Recipe 44 - The Forecast Variance Commentary Writer: Write Month-End Variance Commentary That Survives Scrutiny

Write Variance Commentary That Survives Scrutiny

From The Spiral Sales Office Recipe Vault | By Paul Woodley

There is a standard format that distinguishes variance commentary that gets believed from variance commentary that generates follow-up questions. It is not a long format. It is a five-part structure: number, cause, action, owner, checkpoint. A variance written in that structure is defensible. A variance written as a general narrative of what happened is not, because it omits the three things an owner or management company actually wants to know: what is being done, who is responsible, and when will we know whether it worked.

Recipe 44 is the final recipe in the vault. It builds the commentary from the evidence you provide rather than from the reconstruction you do on the morning it is due. That distinction is where credibility is built or lost at the most consequential moment in the reporting cycle.

What This Recipe Makes

Variance commentary for every line that breached your commentary threshold, structured as number, cause, action, owner, and checkpoint, written from evidence rather than from memory, in language suitable for an ownership or management company review.

• Head Chef: DOSM with the General Manager and Director of Finance
• When to Cook It: Monthly at close, on every line that breaches the commentary threshold
• Ingredients: Forecast versus actual by line and revenue stream for the closed month; the accuracy output from Recipe 42; what you know about why each variance happened

The Five-Part Structure That Survives Scrutiny

Every line of variance commentary produced by Recipe 44 follows the same structure. When every commentary field in a reporting package uses this structure, the package becomes reviewable in a consistent, efficient way. The reviewer can move through it without having to interpret different formats for different lines.

1. Number - The variance stated plainly: actual versus forecast, and the size of the gap in the relevant unit. Not a softened description. A number
2. Cause - One sentence on what drove the variance. Not a list of contributing factors. The primary cause, stated specifically. If two causes contributed equally, both are named. If the cause is not known, that is stated plainly, which is itself information and leads directly to the action
3. Action - What is being done in response to the variance. Specific enough to be verifiable. Not we are monitoring the situation or working to improve performance. A specific action with enough detail that anyone reading it could confirm whether it happened
4. Owner - The person responsible for the action. Named by role, at minimum, and by name where the hotel's reporting conventions permit. Ownership without an owner is shared responsibility, which in practice is no responsibility
5. Checkpoint - When will the next data point be available to assess whether the action is working? A specific date or a specific event, such as the close of the following period or the completion of a specific conversation, not a general statement about ongoing monitoring

How Recipe 42 Feeds Recipe 44

Recipe 42, the Forecast Accuracy Post-Mortem, produces the directional-bias read: whether the missed lines were consistently high or consistently low, and whether the misses cluster by stream. That output is the context that makes Recipe 44 commentary more specific.

A variance in beverage revenue that appears in the context of a consistent history of overforecasting beverage is a different commentary from one that appears in isolation. The first points toward an assumption that is structurally wrong and needs a systematic fix. The second may be a single-event explanation. Recipe 42 tells you which situation you are in; Recipe 44 writes the commentary appropriate to it.

The Threshold Question

The prompt applies your commentary threshold, the dollar or percentage variance above which commentary is required. If no threshold has been defined, it flags that as a gap and uses an illustrative benchmark labelled as an assumption. Setting a threshold matters because threshold-based commentary is reviewable: every line above the threshold has a five-part explanation, every line below it does not, and the reviewer knows what to expect. A reporting package without a defined threshold produces inconsistent commentary, which reads as a commentary-when-it-seems-necessary approach that undermines the credibility of the package.

Where the Output Goes

The month-end reporting package, where the commentary is pasted into each variance field. The one-on-one with the General Manager before the package goes to ownership, where any commentary that requires GM awareness before it is submitted is flagged. The accuracy log from Recipe 42, where each variance and its explanation become the input to the next cycle's forecasting assumptions. The monthly critique, where the variance commentary is the basis for the discussion of what changed and why rather than a starting point for reconstructing the story.

The Mirror

For the last month-end variance that required commentary: was the cause specific or general, was there an identified action with an owner, and was a checkpoint named? If any of those three were absent, the commentary was incomplete by the standard that makes a reporting package defensible. Recipe 44 builds the complete version.

The Spiral Closes Here

Recipe 44 is the last step in the measurement cycle. It comes after the forecast was built in Recipe 39, validated in Recipe 40, reconciled in Recipe 41, and its accuracy assessed in Recipe 42. The variance commentary closes the loop: it explains what happened, names what is changing, and sets the date by which the explanation will be tested against the next set of numbers.

That loop is the spiral moving upward. The forecast that is built with method, submitted with reconciliation, assessed with honesty, and explained with evidence becomes a better forecast the next time. The DOSM who runs that cycle every month, month after month, builds the kind of credibility that makes every other conversation easier: with the GM, with ownership, with the team, and with the numbers themselves.

A Note on Data Handling

Confirm what your employer, owner, brand, and management company permit before pasting forecast and actual revenue data into any AI platform. Month-end revenue figures, ownership commentary, and variance analysis may be commercially sensitive. When in doubt, describe the variance in relative terms rather than absolute revenue figures. Every output is a starting position. The commentary and the decisions it reflects still belong to you.

Your Next Step

After next month closes, before you write a single line of variance commentary, pull your forecast versus actual by line and stream, and the accuracy output from Recipe 42. Run Recipe 44. Paste the five-part commentary into every field that breached your threshold. Then submit a reporting package where every variance has a number, a cause, an action, an owner, and a checkpoint.

Variance commentary written from evidence, the day after close, in a consistent format, is not just a reporting requirement. It is the discipline that improves the next forecast, and the one after that. That is the spiral moving in the right direction.

Paul Woodley | The Spiral Sales Office