Financial Analyst

Financial accounting is about recording and reporting what already happened, following standards like GAAP or IFRS so the numbers are consistent and auditable. Financial analysis takes those numbers and interprets them — explaining performance, spotting trends, and supporting decisions about what to do next. Accounting produces the raw material; analysis is what you do with it.

The income statement shows revenue, expenses, and net income over a period. Net income flows into the balance sheet through retained earnings, and the balance sheet shows what the company owns and owes at a point in time. The cash flow statement bridges the two — it starts with net income and adjusts for non-cash items and changes in balance sheet accounts to show how cash actually moved, split into operating, investing, and financing activities.

Look beyond a single metric: profitability (margins, trend direction), liquidity (can it cover short-term obligations — current ratio, cash position), leverage (how much debt relative to equity or EBITDA, and can it service that debt), and cash generation (is net income backed by real cash flow, or is it propped up by receivables or one-time items). A company can look profitable on paper and still be in trouble if it's burning cash.

EBITDA strips out interest, taxes, depreciation, and amortization to approximate operating cash-generating ability, which makes it useful for comparing companies with different capital structures or tax situations. Net income is the actual bottom-line profit after everything, which is what actually accrues to shareholders. Use EBITDA for operational comparisons and valuation multiples; use net income when you care about the real, final profitability of the business.

Start by comparing actuals to budget/forecast at a granular enough level to isolate the driver — volume vs. price vs. mix, or department by department — rather than just flagging the total variance. When a number looks off, I check for a data or timing issue first (a misclassified entry, an accrual timing difference) before assuming it's a real business change, then go talk to whoever owns that line item rather than guessing.

Fixed costs (rent, salaried headcount, insurance) don't move with volume in the short run; variable costs (materials, sales commissions, shipping) scale with activity. It matters for forecasting because it tells you how profit responds to changes in volume — a business with high fixed costs has more operating leverage, so revenue swings hit profit harder in both directions, which changes how conservative your forecast assumptions should be.

First understand the revenue driver — units × price, customers × average revenue per customer, or something similar — then find historical trend data for each driver. Build a base case from the trend, then sanity-check it against known events (a price change, a new product launch, seasonality) and stress-test it with a downside and upside scenario rather than presenting one number as certain.

Working capital is current assets minus current liabilities — essentially the cash tied up in running day-to-day operations (receivables, inventory, minus payables). A fast-growing company often has to pay suppliers and build inventory or extend credit to customers well before it collects cash from sales, so profit on the income statement doesn't show up as cash in the bank yet — that gap is what causes 'profitable but cash-poor' situations.

OpEx (operating expenses) is recognized immediately on the income statement in the period incurred. CapEx (capital expenditure) is for assets expected to provide value over multiple years — it goes on the balance sheet as an asset and is expensed gradually over its useful life through depreciation, rather than hitting the income statement all at once.

XLOOKUP/INDEX-MATCH for flexible lookups that don't break when columns move, SUMIFS/COUNTIFS for conditional aggregation, and pivot tables for quickly slicing large datasets without writing formulas. For financial models specifically, I lean on named ranges and clear formula structure over cleverness — a model other people can audit is more valuable than one that's technically impressive but opaque.

I'd clarify what they actually mean — often it's a legitimate ask about which real levers (timing, presentation, which metric to lead with) are being used, not a request to misstate anything. If it crosses into misrepresenting actual results, I'd push back directly and explain why, since numbers that don't hold up under scrutiny cause bigger problems later than an uncomfortable conversation now.

A budget is a fixed financial plan set once, usually annually, that becomes the benchmark for performance and often ties to incentives. A forecast is a living estimate that gets updated regularly (monthly or quarterly) to reflect the latest information, so it's meant to be as accurate as possible right now rather than fixed as a target.

I'd break it into the actual drivers — did input prices rise, did volume grow, did the product mix shift toward lower-margin items — rather than stating the percentage alone. I'd frame it in terms that connect to what they can act on: 'volume grew 10% which explains most of this, but input costs also rose 5%, which is the part worth watching.'

SaaS: MRR/ARR, churn rate, customer acquisition cost, LTV/CAC ratio, and net revenue retention — since the business is about recurring value over time. Retail: same-store sales growth, inventory turnover, gross margin, and average transaction value — since the business is about repeated transactions and inventory efficiency. The right KPI set follows from how each business actually creates and captures value.

A strong answer names the specific error (a broken formula reference, a double-count, a wrong sign), how you caught it (a sanity check against a prior period, a total that didn't tie out), and what you did afterward — not just fixing it quietly, but flagging it to whoever relied on the earlier version and, ideally, adding a check that would catch the same class of error next time.

I check that totals tie out across tabs, sanity-check outputs against a rough independent estimate, stress-test extreme inputs to see if the model breaks or produces nonsense, and have someone else review the key assumptions and formulas — a model that only makes sense to the person who built it is a risk.

I'd assess actual urgency and impact rather than just recency — a recurring deadline usually has a hard external dependency (board meeting, close calendar) that ad hoc requests often don't. I'd communicate proactively with whoever made the ad hoc request about timing rather than silently deprioritizing it, since most people would rather know upfront than be surprised later.