Ask ten hotel department heads to explain a variance and you will get ten answers of wildly different quality. Some will write three careful sentences that tell you exactly what happened and what it means for next month. Others will write "timing" and consider the matter closed.

The difference is rarely effort or ability. It is almost always the question. "Please explain your variance" is an invitation to describe the number, and describing the number is what most people do: revenue was down, payroll was up, expenses were higher than budget. All true, all visible on the report already, and none of it useful.

This article sets out what a variance explanation actually needs to contain, how the requirements differ between revenue, payroll and other expenses, and what separates a weak comment from a strong one. The examples use invented figures for a fictional property, the Harbor View Hotel.

What a useful explanation contains

A variance explanation is doing a specific job. It is telling a reader who was not in the department what caused a difference between two numbers, and what that means for the numbers that have not happened yet. To do that job it needs five things.

  1. What happened, stated as an event rather than a number.
  2. Why it happened, the operational cause behind the event.
  3. Whether it continues, meaning timing or permanent.
  4. What it does to the forecast, in figures where possible.
  5. What is being done about it, where action is warranted.

Notice that only the fourth item involves numbers. Finance already has the numbers. What finance cannot see from a trial balance is the event, the cause, the persistence and the response, and that is exactly what a department head knows and nobody else does.

A comment that answers all five in three sentences is worth more than a paragraph that answers none of them at length.

Revenue variances

Revenue variances are the most likely to be misread, because a single revenue number hides at least two moving parts. Rooms revenue is volume multiplied by rate. Food and beverage revenue is covers multiplied by average check, or events multiplied by event value. A total that lands on budget can conceal a large volume shortfall offset by rate, and the two have entirely different consequences.

So a rooms revenue explanation should separate the drivers before it explains anything:

  • Was occupancy above or below plan, and by how many room nights?
  • Was ADR above or below plan, and in which segment?
  • Did the segment mix shift, and did that shift help or hurt rate?
  • Was there a specific event, group, cancellation or closure involved?

A shift in segment mix deserves particular attention, because it often explains a variance that otherwise looks contradictory. A month can finish ahead on occupancy and behind on revenue if the extra room nights came in at a discounted rate, and that situation calls for a different conversation than a straightforward shortfall.

For food and beverage, the equivalent split is covers against average check, and outlet against banquet. Banquet is worth isolating because it moves in discrete blocks: one rescheduled event can create a variance that no operational change would have prevented.

Payroll variances

Payroll is where an explanation most often stops one step short. "Overtime was higher" is an observation, not a cause. Overtime is a symptom, and the question is what produced it.

The usual candidates are worth asking about directly:

  • Volume. Did occupancy or covers run ahead of plan, making the hours legitimate?
  • Vacancies. Were open positions covered by existing staff on overtime, or by agency labor at a premium?
  • Absence. Did sickness or leave force cover at short notice?
  • Productivity. Did hours per occupied room or per cover move, and if so, why?
  • Rate. Did a wage increase, a shift differential or a new benefit land in this period?
  • Calendar. Did the period contain an extra pay date, or an accrual adjustment?

These have very different implications. Overtime driven by volume may be perfectly efficient and should be explained alongside the revenue it supported. Overtime driven by a vacancy is a recruitment problem with a known end date. Overtime driven by productivity drift is the one that needs management attention, and it is also the one most likely to be described as "higher than expected" and left there.

Wherever possible, a payroll comment should express the variance in relation to volume, not in isolation. Hours per occupied room, or labor as a percentage of departmental revenue, turns an unhelpful absolute number into a comparison that means something.

Other expense variances

Expense variances split cleanly into three types, and naming the type does most of the explaining.

Volume-driven

Guest supplies, laundry, food cost and credit card commissions move with business levels. If the driver moved, the expense should have moved, and the interesting question is whether it moved proportionally. An expense that rose faster than the volume behind it is the variance worth writing about.

Timing

An invoice posted in the wrong period, a contract paid annually, a repair that slipped a month. These reverse. They should be identified as timing, and the period the offset lands in should be stated, because otherwise the same variance appears twice with opposite signs and nobody remembers why.

Structural

A new contract, a rate increase from a supplier, a change in service standard, an added headcount. These do not reverse. They change the run rate, and the forecast has to move with them. Structural variances are the most important to identify early and the easiest to mistake for timing in the first month they appear.

Getting to the root cause

A reliable way to test whether an explanation has reached the cause is to ask "why" again and see whether there is an answer.

Working down to the cause

Laundry expense finished $4,200 over budget. Why?

Linen volume was higher than planned. Why?

A larger share of stays were one-night, and every stay generates a full linen change. Why?

Two group bookings that would have been three-night stays canceled, and the rooms were filled with one-night transient business.

The last answer is the one worth writing down. It explains the expense, links it to a revenue event, tells the reader whether it repeats, and points at what to watch. The first answer, "linen volume was higher", would have been technically accurate and entirely useless.

Two or three "whys" is usually enough. The aim is a cause a manager can act on or plan around, not an infinite regress.

Timing, permanent and forecast effects

This is the distinction that makes a variance report useful for anything other than looking backwards, and it is the one most frequently skipped.

A timing variance means the money moves between periods. Nothing has changed about the year, only about which month carries the figure. The correct treatment is to say which period the offset falls in and by how much.

A permanent variance means the money is gone or gained for good. Lost business that will not be recovered, a supplier increase that continues, a cost saving that repeats. The correct treatment is to adjust the forecast for every remaining period it affects.

The word "timing" on its own is not an explanation. It is a category label, and it is only useful when it comes with the offsetting period and amount attached. A comment reading "timing, expect the offset in July" is dramatically more useful than "timing", and it takes four more words.

Corrective action

Not every variance needs corrective action. A favourable revenue variance driven by strong demand needs no correction, and inventing one wastes everyone's time. Action belongs in the comment when the cause is controllable and repeating.

Where it applies, an action statement needs three parts: what will be done, by whom, and by when. Without those, it is an intention rather than a plan.

Weak and strong examples

These use invented figures for the fictional Harbor View Hotel. The weak versions are not exaggerated. They are the kind of comment that arrives in most inboxes.

Weak · Banquet revenue

Banquet revenue was below budget due to lower event volume.

Strong · Banquet revenue

Banquet revenue finished $6,690 below budget on two fewer corporate dinners. Both were rescheduled by the client after a change to their own travel plans, and both are confirmed for July. This is timing, not lost business: July banquet forecast is increased by $11,400 and the full-year outlook is unchanged. To reduce short-notice moves, deposit terms for corporate bookings move to 60 days before the event from August, agreed with the director of sales.

Weak · Rooms payroll

Payroll was over budget because of overtime in housekeeping.

Strong · Rooms payroll

Rooms payroll was $6,420 over budget, all of it housekeeping overtime. Two room attendant positions have been vacant since May, and the hours were covered by existing staff rather than agency labor, which was the cheaper of the two options. Hours per occupied room were 0.48 against a standard of 0.45. Both positions start on 14 July, so the overrun continues at roughly $3,000 in July and ends from August.

Weak · Food cost

Food cost percentage was unfavourable. We are monitoring the situation.

Strong · Food cost

Food cost ran at 31.4% against a budget of 29.0%, worth about $2,900 on the month. Two causes: our main protein supplier applied a 6% increase from 1 June, which is permanent and adds roughly $1,700 a month at current volumes, and a walk-in cooler fault spoiled stock in week two, which was a one-off of about $1,200. The supplier increase is built into the forecast from July. Three competing quotes have been requested, due back by 31 July.

The strong versions are not longer because they are padded. Every extra sentence carries a fact that the reader could not have got from the report.

How to ask so you get good answers

Comment quality is largely determined before anyone writes anything, by how finance asks and what finance supplies. Four things make the difference.

Ask specific questions, not for an explanation. Replace the open request with the five questions above, each in its own field. People answer questions readily and find open-ended requests difficult, and a form makes an incomplete answer visible.

Set a threshold and stick to it. Commenting on every line teaches people that none of it matters. Pick a threshold in both currency and percentage terms, apply it consistently, and only request comment where it is crossed.

Send the figures with the request. A department head should not have to find the variance before explaining it. Send the line, the budget, the actual and the difference already calculated.

Show what good looks like. One strong example attached to the request will do more than a page of guidance. If a comment comes back thin, send it back with a specific question rather than rewriting it yourself. Rewriting it yourself is faster this month and guarantees the same problem next month.

Key takeaways

  • A useful explanation answers five things: what happened, why, whether it continues, the effect on the forecast, and what is being done.
  • Split revenue into volume and rate before explaining it, and split payroll into volume, vacancy, absence, productivity, rate and calendar.
  • Classify expense variances as volume-driven, timing or structural. Naming the type does most of the explaining.
  • "Timing" is only an explanation when the offsetting period and amount come with it.
  • Ask specific questions, set a threshold, send the figures, and show one strong example. Comment quality is set by the request.

This article describes general practice. It is not accounting, tax, audit or legal advice, and it does not replace your own professional judgment or your company's procedures. Every example uses invented figures for a fictional property. Read the disclaimer.