire is a modeling or case exercise, usually a timed Excel build, a budget-versus-actual variance pack, or a 24 to 72 hour take-home with a one-page recommendation, graded on whether the model balances, whether cash and debt are linked correctly, whether assumptions are separated from formulas, and whether you answered the business question instead of only building. Your resume exists to earn that exercise, so state what you forecast, at what scale, in which system, on what cadence, and what decision your number changed.

License requiredNone for corporate finance, FP&A, treasury or government budget analyst roles. Securities licenses gate specific markets-facing seats: an investment banking analyst usually needs the SIE plus FINRA Series 79, a sell-side research associate needs Series 86 and 87, a general securities representative needs Series 7, and most of those also need the Series 63 state exam. Every one except the SIE requires association with a FINRA member firm, so you sit them after you are hired, inside whatever window your firm's policy sets (commonly about 90 to 120 days). The SIE and the Series 65 can be taken with no sponsor if you want to prove intent before you apply.
DegreeA bachelor's degree is the practical floor: finance, accounting, economics, mathematics, statistics, engineering or any business field. Transcripts are rarely checked outside campus pipelines. An MBA matters for a move into banking or a step up to manager, not for a first analyst seat. A master's in finance helps mainly as a career changer's reset and as an on-ramp to a recruiting calendar or a work visa.
CFA, concretelyThree exams taken in order, administered by CFA Institute, plus required practical skills modules at Levels I and II. Level I can be taken in your final undergraduate year. The charter additionally requires 4,000 hours of relevant work experience completed over a minimum of 36 months, plus membership and sponsorship. Level III now has specialized pathways in portfolio management, private markets and private wealth. Realistic time to charter while working full time is three to four years. CFA Institute publishes the pass rate for every level and sitting, and the current fee table: read both on cfainstitute.org rather than a forum, and budget four figures per sitting plus a one-time enrollment fee.
Where the CFA pays and where it does notIt is close to a requirement in sell-side research, asset management, portfolio and investment analysis, and much of credit and private markets. In corporate FP&A it is neutral to mildly positive and will not get you hired on its own. The credentials FP&A employers actually name are the CPA, the IMA's CMA, and AFP's FPAC.
The deciding stageA modeling or case exercise. Four common formats: a timed in-office or screen-shared Excel build with internet and AI tools blocked, roughly 60 to 180 minutes; a 24 to 72 hour take-home model plus a one-page memo; a paper LBO or mental-math screen for banking and private credit, done with no computer; or an existing model handed to you to walk through, critique, or find the planted error in. Supervised formats have become more common precisely because an unsupervised take-home no longer proves much.
Typical loopCorporate FP&A: recruiter or HR screen, hiring manager, an Excel or case exercise, a panel with the business partners you would support, decision. Two to five weeks. Investment banking: campus or recruiter screen, then a superday of back-to-back technical interviews, sometimes compressed into days. Commercial credit at a bank: screen, credit manager, a spreading or credit-memo exercise, then often a cohort start date tied to a formal training class. Government budget analyst: a posted vacancy, a scored rating of your written application, a structured panel, and months of elapsed time.
Who screens youFirst a recruiter or HR generalist matching nouns: Excel, modeling, forecasting, the ERP and planning system by name, the credential. Then the person whose numbers you would own: an FP&A manager, finance director or controller, or in investment seats a senior analyst or portfolio manager. The business-partner panel, a sales VP or a plant manager, is often the one who vetoes, because they are testing whether you will be useful rather than merely accurate.
Where to check payUS Bureau of Labor Statistics Occupational Employment and Wage Statistics, SOC 13-2051 Financial and Investment Analysts, published nationally and by state, metro area and industry. Use the right adjacent code if your title differs: 13-2031 Budget Analysts, 13-2054 Financial Risk Specialists, 11-3031 Financial Managers. Then read posted ranges in pay-transparency jurisdictions, and AFP's compensation survey for FP&A and treasury. In banking and on the buy side, base salary is the smaller half: ask for the target bonus as a percentage of base and the last two years' actual payout against target.

"Financial analyst" is at least six different jobs. Identify which one the posting means

The title is one of the least informative in hiring. A corporate FP&A analyst at a manufacturer, an investment banking analyst, a sell-side research associate, a commercial bank credit analyst, a treasury analyst and a government budget analyst all answer to "financial analyst", and they are hired through almost unrelated processes, by different people, against different tests. Applying to all of them with one resume is the most common reason a qualified candidate gets no replies.

You can identify the variant in under a minute from the posting's nouns. The vocabulary is not decoration; it is the hiring manager describing the work they actually do. Match the top third of your resume to that vocabulary and nothing else.

Two things to know about volume and about where the mythology lives. By headcount, corporate FP&A and credit analysis dwarf the investment banking and research seats that dominate the internet's advice, and they are the variants where school pedigree matters least. And the pipeline myth runs one way: an FP&A analyst can move into corporate development and a credit analyst can move into private credit, but the banking-to-anything path that recruiting forums assume is specific to a small, campus-recruited slice of the market. Pick the variant you can actually get into, then move.

How financial analyst hiring actually works in 2026-27

The common spine is short: a screen, the manager, an exercise, a panel, a decision. What varies is the calendar, the channel, and what the exercise is for.

Corporate FP&A hiring is manager-led and budget-cycle driven. Requisitions cluster when the annual plan is being built or when someone leaves mid-cycle and the close cannot absorb the gap, so the urgency is real and the loop is often two to four weeks. The recruiter screen is a keyword match. The manager conversation is about scope: what you owned, against how much revenue or spend, how many business partners, and whether you produced numbers or produced decisions. The exercise is usually Excel. The panel almost always includes a non-finance business partner, and their veto is about usefulness rather than technique.

Two filters now sit in front of the human stages at larger employers. An online timed Excel or numerical-reasoning assessment sent with the application, which you should treat as a real gate and practice for, and in banking and on the buy side a recorded video interview with fixed questions and no interviewer. Neither is where you win the job, and both are where candidates quietly get dropped.

A large and under-noticed share of FP&A seats are filled contract-first through finance and accounting staffing firms, then converted. If you are changing careers this is frequently the shortest route in: a three-month contract covering a close or a system implementation puts your work in front of the manager who owns the requisition, and the conversion conversation replaces the interview loop. The trade is no benefits and no notice protection for the contract period. Ask at the start what the conversion history on that specific team has actually been.

Investment banking runs on a recruiting calendar rather than on need. Internship recruiting for the following summer starts absurdly early, full-time offers mostly convert from internships, and off-cycle and lateral hiring runs through specialist recruiters. If you are outside the calendar, your realistic targets are middle-market and boutique banks, valuation and transaction advisory firms, and corporate development, all of which hire when they have work.

Credit analyst hiring at banks frequently runs through a formal credit training program with a cohort start date, which means a hard application window and a slow decision, but structured teaching once you are in. These programs take candidates with accounting coursework and no banking background, which is exactly why they are worth targeting.

The public sector is a different machine. A posted vacancy, a self-rating questionnaire, a numeric score, a referral list, then a structured panel where everyone is asked identical pre-written questions. Nothing you say informally to a hiring manager changes the score. The lever is writing your application against the posted criteria, literally and in their words, and the elapsed time from posting to start date is commonly measured in months.

The modeling test: what it actually asks, and how it is graded

This is the stage that decides the hire, and most candidates prepare for the wrong part of it. They practice building faster and neglect the two things graders mark hardest: whether the model is internally consistent, and whether you answered the business question.

The most common corporate exercise is a short three-statement or driver-based build. A typical shape: here are two years of actuals and a set of assumptions, build a 12-month or three-year forecast, flag the risks, and tell us the one thing you would ask management. The most common FP&A exercise is a variance pack: here is budget versus actual by cost center, explain the gap. The most common banking screen is a paper LBO with no computer. The most common senior-analyst exercise is an existing file you are asked to walk through, critique, or find the planted error in.

What graders check, in roughly this order. Does the balance sheet balance, as a visible check row rather than as a claim. Does ending cash on the cash flow statement tie to the balance sheet. Is interest driven off a debt schedule, and if that creates circularity, did you handle it deliberately with an iterative-calculation toggle or a circuit breaker rather than by hardcoding a number. Are inputs separated from calculations, with one consistent convention for hardcoded assumptions. Is each formula consistent across its row, so a reviewer can check one cell and trust the other eleven. Are units and sign conventions stated and obeyed. Is there an error-check row that sums to zero. And finally: did you write the two or three sentences of recommendation they asked for, or did you hand back a spreadsheet and leave the thinking to them.

What loses offers, in rough order of frequency. Spending the first twenty minutes on formatting and running out of time. A hardcoded plug that forces the balance sheet to balance. Formulas that differ along a row. No assumptions area, so a reviewer cannot flex your case. Clicking cell by cell with a mouse while a banker watches. And the quiet killer: a technically clean model with no point of view, submitted by someone who never said which number they were least confident in.

On the paper LBO, the skill being tested is arithmetic fluency under pressure, not insight. Build the muscle: sources and uses, entry enterprise value from an EBITDA multiple, a simple debt schedule paid down with free cash flow, exit at a stated multiple, equity at exit, and an IRR you approximate rather than compute. Memorize the compounding anchors, because they are what fluency sounds like: roughly 2.0x your money over five years is about a 15% IRR, 2.5x over five years about 20%, 3.0x over five years about 25%, and 2.0x over three years about 26%. These are just compound growth, so check them yourself once and then trust them. If you can land within a point of those with no calculator, you sound like someone who has done this.

How to practice, concretely, for free. Pick three public companies in different industries, pull the 10-K from SEC EDGAR, and build a three-statement model from a blank sheet, timed, start to finish. Do it again a week later on the same company without opening your old file. The second build is where the learning is. Then write a one-page memo on each: what you forecast, the two assumptions that matter most, what would have to be true for you to be wrong, and what you would ask management. That memo is also the artifact that substitutes for a target school on your resume. Paid courses from the established modeling training providers are fine and will save you time on conventions, but no certificate from one has ever been the reason someone was hired. The test is the reason.

Getting in without a target school

First, size the problem honestly. School pedigree is a real filter in a narrow band: bulge-bracket and elite-boutique investment banking, a handful of large asset managers and hedge funds that recruit on campus, and the private equity pipeline fed by those. Outside that band, for corporate FP&A at mid-market and large companies, for commercial credit, for valuation and transaction advisory, for insurance, for treasury, for government, for middle-market banking and for corporate development at most operating companies, the hiring manager cares about what you can do in Excel and whether their business partners will trust you. Many of them did not attend a target school either.

Second, stop applying exclusively to postings titled Financial Analyst. The reliable routes in are mostly adjacent.

The internal move is the single highest-probability route and the most neglected. If you are already inside a company in accounting, operations, sales support, billing, procurement or data, you have the thing an external candidate cannot buy: knowledge of the business, and a manager who can vouch for your numbers. Make the move deliberately. Ask to own the forecast for one cost center or one product line, build the variance pack nobody wants to build, present it once to a business partner, then tell the FP&A manager you want the next opening. Twelve months of that beats any certificate.

Third, make the artifact. For a non-target candidate the model-plus-memo is the closest thing to a credential: a three-statement model of a real public company and a one-page written recommendation, as a PDF you can attach and an xlsx you can hand over. Put one line and a link in your resume header. It converts the question from "where did you study" into "walk me through your assumptions", which is the conversation you want.

Fourth, network where this profession actually gathers, which is not a LinkedIn comment section. Local CFA Society chapters run events open to candidates, not only charterholders. AFP chapters are full of exactly the FP&A managers who hire. Both have members who will take a twenty-minute call about a model you built, because talking about a model is the one cold-outreach topic a finance person does not find tedious.

A word on the entry-level squeeze, stated without inflation. Entry-level finance has got harder to enter in the usual two ways: more applicants per posting, and more employers asking for one to two years of experience for jobs that used to take none. What has not happened is the work disappearing. The jobs hardest to fill right now are the unglamorous ones, credit analysis at regional banks, FP&A at manufacturers and healthcare systems, budget analysis in the public sector, cost accounting, which is precisely why they are the ones to target first.

Credentials and licenses: which ones pay, and which ones are training in a trophy case

No credential is required to be a financial analyst in a corporate finance team. Securities licenses gate specific seats, and almost all of them require an employer to sponsor you, which means they come after the offer. Everything else is a choice about which doors you want opened.

The CFA is the most consequential and the most mis-sold. It is three sequential exams from CFA Institute, with required practical skills modules at Levels I and II, and the charter additionally requires 4,000 hours of relevant work experience over a minimum of 36 months plus membership and sponsorship. Level I can be taken in your final undergraduate year. Level III now offers specialized pathways in portfolio management, private markets and private wealth. Realistically it is three to four years to charter while working. Before you spend the money, answer one question with evidence: read thirty postings in the exact variant and geography you want, and count how many name it. In research, asset management and credit you will find it constantly. In corporate FP&A you will find it occasionally, as a preference. Candidacy itself carries signal, because "CFA Level II candidate, 2027" tells an investment employer you are serious, but only until you stall, at which point a three-year-old Level I pass reads as an abandoned project.

For corporate finance the ranking is different. The CPA is the strongest, especially where the role touches technical accounting, close, audit support or a controller track; note that a number of states have added a bachelor's-plus-experience licensure pathway alongside the traditional 150-credit-hour route, so check your own state board rather than advice written a few years ago. The IMA's CMA is well targeted at management accounting, costing and FP&A: two exams, a bachelor's degree and two years of relevant experience. AFP's FPAC is the only credential built specifically for FP&A, and it is named often enough in postings to be worth considering once you already hold the job.

Modeling certificates from commercial training providers are training, not signaling. They are a reasonable way for a non-target or career-changing candidate to learn conventions quickly, and worth the money for that. They are not why anyone gets hired, and listing one as though it were a license invites a harder modeling test. Spend the money if you need the teaching, and expect the test to still decide.

On FINRA licenses, the practical details. The SIE can be taken by anyone with no sponsor, and passing it before you apply is a cheap, legible signal for a markets-facing seat; your pass stays valid for four years. The Series 79 for investment banking, Series 7 for general securities, Series 86 and 87 for research and the Series 63 state exam all require association with a member firm, and firms typically expect you to pass within a set window after you start. The Series 65 can be taken with no sponsor if you are heading toward an investment adviser representative role. Exemptions and waivers exist, including a research-exam exemption tied to CFA progress, and the conditions matter, so read FINRA's current rule rather than a forum summary.

The resume: what belongs on it and what is read as noise

A financial analyst resume has one job: convince a stranger to spend 90 minutes giving you a modeling test. It does that by making the scale and the ownership of your work unambiguous. Most analyst resumes fail because they describe a function rather than a scope, and every candidate's function sounds identical.

Replace the verb list with a fact block per role, in a fixed order so a reader can compare: what you forecast, at what scale, in which system, on what cadence, with how many business partners, and what decision your number changed. Then one or two outcome lines with a number you can defend under questioning. Every example below is a shape, not a benchmark; use your own figures and expect each one to be probed.

On scale, give units a stranger can calibrate. "Owned the OpEx forecast for a $180M revenue business unit, four cost centers, 11 cost-center owners, monthly reforecast on a D+5 close" tells a hiring manager more than any adjective. If you cannot disclose the company's numbers, give the shape and say why: "mid-nine-figure revenue, private, confidential".

On systems, mirror the posting's exact words. If it says Workday Adaptive Planning, write Workday Adaptive Planning, not "planning software". Name the ERP (NetSuite, SAP, Oracle, Dynamics 365, Sage Intacct), the planning tool if there is one (Adaptive, Anaplan, Pigment, Planful, OneStream, Vena, Hyperion), the BI layer (Power BI, Tableau, Looker), and SQL if you genuinely write it. Spell out and abbreviate the big terms once, as in "financial planning and analysis (FP&A)", because recruiter searches use both forms.

On accuracy claims, be precise or silent. "Improved forecast accuracy" is unverifiable and invites the question that sinks it. "Reduced revenue forecast error from 8% to 3% on a one-quarter horizon, measured monthly against actuals" states the metric and the horizon, which is what a finance manager needs in order to believe it.

On what gets ignored: an objective statement, "detail-oriented", "proficient in Microsoft Office", "advanced Excel" with nothing demonstrating it, a twenty-item tool salad, soft-skill paragraphs, and responsibilities phrased as "assisted with". GPA is the exception worth getting right by variant: include it in campus and banking pipelines if it is strong, drop it once you have two or three years of work, and never include it for a mid-career FP&A role.

One page until roughly eight years in. Reverse chronological. No photo, no graphics, no skills bar charts, no two-column layout that an ATS will read in the wrong order. And name the model types you have actually built, such as three-statement, driver-based revenue, headcount, 13-week cash flow, DCF, LBO, accretion and dilution, because naming one is an invitation to be asked about it, and that is the invitation you want.

Pay, and how to find a number you can actually rely on

Pay for this title spans a range so wide that any single quoted band misleads, because the variants compete in different labor markets. A government budget analyst, an FP&A analyst at a mid-market manufacturer, a credit analyst at a regional bank and a first-year investment banking analyst do not price against each other at all. Use sources rather than bands.

Start with the US Bureau of Labor Statistics Occupational Employment and Wage Statistics. The code covering most of this work is 13-2051, Financial and Investment Analysts, published nationally and by state, metro area and industry, which is the breakdown that matters: the same title pays differently in commercial banking, in manufacturing and in local government. If your actual title sits elsewhere, use the right code: 13-2031 Budget Analysts, 13-2054 Financial Risk Specialists, 11-3031 Financial Managers for the step up. BLS figures lag and exclude bonus, so treat them as a floor and a geographic ratio rather than as your offer.

Then read posted ranges. Pay-transparency laws require a range in the posting in Colorado, California, Washington, New York, Illinois, Minnesota, Maryland, Massachusetts, New Jersey, Vermont, Hawaii and the District of Columbia among others, and because large employers post nationally, those ranges tell you what a company pays for your variant even if you are applying elsewhere. Twenty postings in your variant and metro is a better estimate than any aggregator.

Then the specialist sources. AFP publishes a compensation survey focused on FP&A and treasury roles. Staffing firms publish annual salary guides, useful for the shape of a market with the caveat that they reflect the roles those firms place. Public employers publish actual schedules: a federal posting states a GS grade and locality table, and state and municipal scales are published line by line, which makes those offers the most predictable in finance and the least negotiable.

In banking and on the buy side, base salary is the smaller and more public half of compensation. Recruiting-firm surveys circulate base and bonus ranges every year; use them for shape only. The questions that get you a real number are specific: what is the target bonus as a percentage of base, what did this team actually pay out against target for the last two years, when is it paid, and is any of it deferred or clawback-eligible.

What consistently moves an offer for this role: the variant and the industry, the scale of what you have forecast or underwritten, demonstrated system ownership (having been the finance lead on an ERP or planning implementation is worth more than most credentials), SQL or Python you actually use on the job, the CFA in investment seats, and geography. What moves it less than candidates expect: years of experience past about five in the same variant, and credentials unaccompanied by scope.

The interview, and the late-stage checks that derail finance hires

There are four interview strands, and candidates over-prepare the first and under-prepare the fourth: accounting and technical mechanics, the case or model, the business-sense question, and the question about a time your number was wrong.

The technical bank is finite and knowable. Walk me through the three statements and how they connect. If depreciation increases by 10, what happens to each statement. Why is EBITDA not cash flow, and when is it most misleading. What happens to free cash flow if DSO increases by five days. Walk me through a DCF, and where are you most likely to be wrong in it. How do you get from enterprise value to equity value. What is deferred revenue and where does it show up. Walk me through a paper LBO. For FP&A: what is in a variance bridge, how do you build a driver-based forecast, what is the difference between a budget, a forecast and a latest estimate, how do you build a 13-week cash flow, what is in your close calendar. For credit: how do you spread a borrower, what covenant would you set and why, what is the difference between leverage and coverage, when would you decline a profitable borrower. None of these are trick questions. They check that you have done the work, and fluency is the whole signal.

The business-sense question is the discriminator, and it usually arrives as a scenario: revenue came in 4% under plan, you have the file, what do you look at first. The answer that gets hired has a shape. Decompose before you theorize: price, volume, mix, timing, FX, one-offs. Compare against the driver that was forecast rather than against last year. Isolate whether it is a miss or a shift. Then say who you would call and what you would ask them, because the data will not tell you whether a large customer delayed an order. A candidate who starts naming causes before decomposing is telling the interviewer how they will behave during a close.

If you are interviewing for research or the buy side, prepare a stock pitch and expect to deliver it in two minutes: what the business is, what you believe that the market does not, what the valuation implies, what the catalyst is, what would prove you wrong, and how large a position you would take. Pitching a consensus mega-cap long with no variant view is the most common way to fail this.

The behavioral question that decides more of these interviews than any technical one: tell me about a time your number was wrong. Every working analyst has published a wrong number. The answer must name the error, name how it was found, name who you told and how fast, and name what you changed in your process so it cannot recur. Candidates who cannot produce one are either inexperienced or dishonest, and experienced interviewers read it that way.

What gets people rejected, in rough order of frequency: being unable to explain a number on your own resume; a model that does not balance; no point of view, having built something and formed no opinion about it; treating finance as reporting rather than as decision support; no curiosity about how the business makes money, which shows up instantly in the questions you ask at the end; blaming systems or other teams for a bad number; and answers that sound generated, meaning fluent, general and free of specifics. That last one has become its own rejection reason, which is worth knowing before you rehearse with a chatbot.

Then there is the part nobody warns candidates about: the checks. For any seat at a bank, a credit union or a broker-dealer, expect fingerprinting and a background check, often a credit check, and for a registered role a Form U4 that asks you to disclose bankruptcies, judgments, liens and certain charges. Banks are also barred by statute from employing people with specific dishonesty-related convictions without a waiver. None of this is automatically disqualifying, and almost all of it is survivable if you raise it yourself before it surfaces. Reference calls in finance are also unusually specific: the question asked is "did you own that forecast, and was it any good", so line up the manager who can answer that rather than the most senior person who knows your name.

Your own questions should sound like someone who intends to own a forecast. What does the close calendar look like and where does it hurt. Who are the business partners I would support, and which of them currently distrusts the numbers. What is forecast accuracy on this business today, and measured against what. What is the one report everybody complains about. What would you want me to have changed in six months. Those questions also tell you whether the job is decision support or spreadsheet maintenance, which determines whether the role is still interesting in two years.

Working with AI in this role

What a financial analyst has to know about AI in 2026-27

The honest version first, because the hype on this role has outrun the facts. The core of the job, owning a number you will personally defend in front of someone who can act on it, has not been automated and is not close to it. Excel is still the deliverable. The model is still the artifact. What changed is everything around the model: gathering data collapsed from hours to minutes, first drafts of commentary and memos are now effectively free, and the burden of proof shifted from producing the analysis to verifying it.

Three concrete shifts you can verify yourself by reading postings and talking to people doing the job. First, data extraction and spreading got absorbed. Tools that pull historical financials out of filings into a model, and that answer questions across filings, transcripts and research, are in normal use: Microsoft 365 Copilot and Python in Excel in corporate finance teams, AlphaSense and the AI features inside S&P Capital IQ and FactSet in research and corporate development, and firmwide internal assistants at large banks that were announced as deployments rather than pilots, including JPMorgan's LLM Suite, Morgan Stanley's OpenAI-built assistant for advisors and Goldman Sachs's GS AI Assistant. The junior work that consisted of retyping numbers out of a PDF into a sheet is shrinking, and it was a meaningful share of a first-year's week.

Second, hiring got more suspicious rather than less rigorous. Because an unsupervised take-home now proves very little, more employers have moved to supervised, in-person or screen-shared exercises with internet and assistants blocked, and to follow-up questions designed to establish that you can rebuild what you submitted. If anything the modeling test matters more than it did three years ago. Expect to be asked how you use these tools and what you do not let them do. A candidate with no answer reads as incurious, and a candidate who says they paste company financials into a consumer chatbot has just failed on controls.

On headcount, be careful what you believe. Senior people at several banks have said publicly that they expect to need fewer junior analysts. What is actually visible in postings is slower entry-level hiring and more employers asking for one to two years of experience, not the work vanishing. Treat the headcount claim as contested rather than settled, and notice that the variants struggling to fill seats, regional-bank credit, manufacturing and healthcare FP&A, cost accounting, public-sector budget, are the ones nobody writes threads about.

Third, where the value moved. A model will write a formula, a Python block, a VBA macro, a variance narrative and a first-pass memo quickly and mostly plausibly. It will not tell you which driver set is right for a business you have not interrogated, which of two sources is authoritative when they disagree, whether an adjustment is defensible to an auditor, what a cost-center owner is not telling you, or how much reversal cost a decision carries. It is also unreliable at long-chain arithmetic consistency, which is exactly the property a financial model depends on. And accountability did not move: if the number is in your deck, it is your number.

Using AI in Excel and in your data stack to delete work, with one real before-and-after

Employers are not hiring a financial analyst for AI expertise. They are hiring someone who will not be a drag on tools they already pay for, and who uses them to shorten a cycle rather than to add another dashboard. Close and reporting overhead is the cost a finance leader feels most directly, so a candidate who has measurably reduced it arrives with a contribution already quantified.

Show it: Name the stack in the posting's own words, such as Excel with Copilot, Python in Excel, Power BI, SQL against the warehouse, and the planning tool by name. Then give one before-and-after with your own measured number: "moved the first draft of monthly variance commentary to a generated pack reviewed against the ledger; commentary prep went from two days to four hours, and the exceptions actually got read." Also say what you deliberately kept manual and why, because indiscriminate automation is its own red flag in finance.

Verifying a generated number before it leaves your name, as a stated habit

The characteristic failure now is a fluent, confident, slightly wrong artifact: a summary that omits the one-off, an extracted figure that predates a restatement, a narrative explaining a variance that is actually a timing difference. An analyst who has been burned once behaves visibly differently from one who has not, and interviewers probe for it directly.

Show it: Describe the check as a procedure. Every extracted figure ties back to the filing or the ledger before it enters the model. Every variance explanation is confirmed with the cost-center owner before it is published. Anything in a CFO-facing pack traces to a source. One concrete catch beats any policy statement: "the extraction pulled a pre-restatement revenue line, and we caught it because the tie-out to the 10-K is a step in the process, not a courtesy."

Knowing your employer's data and MNPI rules well enough to say them out loud

In finance this is not a formality. Non-public financials, deal information, customer data and material non-public information carry legal and contractual consequences, and in regulated firms the use of unapproved tools is a supervisory issue. A candidate who distinguishes an enterprise tool inside the firm's boundary from a consumer chatbot is demonstrating the judgment compliance cares about most.

Show it: State the distinction unprompted when AI comes up: which tools you use on internal data and on whose approval, what you never put into an external tool, and how you handle a deal or an earnings blackout. If you have worked under an information barrier or a restricted list, say so plainly. That experience transfers and is scarce.

Writing the memo, because judgment is now the scarce half of the work

When drafting is cheap, the differentiator is a short, specific, defensible recommendation: what you believe, why, and what would change your mind. This is also the part of an analyst's output a senior person reads first and remembers. Building is the commodity skill; taking a position and owning the consequence is the scarce one.

Show it: Attach a real one-pager to your application: a public-company model with a recommendation, the two assumptions that matter most, and what would have to be true for you to be wrong. In the interview, volunteer which of your own numbers you trust least. Interviewers consistently read stated uncertainty as seniority.

Enough SQL and Python to go to the source instead of waiting for an extract

The constraint on an analyst is increasingly access to data rather than ability to analyze it. An analyst who can pull their own transaction-level data answers a question in an afternoon instead of filing a ticket, and generated code has lowered that bar a long way. Corporate FP&A postings now name SQL in roles that would not have mentioned it a few years ago.

Show it: Give one specific example of a question you answered with your own query that you would otherwise have queued for a data team, and say how long it took. Name the actual environment, such as Snowflake, BigQuery, SQL Server or Databricks, and say that generated code was reviewed before it ran if that is the team norm. Do not claim Python if you have only run notebooks someone else wrote.

Being able to rebuild, unaided, anything you submit

This is now an explicit hiring control. Supervised exercises and follow-up probes exist because employers assume take-homes were assisted. The candidate who can reproduce a debt schedule on a whiteboard passes a test the submitted file no longer passes on its own.

Show it: Practice the second build. After any assisted model, rebuild the mechanics from a blank sheet with no help, timed. In interviews, offer it: "happy to rebuild the debt schedule here if that is useful." Very few candidates say that, and it settles the question immediately.

What a screen is looking for

These are the terms that a resume screen, human or automated, is matching against for this role. Use the ones that are true of you, in the words the posting uses.

Mistakes that cost people this job

Sending one resume to FP&A, investment banking, credit analyst and budget analyst postings because they all say "financial analyst".

Pick the variant from the posting's nouns and rewrite the top third. FP&A leads with forecast scope, systems, close cadence and business partners. Credit leads with facility sizes, industries, memos and approval authority. Banking leads with deals, size and your role on them. Public sector leads with the posted criteria in the posting's own words. Same career, four documents.

Treating a target school as the gate on the whole profession and giving up before applying.

Name the band where school actually filters: bulge-bracket banking, campus-recruited asset management, and the private equity pipeline fed off them. Then apply to the rest of the market, which is most of it. Credit training programs, mid-market FP&A, valuation firms, rotational programs and the public sector hire on demonstrated ability.

Preparing for the modeling test by practicing speed and formatting, then producing a pretty model that does not balance.

Build in this order every time: structure, links, check row, then formatting if time remains. A plain model that balances, ties cash, drives interest off a debt schedule and separates inputs beats a formatted one with a plug. Put the balance check in the first screen the reviewer sees.

Hardcoding a plug to make the balance sheet balance and hoping nobody opens the cell.

They open the cell. It is the first thing an experienced reviewer checks. If you cannot find the break in the time allowed, say so in your cover note and name where you think it is. A candidate who flags an unresolved tie-out gets hired over one who hides it.

Handing back a spreadsheet when the exercise asked for a recommendation.

Write the two or three sentences: what you would do, the assumption it hinges on, and what you would ask management. The model is the evidence; the recommendation is the job. This is also the clearest way a non-target candidate outperforms a pedigreed one.

Writing "improved forecast accuracy" or "built financial models" and nothing checkable.

State the metric, the horizon and the scale, in the shape of "reduced revenue forecast error from 8% to 3% on a one-quarter horizon" and "owned the OpEx forecast for a $180M business unit across four cost centers in Adaptive". Both can be questioned, which is the point.

Listing a tool salad: twenty systems, with no indication of which you used daily and which you saw in a demo.

List what you have worked in, in the posting's own words, and be ready to describe one thing you built in each. A short honest list survives questioning; a long one collapses on the first follow-up.

Waiting for a Financial Analyst posting while sitting inside a company in accounting, operations or billing.

Make the internal move. Ask to own the forecast for one cost center or product line, build the variance pack nobody wants, present it to a business partner once, then tell the FP&A manager you want the next seat. Internal knowledge plus a manager who vouches for you beats external applications.

Deferring the job search to finish the CFA, then arriving with three exams and no analyst experience.

Start the search now and the exams alongside. The charter needs 4,000 hours over at least 36 months of relevant experience anyway, so a job is a prerequisite rather than a reward. Most employers will also pay for part of it.

Being unable to explain a number on your own resume when asked.

Before any interview, go line by line and answer out loud: where did that figure come from, over what period, what was the denominator, and what did I personally do versus my team. If you cannot source it, remove it.

Rehearsing with a chatbot until every answer is fluent, general and indistinguishable from everyone else's.

Prepare five specific stories with real numbers, real systems, the roles of the real business partners involved, and one real failure. Specificity is what interviewers now use to tell candidates apart, precisely because fluency has stopped being scarce.

Submitting an AI-assisted take-home you cannot rebuild, then being asked to reproduce the debt schedule live.

Rebuild anything you submit, unaided and timed, before you send it. Then offer the rebuild in the interview. Supervised follow-ups exist specifically to test this, and offering first removes the suspicion.

Questions people ask

What does a financial analyst actually do?

A financial analyst turns financial and operational data into a number someone else will act on, then defends that number. In corporate FP&A the week is built around a cycle: close the month, explain the variance against plan, reforecast the rest of the year, and answer whatever the business asks in between, which is usually a pricing, hiring or spending decision. In an investment role it is building and maintaining models of companies, forming a view on value, and writing it up. In commercial credit it is spreading a borrower's financials, sizing what they can safely carry, and recommending approval or decline in writing. The common core is the same in all three: own a forecast or a valuation, be the person who knows why it is what it is, and be useful to someone who has to choose.

Can I become a financial analyst without a finance degree?

Yes, and it is common. Hiring managers in corporate FP&A, commercial credit and the public sector care about demonstrated modeling ability, accounting literacy and judgment, not the name on the diploma. The two gaps worth closing deliberately are accounting fundamentals and Excel depth. The cheapest credible fix for accounting is two courses with a transcript: intermediate financial accounting, and managerial or cost accounting. The fix for Excel is building three-statement models from real 10-K filings until you can do it from a blank sheet under time pressure. Where a non-finance background genuinely constrains you is campus-recruited investment banking, which screens on degree, school and GPA before anyone looks at what you can do.

Do I need a CFA to be a financial analyst?

No. No credential is required for a corporate financial analyst role. Whether the CFA is worth it depends on the variant, and the answer is checkable in twenty minutes: read thirty postings in the exact job and geography you want and count how many name it. In sell-side research, asset management, portfolio analysis, private markets and much of credit it appears constantly and is often a hard filter. In corporate FP&A it appears occasionally as a preference and will not get you hired on its own; there the credentials employers name are the CPA, the CMA and AFP's FPAC. Note also that the charter requires 4,000 hours of relevant work experience over a minimum of 36 months, so it follows a job rather than replacing one.

What does the financial analyst modeling test look like?

Four formats dominate. A timed in-office or screen-shared Excel build, commonly 60 to 180 minutes with internet and AI tools blocked, where you forecast from supplied actuals and assumptions and write a short recommendation. A 24 to 72 hour take-home model with a one-page memo, increasingly followed by live questions to confirm you built it. A paper LBO or mental-math screen for banking and private credit, done with no computer. Or an existing model handed to you to walk through, critique, or find the planted error in. Graders check whether the balance sheet balances, whether ending cash ties, whether interest is driven off a debt schedule with circularity handled deliberately, whether inputs are separated from formulas, whether each formula is consistent along its row, and whether you answered the business question.

How do I prepare for a financial modeling test for free?

Pull three 10-K filings from SEC EDGAR for companies in different industries and build a full three-statement model for each from a blank sheet, timed. Then rebuild one of them a week later without opening your old file, because the second build is where the skill forms. Add a one-page memo to each: your forecast, the two assumptions that matter most, what would have to be true for you to be wrong, and what you would ask management. Separately, drill paper LBOs until you can approximate an IRR with no calculator, using the compounding anchors that roughly 2.0x over five years is about 15%, 2.5x about 20%, 3.0x about 25%, and 2.0x over three years about 26%. Paid modeling courses save time on conventions, but no certificate from one is why anyone gets hired.

Which licenses does a financial analyst need?

None for corporate FP&A, treasury, budget analysis or most buy-side analytical work. Securities licenses gate specific markets-facing seats: the SIE as a prerequisite, then FINRA Series 79 for investment banking representatives, Series 7 for general securities representatives, Series 86 and 87 for sell-side research analysts, and the Series 63 state exam. All of those except the SIE require association with a FINRA member firm, so they come after you are hired, usually within a window your firm sets. The SIE and the Series 65 can be taken with no sponsor, and an SIE pass stays valid for four years. Exemptions exist, including a research-exam exemption tied to CFA progress, so read FINRA's current rule rather than a forum summary.

How much does a financial analyst earn?

Look it up for your variant and metro rather than trusting a single band, because a government budget analyst, a mid-market FP&A analyst, a regional-bank credit analyst and a first-year investment banking analyst price in different labor markets. Use the US Bureau of Labor Statistics Occupational Employment and Wage Statistics for SOC 13-2051, Financial and Investment Analysts, which breaks down by state, metro area and industry, and use 13-2031 for Budget Analysts or 13-2054 for Financial Risk Specialists if that is your actual title. Then read twenty posted ranges in pay-transparency jurisdictions for your variant, plus AFP's compensation survey for FP&A. In banking and on the buy side base is the smaller half: ask for the target bonus as a percentage of base and the last two years' actual payout against target.

Is AI replacing financial analysts?

Not at the core of the job, and claims that it is are overstated. What has been absorbed is data extraction, spreading financials out of filings, first-pass commentary, deck formatting and routine reconciliation, which was a real share of a junior analyst's week. What has not been absorbed is deciding which drivers a business should be forecast on, judging which source is authoritative when two disagree, knowing whether an adjustment is defensible to an auditor, and standing behind a number in front of someone who will act on it. The practical effects on hiring are three: postings name tooling more often, employers ask what you let these tools do and what you do not, and modeling tests have moved toward supervised formats because unsupervised take-homes no longer prove much. Entry-level hiring has slowed and experience asks have risen, but the work itself has not disappeared.

How do I move from accounting into FP&A?

It is the most common and most reliable transition in finance, and it is usually made internally. Accounting gives you what self-taught modelers lack, a defensible balance sheet and close mechanics, so the gap you are closing is forward-looking work and business partnering rather than technical ability. Do it in a specific order: ask to own the forecast for one cost center or product line, build the variance pack nobody wants to build, take it to a business partner and talk them through it once, then tell the FP&A manager you want the next opening. If your employer has no FP&A function, use interim or contract FP&A work through a finance staffing firm to get forecast ownership onto paper.

What is the best entry-level route into financial analysis right now?

The routes with the least competition relative to their quality are commercial bank credit training programs, which teach a durable skill and take candidates with accounting coursework and no banking background; valuation and transaction advisory firms, which hire in volume and teach modeling under review; rotational finance development programs at large industrials, insurers, utilities, defense contractors and healthcare systems; contract or interim FP&A through finance staffing firms, which converts often; and the public sector, which is slow but published, scored and genuinely open. All five are easier to enter than the investment banking pipeline most online advice is written about, and all five lead somewhere.

Put this on a resume in about a minute

Paste your history once and point it at the Financial Analyst posting you are looking at. No account, no card.

Build my resume free More roles

Related roles