Restaurant Revenue Management: Could it work?
Sheryl E. Kimes
Journal of Revenue and Pricing Management
Article is available here
Revenue Management is pretty well-used in modern hotels and its benefits are widely recognized. Therefore, I wonder, is it possible to apply some of the ideas from airlines/hotel revenue management into restaurant industry? As I always emphasize, forecasting is the core part of Revenue Management that enable dynamic pricing. I hardly imagine dynamic pricing in Restaurants, although can we somehow use the knowledge about probable demand in certain days/hours? I wanted to confront my own ideas with some academic conclusions and I found this article. Following text is the most important arguments I learnt from it.
Sheryl E. Kimes – author of this article – is well known in RM society. She has written numerous of articles in hospitality journals as well as she is invited for international conferences. In this article she point out main concepts of hotels/airlines revenue management that could be applied to restaurant industry.
The restaurants are similar to hotels is a sense that both has basically two types of demand. Reservation requests and walk-ins. However, number of walk ins varies depending on the popularity and class of restaurant, but usually it is still bigger issue than in hotels.As a result, if we want to apply revenue management in restaurants, we need to make decisions in 5 areas:
- forecasting – the expected number of reservations and walk-ins
- allocations – the number of different-sized tables
- restrictions – who is litigable to book a table (e.g. size of the group)
- protection levels – who many tables should leave for VIP clients or walk-ins
- overbooking – how to prepare yourself for “no shows”
Let’s now look more carefully inside of those areas and think whether RM tools and methods used in hotels could be applied in this case.
The revenue management doesn’t exist without proper forecasting methods. However, it might be more difficult than in hotels, where we only forecast the final number of bookings at particular class. In restaurants, it is more complicated as we want to forecast not only daily number of covers (customers) but also hour and party size. Moreover, as we know, demand should be assess based on unconstrained data, that’s way forecasting is even more problematic, because restaurants do not track turned-down reservations or walk-ins. Additionally, there is also a group of potential customers that didn’t bother to do reservation because they assumed that there will be no tables available. Unconstraining the demand might to be done by tracking the denials or statistical methods. More about uncontraining the demand is available: here. When the demand is unconstrained, the second big task is choose the appropriate forecasting method. Some of them are explained: here, but the real issue is hourly forecast, which neither me nor the author of this article have a clear idea how to do it.
The second difference between hotels and restaurants (in context of revenue management) is fact, that restaurant’s capacity is somehow flexible. Imagine, that in certain restaurant there is place for 20 tables. Every table has place for 4 guests. But the way we allocate tables might vary. There might be 10 double tables for 6 guests each or even bigger. The more tables we allocate, the less capacity we have. Of course, restaurant has strategy for this allocation, but applying historical knowledge would improve expectations and as result increase revenues.
Restaurants have different possibilities of restricting demand compared to hotels (where minimum length of stay, or pre-pay restrictions are used). Those restrictions (especially maximum length of stay) would have negative impact on customer satisfaction. However, the restriction that customers would understand and accept is that a table for four people might be occupied by at least three (or even four). This way, restaurant would increase it occupancy and revenues.
Protection level is a very important RM issue, however, it also creates a lot of controversy. Basically, it means, that restaurant wants to secure some of its resources for valuable customers (and this creates controversy as we assume that richer customer is more valuable). Restaurants also deal with this problem, as they need to decide how many tables will be set aside for important customers or walk-ins. I believe (and this is my personal opinion) that certain restaurant should very carefully study behavior and customs of its customers. To do so, I suggest following steps:
- study historical data from at least a year period
- try to divide customers into categories, based on:
– total spendings
– spendings per person (useful when groups occur)
– time of the day
- Find whether there are customers that are valuable enough to create protection levels for them
- If so, study their buying behavior to to predict when on average they appear
- Protect some of the seats at the time they usually appear.
Overbooking is the last but not least part of RM. Because of no shows, restaurants lose much more money than hotels or airlines because hotels are pre-paid and the owner only lose money because he/she have empty room of capacity. However, restaurants not only lose this opportunity of extra revenue, but also revenue from customers that doesn’t appear. Moreover, there is higher chance that someone will not appear when this person had no pre-payment. What is also important, overbooking in case of restaurants has lower negative impact when a customer is turned down in restaurant than in hotel. If the capacity of hotel is full, the fair thing to do is to find another hotel for this guest and pay for his accommodation. In case of restaurant, we have chance to convince the customer to wait a little longer for the next available table. Therefore, there is a huge potential for overbooking in restaurant industry.
The restaurant industry is about three times as large as the hotel industry
Therefore, there is a huge potential for it. However, as airline/hotel customers are used to price – discrimination or overbooking practices, this is still new concept for restaurants. I believe that bigger challenge will be to accustom customers than implement sophisticated forecasting algorithms or whole revenue management systems.
But the world is changing rapidly, and what is now perceived as impossible, might be seen as common sense in the future.
Kimes, S. (2005). Restaurant revenue management: Could it work?. Journal of Revenue & Pricing Management, 4(1), pp.95-97.