Revenue Management 101.

What revenue management is

Revenue management is selling the right room to the right guest at the right price at the right time, so that a fixed number of rooms earns the most it can over the long term. The definition is Sheryl Kimes’s.

There are three levers. Price is what you charge. Duration is how long the guest stays, and when. Reach is how and where you sell. Most hotels pull the first and neglect the other two.

It is not discounting. Across industries Kimes puts the gain from doing it properly at 2 to 5 percent of revenue, most of which reaches profit.

Does it suit your hotel? Six conditions

• Capacity is relatively fixed
• Fixed costs are high and variable costs are low
• Unsold rooms perish every night
• Guests book in advance, so a request can be accepted or declined
• Demand varies by season, day and hour
• Different guests will pay different prices, and you can tell them apart

Score each from 1 to 5. Above 80 percent of the maximum, run the full discipline. Between 60 and 80, start with duration controls and pricing. Below 60, fix the product first.

The numbers

ADR is rooms revenue divided by rooms sold. Occupancy is rooms sold divided by rooms available. RevPAR is the two multiplied.

RevPAR rewards volume bought at any price. Net RevPAR deducts the cost of winning the booking. GOPPAR shows what is left after running the hotel. Report all three together.

Against competitors, read MPI, ARI and RGI as a set. Hotels priced above their competitive set earn a RevPAR premium on average (Enz, Canina and Lomanno).

Net RevPAR = (rooms revenue − acquisition cost) ÷ rooms available. RGI = MPI × ARI ÷ 100.

One price leaves money behind

Take 100 rooms and demand that falls in a straight line, from 100 rooms at a price of zero to none at a price of 100. One price of 50 sells 50 rooms and earns 2,500. Four prices of 80, 60, 40 and 20, each selling 20 rooms, earn 4,000 from the same demand. That is 60 percent more (Hanks, Cross and Noland).

The gain holds only if a guest who would pay 80 cannot buy at 20. The rules that keep them apart are called fences.

Fences

Physical

View, floor, room size, club access.

Product line

Good, better, best.

Controlled availability

Advance purchase, non-refundable, member rate, book-by date.

Buyer

Corporate, government, loyalty tier.

A good fence rests on something the guest controls: when they book and how firmly they commit. It is clear to the guest and to the reservations agent. The discount is deep enough to attract the guest you want, and the restriction tight enough to keep the others out.

Build fences at the top as well as the bottom. Paid extras that today’s guests value earn more than new discounts for guests who were not coming.

Fair pricing

Guests judge a price against a reference price and against the way they expect to buy. They accept that a firm deserves a reasonable profit. They resent a price raised simply because demand is high (Kahneman, Knetsch and Thaler).

• Show the full rate, so the discount is visible
• Attach a benefit to every higher price and a restriction to every lower one
• Frame a difference as a discount: 20 percent off on weekdays reads better than 20 percent more at weekends
• Quote each night at its own rate
• Never let the price depend on whether the guest asks

Forecasting

The forecast comes first. Without one there is no revenue management. With a poor one there is poor revenue management (Kimes).

Build a booking curve: rooms on the books by days before arrival, for like days. Then add the pickup you normally see between now and arrival. If 55 rooms are booked seven days out and Fridays normally pick up 40 more, forecast 95.

Forecast what demand would have been, including the guests you turned away on full nights. History taken from sold-out nights understates demand for your best dates for ever.

Measure the error. Report the weighted error and the bias together. A forecast can be close on average and still wrong in the same direction every week.

Forecast = rooms on the books + average pickup from this point. Weighted error (WAPE) = sum of absolute errors ÷ sum of actuals.

Hot, warm and cold

Divide forecast demand by capacity for each night. At 100 percent or more the night is hot. Between 80 and 99 it is warm. Below 80 it is cold (Kimes).

Hot means more demand than rooms. A hotel that calls 40 percent occupancy its peak has a product problem, and pricing will not solve it.

Protecting rooms for guests who book late

The guests who pay most tend to book last. If every room is sold early at a low rate, nothing is left for them. Littlewood’s rule: accept a low-rate booking while the low rate is at least the high rate multiplied by the chance of still selling that room at the high rate.

With rates of 250 and 200, and late demand averaging 15 rooms with a standard deviation of 5, the rule protects 10 rooms. The calculator below does it for your figures. With several rates the same idea is applied in layers (Belobaba’s EMSR-b).

Protect y rooms, where the chance that high-rate demand exceeds y equals low rate ÷ high rate.

Length-of-stay controls

Use them on hot dates only, and forecast by length of stay before setting them. Stacking many controls has damaged hotels (Kimes).

Overbooking

Some guests with reservations do not arrive. An empty room costs its contribution. A guest turned away costs the room elsewhere, the transport, the compensation and the goodwill. Overbook one more room while the chance of enough no-shows is at least the walk cost divided by the walk cost plus the empty-room cost.

Reduce no-shows first, with card guarantees and reminders, and overbook on what remains. Keep separate tables by day of week and by channel. Track walks per 1,000 arrivals beside occupancy.

Plan the walk in the morning, when it is already visible. Protect stayovers, loyal guests and late arrivals. A manager handles it: a comparable room nearby, a car both ways, and the telephone call.

Overbook the k-th room while P(no-shows ≥ k) ≥ W ÷ (W + E). W is the cost of a walk, E the contribution of a room.

Groups and displacement

A group is worth taking when its contribution exceeds the contribution of the guests it displaces. Count only the rooms actually displaced, at the rates those guests would have paid, and add the group’s own food, beverage and room hire.

A 300-room hotel expects 150 transient guests at 100, with a variable cost of 25 a room. A group asks for 220 rooms. It displaces 70. The break-even group rate is 48.86. That is the floor, and the quote should sit above it.

Give a corporate account last-room availability only if its rate clears what the last room is worth on your peak nights.

Break-even rate = variable cost − group ancillary contribution per room + displaced rooms × (transient rate − variable cost) ÷ group rooms.

Distribution

Rank channels on what you keep, not on what the guest pays. Net ADR is the rate after commission and the fixed cost of the booking. Acquisition cost has risen to between 15 and 25 percent of what guests pay (Kalibri Labs).

An online travel agent is a paid acquisition channel. The claim that a listing lifts direct bookings is true but small: about 7 percent of shoppers switch to the hotel’s own site. It is no reason to buy volume.

Keep the public rate in parity. Win direct bookings with member rates and with benefits that only the hotel’s own site carries. The arithmetic is in digital marketing for hotels.

Net ADR = ADR × (1 − commission) − fixed cost per booking ÷ length of stay.

Total revenue

Optimise the guest, not the room. A wedding group at a lower room rate with heavy banqueting can be worth more than a full-rate guest who eats elsewhere.

Departments earn very different margins: rooms about 80 percent, food and beverage 30 to 40, room hire 85 to 95. Rank every enquiry on contribution per unit of the scarce thing: the room-night, the seat-hour, the square foot per session.

The rhythm

Pay the sales team on contribution after displacement, and the general manager on GOPPAR and RGI. A team paid on volume will deliver volume.