Building and Maintaining an Earnings Model

Initiating coverage (previous chapter) builds the model once. This chapter covers the ongoing work: updating that model every quarter, and what genuinely happens around an earnings report.

Consensus estimates — the number a company is actually measured against

Consensus is the average (or median) of all covering analysts' individual estimates for a specific metric — usually EPS and revenue — for an upcoming quarter. A company's actual reported results get measured against this consensus number, not against some absolute standard of "good" or "bad" — which is exactly why a company can report genuinely strong year-over-year growth and still see its stock fall, if that growth came in below what the market had already priced in via consensus.

Beat, miss, and the real stock reaction

Worked example: consensus EPS estimate is $2.00. The company reports actual EPS of $2.15 — a beat of $0.15, or 7.5% above consensus. All else equal, this typically supports a positive stock reaction. But if the company also guides next quarter's expected EPS to $1.90 (below the new forward consensus that had been building toward $2.05), the stock can still fall on the news — the market is forward-looking, and a strong quarter paired with weak forward guidance is a genuinely common pattern that surprises people expecting the quarter's own results alone to drive the reaction.

Updating the model after each earnings report

After each quarter, the analyst updates the financial model with real, reported figures (replacing what were previously just estimates for that quarter), rolls the forecast forward, and — based on the update — issues a revision: raising, lowering, or maintaining prior estimates for future quarters, and correspondingly, the DCF/comps-derived target price itself may move even without a formal ratings change. A meaningful, sustained pattern of upward estimate revisions across the sector (not just one company) is itself a real, closely-watched signal — often used as a leading indicator that an industry is doing better than the market had broadly priced in, sometimes before it's even fully reflected in stock prices.

Guidance — management's own forward estimate, and why it's not just taken at face value

Guidance is management's own stated expectation for future results — genuinely useful information, since management has real, direct visibility into the business that outside analysts don't. But experienced analysts also weigh a company's own historical track record of hitting, missing, or reliably sandbagging (deliberately setting conservative guidance to consistently "beat" it later) its own guidance — a company with a pattern of reliably beating its own conservative guidance is being read differently by the market than an identical guidance number from a company with a track record of missing its own targets.

Why this ongoing maintenance work is the bulk of a research analyst's actual job

Initiating coverage happens once per company; earnings updates happen every single quarter, for every company under coverage, indefinitely — the genuine bulk of a sell-side analyst's real day-to-day time, and the actual mechanism by which a research view stays current rather than becoming a stale, one-time snapshot.