Live Modeling

A full three-statement operating model links the income statement, balance sheet, and cash flow statement so that every projected period flows correctly and automatically from the prior one, built typically with a clean separation between hardcoded historical inputs (blue font, by banking convention), formula-driven calculations (black font), and links pulling from another sheet (green font), so anyone reviewing the model can tell at a glance what's an assumption versus what's derived.

The build order matters: revenue and operating expense assumptions drive the income statement down to net income; net income flows into retained earnings on the balance sheet; a supporting debt schedule (interest expense, revolver draws/paydowns) and a working capital schedule feed both the income statement (interest) and the balance sheet (receivables, payables, debt balances); and the cash flow statement is built last, reconciling the change in every balance sheet account back to the change in cash.

A three-statement model is inherently circular: interest expense depends on the debt balance, but the debt balance (via a revolver draw or paydown, or cash sweep) depends on how much cash the business generates, which itself depends on interest expense. Excel handles this either by enabling iterative calculation (allowing the circular reference to resolve numerically) or by inserting a circularity breaker/switch (a toggle that can zero out the circular interest calculation to unfreeze the model if it breaks), which is standard practice specifically to avoid a model getting stuck or showing errors that are hard to trace back to their source.

The 'cash flow plug', typically a revolving credit facility that draws or repays automatically to keep cash from ever going negative, is what makes the model balance dynamically each period without manual intervention, and understanding exactly how that plug mechanism works is one of the clearer signs of real modeling fluency.

Every complete three-statement model should include an explicit, visible balance check confirming assets equal liabilities plus equity in every single projected period, if it doesn't tie out, there's a linkage error somewhere (a common culprit: net income not flowing correctly into retained earnings, or a working capital change not correctly reflected on both the balance sheet and cash flow statement). Building this check in from the start, rather than adding it only once something looks wrong, catches errors immediately rather than after a much longer, harder-to-trace build has compounded on top of the mistake.

A well-built model also includes sanity checks beyond the pure balance identity, does projected revenue growth look reasonable, does the debt schedule correctly stop amortizing once debt hits zero rather than going negative, since a model can balance mechanically while still containing an error that produces an implausible, wrong output.

Unlike a paper interview version, a live comps build starts from pulling real, current data, typically from Capital IQ, FactSet, or Bloomberg, for every company in the set: current share price, diluted shares outstanding, balance sheet items for the EV bridge, and the relevant financial metrics (revenue, EBITDA, net income) for the trailing twelve months and forward estimates. Getting the comp set itself right (discussed in the interview-level valuation lessons) is only the starting point; the live version is about executing the actual data pull and calculation cleanly and correctly at scale, often across a dozen or more companies at once.

Sources need to be tracked carefully, which filing or data provider each figure came from, since a live comps page in a client deck needs to be defensible if a client or a senior banker questions any individual number.

Scrubbing means adjusting reported financials to remove one-time or non-recurring items (a legal settlement, a restructuring charge, a divestiture gain) so the resulting multiple reflects each company's ongoing, comparable operating performance rather than being distorted by a one-off event, a live comps build requires actually reading each company's earnings release and footnotes to identify these items, not assuming reported figures are already clean.

Every adjustment made needs a clear, visible footnote explaining what was excluded and why, both for internal review (an associate or VP checking the work needs to be able to verify the adjustment was reasonable) and because a scrubbed number that later gets questioned by a client needs a specific, citable justification, not just 'we removed it because it looked unusual.'

A live comps page isn't built once and left alone, it needs to be refreshed as new earnings are reported, share prices move, or estimates are revised, and on an active deal it's often updated daily or weekly leading up to a client meeting or a bid deadline. Building the page with clean, well-organized formulas from the start (rather than hardcoded, one-off calculations) makes this refresh process fast and low-risk; a poorly structured initial build turns every update into a slow, error-prone rebuild.

This is one of the clearest places where the error-checking discipline and formula-auditing habits covered separately matter in practice, a comps page that gets refreshed repeatedly over the life of a deal is exactly the kind of recurring, high-visibility deliverable where a small formula error, if uncaught, compounds and eventually surfaces in front of a client.

An interview-level DCF typically works from simplified, given assumptions; a live, on-the-job DCF usually starts from actual management projections (or, absent those, detailed analyst estimates and historical trend analysis), reconciled against a company's own operating model and recent guidance. Part of the live analyst's job is testing the reasonableness of management's own projections, is the assumed growth rate consistent with the company's historical performance and its industry's growth, are margin assumptions consistent with what similar companies have actually achieved, rather than accepting them uncritically.

Where management projections seem optimistic (common, since management has an incentive to present the most favorable case), a live DCF often builds a separate, more conservative case alongside management's own numbers, so the eventual output isn't solely dependent on assumptions the company itself provided.

A live WACC build pulls real inputs rather than given interview assumptions: current risk-free rates from government bond yields, beta estimated from a genuine unlevered-and-relevered comp set (not a single assumed number), current credit spreads or bond yields for the cost of debt, and current market values (not book values) for the equity and debt weights. Each input requires a defensible, cited source, since WACC is one of the most heavily scrutinized assumptions in any live DCF a client or senior banker will review.

Because WACC has such outsized leverage over the final valuation, live analysts typically build and present a WACC sensitivity range (rather than a single point estimate) directly in the output, explicitly acknowledging the real uncertainty in the discount rate assumption rather than presenting false precision.

A live DCF's output is rarely shown as a single number in client-facing materials, it's typically presented as a sensitized range (varying WACC and terminal growth or exit multiple, as covered in the sensitivity lesson) within a broader 'football field' chart alongside comps and precedents, so the DCF's contribution to the overall valuation view is visible and can be discussed alongside the other methods rather than in isolation.

Being able to explain, clearly and quickly, why the DCF landed where it did relative to comps and precedents, and which specific assumptions are driving any meaningful divergence, is exactly the skill a live analyst needs when a VP or client asks about the output in a real meeting, which is why understanding the DCF's mechanics deeply (not just being able to build it) matters so much on the job.