Why Your CFO Rewrites the Variance Commentary Exactly Three Times Before Earnings

This is satire, although anyone who has spent the week before earnings trying to explain a 180-basis-point gross-margin movement without using the phrase “one-time” seven times may find it improperly documentary. The CFO does not rewrite variance commentary three times because finance cannot write. Finance can write. Finance has written 14 pages of source-linked analysis, with bridge tables, customer cohorts, FX normalization, purchase-price variance support, and a footnote explaining why the footnote is not material. The CFO rewrites it three times because earnings commentary must pass through three separate systems of organizational physics before it can be released into the wild.

Revision One: The Version That Is True

The first draft is usually prepared by the person closest to the numbers, often an FP&A manager, senior director, controller, investor-relations lead, or the analyst who has become the unofficial keeper of the file called FINAL_v23_USE_THIS_ONE.xlsx. It says what happened. Revenue was below plan because implementation slipped at two enterprise customers. Gross margin declined because the mix shifted toward lower-margin services, expedited freight rose, and a vendor rebate posted one month later than forecast. Operating expense rose because the company hired ahead of a product launch that is now expected to occur after the universe cools.

This draft is admired internally for approximately nine minutes. It is specific, supported, and comprehensible to anyone with access to the general ledger, the operating plan, and basic emotional resilience. It also contains the fatal flaw: it assigns causes. Causes imply ownership. Ownership implies someone may be asked a follow-up question by the CEO, the board, or an equity analyst who has printed the prior quarter’s transcript and arrived with a highlighter.

The CFO’s first rewrite therefore removes sentences that are technically correct but insufficiently survivable. “Sales hiring lagged plan” becomes “sales capacity continued to develop.” “The ERP cutover disrupted invoicing” becomes “billing processes experienced a planned transition period.” “We did not forecast this accurately” becomes a sentence so carefully engineered that it is legally a weather pattern.

Revision Two: The Version That Is Strategically True

The second draft is where commentary enters the Executive Translation Layer. This is not the same as concealment. Concealment has a compliance department, external counsel, and a calendar invite with no agenda. Executive translation is the more ordinary process of deciding which true thing belongs in the earnings narrative and which true thing should remain safely in the planning model, beneath a tab named Assumptions_Archive_Do_Not_Delete.

At this stage, each variance is asked to declare its long-term intentions. Is it temporary? Structural? A timing issue? A mix issue? An execution issue? An issue caused by a customer, a supplier, a foreign exchange rate, a former executive, or the ancient curse placed upon all deferred-revenue schedules? The answer matters because each category has different consequences for guidance, headcount, capital allocation, and the number of times someone will say “confidence” during the call.

A good CFO knows that the same $4 million unfavorable variance can mean very different things. In a mature manufacturer, it may trigger a discussion of volume, labor efficiency, inventory, and pricing realization. In SaaS, it may become a debate over bookings, ramp, churn, services mix, and whether the word “durable” can carry an entire slide. In a private-equity-backed business, it may be placed next to the EBITDA bridge and asked, politely but firmly, why it has not yet become an adjustment.

This is also the version where the CEO adds a sentence about momentum. The sentence may be valid. It may even be useful. But it will be reviewed by finance as though it is a volatile derivative held by a subsidiary in a jurisdiction no one remembers approving.

Revision Three: The Version That Can Be Said Out Loud

The final rewrite is for the earnings call itself: a live event in which every adjective acquires a market capitalization. Commentary must now be readable at a measured pace, understandable to investors who have joined late, and narrow enough that it does not accidentally create a new forecast, a new key performance indicator, or a new obligation to explain the same phrase for the next eight quarters.

This is why the final language can sound strangely polished. The commentary has been pressure-tested by finance, legal, investor relations, the CEO, the operating lead, and a person from communications whose job is to ask whether the phrase “margin headwind” sounds too meteorological. It has survived the question, “Could an analyst interpret this as worsening?” It has survived the follow-up, “Could they interpret it as improving?” It has survived the ultimate test: “Can we answer the obvious next question without introducing three new problems?”

By now, the original explanation may have evolved from six crisp bullets into: “Results reflected a combination of expected investment, timing dynamics, and select operational factors, while underlying demand remained consistent with our strategic priorities.” This sentence appears to have been produced by a committee of fog. In fact, it is the last remaining sentence that everyone can live with.

What Job Candidates Should Notice

If you are interviewing for FP&A, strategic finance, investor relations, controllership, or a finance-lead role, do not treat this process as proof that clear analysis is wasted. The opposite is true. The public sentence gets shorter only because someone built a much more detailed private answer underneath it. Hiring managers want people who can trace a variance from the headline to the transaction-level driver, distinguish timing from trend, quantify what is controllable, and explain what management should do next.

• For an FP&A role, be ready to describe how you built an executive-level bridge without losing the operational drivers underneath it.
• For a controller or accounting role, explain how you separated reported results, accounting effects, and genuine business performance without hand-waving any of them away.
• For finance operations or business operations, show how you turned a recurring explanation—billing delays, poor master data, approval bottlenecks—into a process fix rather than a permanent commentary category.
• For director and VP roles, expect questions about judgment: what you elevate, what you monitor, who owns the action, and when an unfavorable variance changes the forecast rather than merely decorating the monthly package.

The joke is that the CFO’s third draft may sound less informative than the first. The true observation is that the company still needs the first draft. Before anyone can communicate a result responsibly, someone in finance has to know exactly what happened, why it happened, what changes next, and which supporting schedule will answer the inevitable question.