Revenue Management and Pricing Analytics: A report from the INFORMS Annual Meeting

Last week I had the opportunity to attend the INFORMS Annual Meeting in San Francisco.  For those of you not familiar with this organization or conference, the Institute for Operations Research and the Management Sciences (INFORMS) is the largest society in the world for professionals in the field of operations research (O.R.), management science, and analytics.  Several thousand papers are presented at a typical INFORMS Annual Meeting, which is held over three to four days, and includes many separate tracks covering distinct areas of focus.  The plurality of presentations at an INFORMS Annual Meeting seem to come from academic sources, but there are also papers presented by vendors and industry.  This mix of paper sources makes the conference an excellent place to see where research is being focused (by academics, industry, and vendors), while at the same time gauging where the current “state of the art” is in a given area.  Unfortunately, because the conference is so large, it is virtually impossible to see more than a very small slice of it, which makes getting general impressions across disciplines difficult.  In my entry today I’m going to focus very specifically on Revenue Management and Pricing, as this was my focus at the event.

This year’s conference included 83 separate tracks, ranging from Aviation Applications to Service Science.  It seemed that every available meeting room was taken up by the conference, and most attendees agreed that this year’s event was larger (in terms of numbers of sessions) and better attended than others in recent years.  There were two (sometimes three) tracks dedicated to Revenue Management and Pricing, but several very relevant presentations on revenue management were held in other tracks, so I found myself shifting tracks quite a bit to catch those topics that interested me most.

My own presentation was held in the Practice track, and all three of the presenters in this section focused on revenue management.  I presented “Revenue Management in the Big Data Era” where I focused on the industry and technology changes that are driving new challenges to revenue management, and how Big Data appears to offer the opportunity to face these challenges more successfully in the future – when that data is paired with new analytic approaches.  In my presentation I referenced several studies that are showing the potential value of big data to revenue management, including Kelly McGuire and Breffni Noone’s studies on Hotel Pricing in a Social World.

The two other presenters in my practice section were Warren Lieberman from Veritec Solutions (Warren also chaired our session, which was titled “Succeeding with Revenue Management”), and Cory Canamo from Disney Cruise Line.  Warren’s presentation “Models and Methods” covered a number of topics, including comparing the benefits of centralized vs. decentralized revenue management organizations, and the availability data and importance of using a range of data to drive pricing decisions – which dovetailed nicely off of my own piece.  Warren then described the benefits that Veritec’s customers are seeing from a multiple signal approach.  Cory’s presentation provided an overview of Disney Cruise Line, and some of the unique challenges that Disney Cruise Line faces as a niche, family-oriented cruise line.  Cory also discussed some of the industry innovations that Disney Cruise Line has introduced (such as the Magical Porthole), and how customer reactions to these innovations has challenged the revenue management function at Disney Cruise Line.

This year saw a large number of papers at the INFORMS Annual Meeting, with 27 papers involving SAS authors, presenters, or topics.  A number of these presentations related to revenue management and pricing, including Matt Maxwell’s two presentations “Customer Choice Model Optimization with Overlapping Consideration Sets”, and “Optimization Challenges with a Customer Choice Model”.  Matt’s presentations highlighted:

  • Limitations in traditional revenue management forecasting approaches - in particular the assumption of independence between demand streams
  • How customer choice model approaches overcome these limitations, and
  • Challenges that must be overcome to optimize revenues using customer choice models

Jason Chen’s (also from SAS) presentation “A Simulation Study of BAR by Day Heuristics” discussed the complexity in optimizing BAR (best available rate) by Day (where room prices are the same for a given date of stay for all occupants, regardless of length of stay).  As Jason explained, BAR by day is attractive due to its simplicity, but it presents challenges when trying to optimize revenues while accounting for length of stay effects.  Jason presented the outcomes of simulation studies used to gauge several different approaches to optimizing BAR by day pricing.

One last presentation that I thought would be interesting to regular readers here at the Analytic Hospitality Executive: “Upgrades and Upsells: Hertz vs. Hilton Models” by Guillermo Gallego from Columbia University.  Upgrades and upsells is an area that has received quite a bit of attention, not only in hotels, but also in airlines (due to the increased number of seat inventory types caused by the addition of extended-legroom coach seating).  Dr. Gallego’s presentation discussed the origin of upgrading / upselling (mismatching demand and inventory types) and compared several different business strategies to maximize revenue.  Dr. Gallego also discussed the issue of “strategic behavior” (i.e. customers choosing not to pay for paid upsells when free upgrades are frequent), and the impact on optimization and outcomes.  One of the things that struck me about this particular presentation was the value that was gained from improving inventory utilization (occupancy) when using just a simple free upgrading policy – value that I think is being lost or forgotten as hotels have begun to follow paid upgrade policies, because the value gained from the payment is simply easier to see.

Overall I had two general impressions regarding revenue management and pricing from this conference:

  1. Choice modeling continues to be a very hot topic amongst researchers, with at least 5 separate sessions (3-4 presentations per session) in the revenue management tracks dedicated to Choice modeling in revenue management – and a handful of other sessions outside of the official RM tracks covered this topic.
  2. I was also intrigued and encouraged at the large number of presentations at the conference regarding the application of game theory in revenue management.  One of the general criticisms of traditional revenue management approaches has been the focus on short-term revenue / profitability, and the inability to capture the complexities of strategic decisions that need to be made in a competitive environment, where competitors respond to actions.  Based on what I saw at INFORMS, researchers are clearly beginning to take a hard look at these very issues.  You might want to study up on “Nash Equilibriums” before attending your next revenue management conference.

The most anticipated session that didn’t actually occur was the planned keynote by Google project leader Anthony Levandowski. Anthony was to present on Google’s work on the driverless car, but unfortunately this session was cancelled at the last minute.  An audible “No!” was heard throughout the conference center when the email went out to all attendees, less than an hour prior to the event, indicating that the session was cancelled – leaving all of us with the question “If they can invent a driverless car, how come they can’t run a speaker-less keynote session?”  (Full credit for this quip goes to my colleague Ed Hughes).

Post a Comment

Big data and big analytics are a big opportunity for hotels – Part 2

Big data is of no use unless you can turn it into information and insight.  For that you need big analytics.  Every piece of the analytics cycle has been impacted by big data, from reporting, with the need to quickly render reports from billions of rows of data, through advanced analytics like forecasting and optimization, which require complex math executed by multiple passes through the data set.

Without changes to the technology infrastructure, analytic processes on big data sets will take longer and longer to execute.  It’s not enough now to push the button and wait hours or days for an answer.  Today’s advanced analytics need to be fast and they need to be accessible.  This means more changes to the technology infrastructure to support these new processes.

Analytics companies like SAS have been developing new methods for executing analytics more quickly.  Below is a high level description of some of these new methodologies, including why they provide an advantage.  Once again, the intention is to provide enough detail to start conversations with IT counterparts (or understand what they are talking about), certainly not to become an expert.  There is a ton of information out there if you want more detail!

  1. Grid computing and parallel processing – Calculations are split across multiple CPUS to solve a bunch of smaller problems in parallel, as opposed to one big problem in sequence.  Think about the difference between adding a series of 8 numbers in a row versus splitting the problem into in four sets of two, and handing them out to four of your friends.   To accomplish this, multiple CPUs are tied together, so the algorithms can access the resources of the entire bank of CPUs.
  2. In-database processing - Most analytic programs lift data sets out of the database, execute the “math” and then dump the data sets back in the database.  The larger the data sets, the more time consuming it is to move them around.  In-database analytics bring the math to the data.  The analytics run in the database with the data, reducing the amount of time-consuming data movement.
  3. In-memory processing – This capability is a bit harder to understand for non-technical people, but it provides a crucial advantage for both reporting and analytics.  Large sets of data are typically stored on the hard drive of a computer, which is the physical disk inside the computer (or server).  It takes time to read the data off of the physical disk space, and every pass through the data adds additional time.  It is much faster to conduct analysis and build reports from the computer’s memory. Memory is becoming cheaper today, so it is now possible to add enough memory to hold “big data” sets for significantly faster reporting and analytics.

To give you an idea of the scale of the impact, applying these methodologies, we have been able to render a summary report (with drill down capability) from a billion rows of data in seconds.  Large scale optimizations like risk calculations for major banks, or price optimization for thousands of retail products across hundreds of stores, have gone from hours or days to minutes and seconds to calculate.  As you can tell, the advantages are tremendous.  Organizations can now run analytics on their entire data set, rather than a sample.  It is possible to run more analyses more frequently, testing scenarios and refining results.

Here are some examples of how innovative companies are applying big analytics to get value from their big data:

  • Airline companies are incorporating the voice of the customer into their analyses, by mining all of the internal and external unstructured text data collected across channels like social media, forums, guest surveys, call center logs, and maintenance records for passenger sentiment and common topics.  With big text analytics, these organizations are able to analyze all of their text data, as opposed to small samples, to better understand the passenger experience and improve their service and product offerings.
  • A major retailer is keeping labor costs down while maintaining service levels by using customer traffic patterns detected by security video to predict in advance when lines will form a the register.  This way, staff can be deployed to various stocking tasks around the store when there are no lines, but given enough notice to open a register as demand increases, but before lines start to form.
  • A major hotel company has deployed a “what if” analysis in their revenue management system which allows users to immediately see the impact of price changes or forecast overrides on their demand, by re-optimizing around the user’s changes.  Revenue managers no longer have to make a change and wait until the overnight optimization runs.

Unlocking the insights in big data with big analytics will require making some investments in modernizing technology environments.  The rewards for the investment are great.  Organizations that are able to use all that big data to improve the guest experience while maximizing revenue and profits will be the ones that get ahead and stay ahead in this highly competitive environment!

Post a Comment

Big data and big analytics are a big opportunity for hotels

By infusing analytics through every phase of the guest journey, hotel managers can help shore up the complicated balance between the guest experience and revenue and profit responsibilities – delivering memorable and personalized guest experiences, while maximizing revenue and profits.  To accomplish this, hotels need to be able to collect, store and analyze the volumes of data generated by their guest interactions, their operations and the broader market.    As the volume and complexity of data increases, wrapping your head around what is available and how it can be useful is becoming challenging.

“Big Data” is a challenge for organizations not just because the volume of data has increased, but also because the variety has increased – it’s gone beyond traditional transactional data into unstructured formats like text, video, email, call logs, images, click-stream.  It is coming at us fast.  Data like tweets or location is stale nearly the minute it is created.

The reason why big data is a “big deal” is because the volume and complexity of the data puts pressure on traditional technology infrastructures, which are set up to handle primarily structured data (and not that much of it).  In these environments, it is difficult, or even impossible, for organizations to access, store and analyze “big data” for accurate and timely decision making.  This problem has driven innovations in data storage and processing, such that it is now possible to access more and different kinds of data.

To a certain extent, big data is forcing business leaders (like our analytic hospitality executives) to get more involved in technology decisions than ever before.  To help with this, in this post, I’ll talk about how technology has evolved to handle big data and give some examples of how companies are innovating with their big data.   Next week, I’ll do the same for big analytics.  This is not intended to make everyone into technology experts, but rather, to provide some basic information that can arm hotel managers to start having conversations with their IT counterparts.

This influx of large amounts of complex data has necessitated changes in the way that data is captured and stored.  To handle the volumes of unstructured data, databases need to be faster, cheaper, scalable and most importantly more flexible.  This is why some have been talking about Hadoop as an emerging platform for storing and accessing big data.  Hadoop is a database that is designed to handle large volumes of unstructured data.    Hadoop works because it is cheap, scalable, flexible and fast.

  1. Cheap & Scalable – Hadoop is built on commodity hardware – which is exactly what it sounds like – really cheap, “generic” hardware.  It is designed as a “cluster”, tying together groups of inexpensive servers.  This means it’s relatively inexpensive to get started, and easy to add more storage space as your data, inevitably, expands.  (it also has built in redundancy – data stored in multiple places,- so if any of the servers  happen to go down, you don’t’ lose all the data)
  2. Flexible – The Hadoop data storage platform does not require a pre-defined data structure, or data schema.  I use the analogy of the silverware drawer in your kitchen.  The insert that sorts place settings is like a traditional relational database.  You had to purchase it ahead of time, planning in advance for the size of the drawer and the kinds of silverware you wanted to put in it.  It makes it easy for you to grab out the four sets of forks and knives you need for a place setting.  However, the pre-defined schema makes it difficult to add additional pieces of silverware should you decide to buy ice tea spoons or butter knives, or if you are looking for a place to store serving utensils.  Hadoop, on the other hand, is more like an empty drawer with no insert – it has no pre-defined schema.  You can put any silverware you want in there without planning ahead of time.  You can see the advantage of this approach with unstructured data.  There is no need to “translate” it into a pre-defined schema, you can just “throw it in there” and figure out the relationships later.
  3.  Fast – Hadoop is fast in two ways.  First, it uses massive parallel processing to comb through the databases to extract the information.   Data is stored in a series of smaller containers, and there are many helpers available to reach in and pull out what you are looking for (extending the drawer metaphor: picture four drawers of silverware with a family member retrieving one place setting from each at the same time).  The work is split up, as opposed to being done in sequence.  The second way that Hadoop is fast is that because the database is not constrained by a pre-defined schema, the process of loading the data into the database is a lot faster.  (picture the time it takes to sort the silverware from the dishwasher rack into the insert, as opposed to dumping the rack contents straight into the drawer).

Many companies have put a lot of effort into organizing structured data over the years, and there are some data sets that make sense to be stored according to traditional methods (like the silverware you use every day).  Because of this, most companies see Hadoop as an addition to their existing technology infrastructure rather than a replacement for their relational, structured, database.

Next week I’ll talk about innovations in the execution of analytics that speed up time to results, allowing organizations to take full advantage of all of that big data.

 

Post a Comment

What can hotels learn from casinos? Focus on non-rooms pricing.

This week I’m going to “fill in” for Natalie to complete our series on “What can Hotels learn from Casinos”?  (Look for parts 1 and 2 here and here, respectively) Today I’m going to address pricing. We all know that great entertainment comes at a price, and we also know that casinos need to manage demand and supply just like any other business. But when we are managing revenue in a casino, there are a lot of interrelated factors at play. The denomination or price of a slot machine can impact demand for that machine. The price and availability of hotel rooms need to be balanced with the ability to offer comp rooms to the casino’s most valuable players, as well as with the overall goal to drive revenue throughout the estate.  After all, the modern casino resort produces revenue across many fronts (rooms, entertainment, gaming, food and beverage, and more), so any decision that impacts room rates and availability has to consider the broader effect on estate-wide revenues.

Predictive analytics, like forecasting and optimization, have been used in revenue management applications in travel, hospitality AND gaming companies for some time. But the traditional revenue management approach is revenue-focused, not customer focused.  Furthermore, typical revenue management applications in hotels (and most hospitality and travel segments) have a notoriously short-term focus: that is, optimization focuses on optimizing revenues for a given date or date range – the longer-term impact of those pricing decisions is not considered.  As a business built for entertainment, casinos have focused their use of predictive analytics in revenue management not only on their patrons, but on a variety of other factors that impact pricing.

Let’s look at a few examples. First let’s consider the pricing of hotel rooms. In Natalie’s first entry in this series, she talked about the use of analytics to derive the customer lifetime value.  This is an important element for a casino when they decide to make a room available for a patron.  Casino patrons tend to play where they stay, and high-roller customers can make several trips to a casino each year. Traditional revenue management would optimize the best mix of business and prices for the hotel for any given day and length of stay. When you consider that the more you play at a casino, the more valuable you are and the less you may pay for a room (because discounting for “high rollers” is a common practice at most casino resorts), a typical rooms-revenue-focused optimization would likely recommend closing out the rates for the casinos’ most valuable, frequent patrons – and cost the casino valuable gaming and entertainment revenue.  Furthermore, if a casino doesn’t have room for its most valuable patrons to stay when they are here in Vegas, those patrons will likely stay and play at another hotel and casino– and this can cause valuable repeat customers’ business to be lost.

Today, the modern casino is factoring customer lifetime value into its revenue management and price optimization processes.  By segmenting their forecasts into patron value buckets that consider both estate-wide spending and frequency of visitation, versus by business types, casinos can ensure that the optimization process leaves rooms available for their most valuable patrons.

Our second example is with the pricing of tickets for shows. As noted earlier, most casino resorts have morphed into multi-facetted entertainment destinations.  Nowhere is this truer than in Las Vegas, where we continue to see an array of entertainment options whose goal is to attract guests to the vicinity and keep them entertained – the longer you stay, the more you play (and pay).  The expanding options in Las Vegas include the recently-constructed High Roller (billed as the world’s tallest observation wheel), and the recent announcement by MGM of a new 20,000-seat arena to be built just off the strip. And of course, when not gaming, watching a show, or going to a concert, you can also eat in a world-class restaurant run by a celebrity chef.

Of course, Las Vegas shows are significant revenue-generators.  With multiple shows per day, often seven days per week, with hundreds of seats available per show, and operating practically year-round, these shows can generate a lot of revenue.  But if casinos try to optimize ticket prices, without considering the impact on the rest of the estate, they may miss an important input into the decision making process; specifically, how do they fill the gaming floor with patrons? As a result, casinos must balance the need to drive ticket revenues with the need to drive patrons onto the property that will use the gaming floor, bars and restaurants. So, when making ticket pricing decisions, they use show occupancy as a goal in their optimization process and end up with a price that meets the twin goals of occupancy AND revenue while factoring in the constraint of the number of seats available in the theatre.

Finally, because casinos have a high transient quotient, casinos are frequently running campaigns to fill rooms during expected low periods – far more frequently than a typical hotel.  Many (though certainly not all) of these campaigns are oriented towards attracting known customers back to the property.  Because such campaigns are so frequent, marketing and revenue management teams in a typical casino environment work hard to:

  • Ensure that campaigns are properly aligned with “need” periods and pricing strategies
  • Estimate the “lift” from each campaign, so that revenue management can properly manage demand and pricing during campaign-affected arrival periods

Does managing all of these disparate revenue streams make casino pricing more difficult?  Certainly!  But I was not surprised when a panel of revenue managers at the G2E conference this year said that they felt that managing revenue for a 200-room hotel was in fact far more difficult than managing revenue for a 2,000 room casino property.  Why?  Because the variability of behavior on the smaller number of rooms makes predictions and revenue management so much more difficult!  The large number of rooms at many casinos leads to much more predictable behaviors across the estate – making most day-to-day pricing decisions much easier.

Casino properties are different from typical hotels – and it isn’t necessary to follow their every practice – but we should consider what can be learned from these practices in managing pricing for hotels, and apply them when and where they make sense.

How much revenue does your property generate from non-rooms revenue sources?  Do you consider the impacts on these sources when determining your rooms pricing decisions?  What about the long-term value of your guests?  Is this also considered in your decisions?

Post a Comment

Text Analytics: The voice of the customer at the speed of business

I’ve written about using text analytics to capture comprehend and act on the voice of the customer on this blog before. Recently, I recorded a short presentation that outlines how text analytics provides insight from all of the customer experience data that you collect, not just review data from social media sites but also internal data such as call logs and surveys.

Natalie Osborn discusses using Text Analytics to discover insights about customers.

If you are attending the Cornell Hospitality Research Summit, I’ll be presenting on the voice of the customer on Tuesday 14 October, 2014 at 1:15pm in Statler 398. If you are not attending, but would like more information on this topic – please download the SAS white paper – “Unleashing the Power of Text Analytics for Hotels”.

Post a Comment

Can big data help revenue management?

Next week I will be presenting at the Cornell Hospitality Research Summit on big data and revenue management. In preparation for the Summit, I have recorded a short presentation that outlines how big data can help augment revenue management.

Alex Dietz talks about how big data can help revenue management.

If you are attending the Cornell Hospitality Research Summit, I’ll be presenting on big data and revenue management on Tuesday 14 October, 2014 at 10:30AM in Statler 398. If you are not attending, but would like more information on this topic – please read my two blogs: Big data and revenue management part one and part two.

 

Post a Comment

Hotel Pricing in a Social World: The unmanaged business traveler

Next week Breffni Noone, Associate Professor, The Pennsylvania State University and I will be hosting a discussion at the Cornell Hospitality Research Summit on how user generated content impacts the purchase choices of the unmanaged business traveler. The discussion will take place on Monday October 13 at 2:15pm in Statler 396.

In preparation for this discussion, Breffni and I recorded a video presentation of our research. If you are attending the conference – you should plan to watch this in advance of the discussion. If you are not attending the conference, but visiting from twitter, Facebook or Linkedin – we’d love to hear your feedback and would be happy to answer any questions that you have in the comments section of this blog.

Kelly McGuire and Breffni Noone present the results of their research.

For further information – see my previous post which outlines the details of this study.

Post a Comment

From the desk of the CHRS: We love innovation, right?

By Cathy Enz, PhD., Lewis G. Schaeneman, Jr. Professor of Innovation & Dynamic Management, Cornell School of Hotel Administration.

It seems that everyone loves to talk about innovation these days, but sadly the overuse of the word to describe minor tweaks and ordinary activities has turned it into a confusing if not meaningless term.  While senior leaders in many different industries use the word at alarmingly frequent intervals, I wonder if they mean the same thing I do when I talk about innovation.  For me, innovation has its roots in the notion of creating a new idea or the rearranging of an old one in such a way that it creates significant customer value.  However this business of new ideas is not as easy to sell as we might think.  You history buffs know it took the British Navy of the 17th century over 150 years to adopt a proven cure for scurvy, but in today’s world we embrace new ideas – right?

I have begun to wonder as we prepare for the Cornell Hospitality Research Summit coming in October (CHRS14) if people really like new ideas.  The catalyst for my reflection is the decision Rohit Verma, my co-chair for the summit, and I made to encourage new methods for the presentation of proposals for the summit.  This year we have asked moderators and speakers to experiment with different session types like posing a “big question” to spark debate, or both showing and telling participants about a new practice, process, program, product, or application.  We have tried to ask folks to step away from the comfort of a traditional formal PowerPoint dependent presentation to be creative, or to more fully explore a devil’s advocacy style to question the positions of others.  In short, we wanted to be innovative and experimental in both the theme of the conference and the way in which we shared information at the conference.

Here is what we have discovered in trying to execute on this idea.  First, many presenters don’t automatically accept, understand, or really want to try this new idea.  In some ways their hesitancy makes perfect sense.  Who likes to take unnecessary risks, or move out of their comfort zone, particularly if the new idea is unproven, or ambiguous, and the probability of success is uncertain?   New for the sake of newness is not necessarily a good thing.   The second thing we have discovered is that how the creator of an idea sees it can be quite different from how it is seen by others.  In fact this gap between the acceptance of a new idea by others and the expectation that people will embrace something new is perhaps one of the biggest challenges innovators face.

It is too soon to tell but our desire to improve on the conference method may not be embraced with enthusiasm, as we had hoped.  Watching from a safe distance, and following once the idea is clearer and proven seems prudent, and we get it.  I’m reminded of the inventor Howard Aiken who once said, “Don’t worry about people stealing your ideas.  If your ideas are any good, you’ll have to ram them down people’s throats.”  We of
course don’t want to ram our idea for an experimental conference down anyone’s throat.  In fact, we want to thank all of the folks who asked for clarity (see our do’s and don’ts video, including our bad acting below) and we are grateful to those of you who bravely promised to give it a go.  We are looking forward to a great conference, but I would venture that just maybe people don’t love new ideas.  See you in October, and let’s find out.

 

As the leading source for research on and for the hospitality and related service industries, the Cornell University School of Hotel Administration invites representatives of industry and academe to the 3rd Cornell Hospitality Research Summit, on the beautiful campus of Cornell University in Ithaca, New York, October 12-14, 2014. A conference unlike any other, the CHRS is designed to create new knowledge through the intentional interaction of industry and academic presenters and participants.

Post a Comment

How Business Travelers Buy: Hotel Pricing in a Social World

With the growing popularity and availability of online reviews and ratings, consumers have more information than ever before when purchasing a hotel stay.  In order to build effective pricing and positioning strategies, hotel managers need to understand how consumers are using all of this user generated content with price to make a purchase decision.  Dr. Breffni Noone, Associate Professor at Penn State, and I have done a series of studies to understand how consumers use all of this information to assess value, and ultimately, make a purchase decision.  Our latest research looked at unmanaged business travelers.

As in previous studies, we used a choice modeling experiment.  In this technique, researchers select a set of attributes of a product or service they wish to test, and subjects are presented with a set of options with varying combinations of levels of those attributes.  They are asked to select the option that is most attractive to them (the one they would buy).  As participants repeat this exercise over different sets of options, it is possible to statistically assess the importance of each attribute in decision making, and the value that they place on the attribute and its levels.

The study design

For this study, we told our business traveler participants that they were attending a meeting in a city center.  We told them we would show them a set of hotels that were “business-friendly” with equivalent class of service and amenities, all with locations equally convenient to the meeting.  We then showed them a set of three hotels in which the price (low, medium, high), review sentiment (positive, negative), review content (physical property, service) review language (emotional, descriptive), aggregate rating (low, medium, high), TripAdvisor Rank (low, mid-range, high) and brand (known, unknown, preferred) varied.  We asked them to select the hotel they would book from the set of three options, and they repeated that exercise three times.

The data was gathered via an online survey distributed to a representative sample of the US population.  We screened for participants who traveled on business at least six times per year, and were able to make the choice of where to stay themselves (i.e. not overly constrained by a corporate travel policy).  Since we knew that loyalty program affiliation could be influential in traveler decision making, we asked them to tell us what loyalty programs they belonged to, and to identify a preferred brand from within those loyalty programs.  We presented a list of brands within these loyalty programs that met the study criteria of “business-friendly” hotel, so for example, the Hilton Honors brands were Hilton, Embassy Suites and Doubletree.  Note that this was not a study about loyalty program membership, so we can’t really make inferences to the general population of business travelers, but it is interesting to see how these business travelers are affiliated with common loyalty programs.

Business traveler loyalty behavior and demographics

Figure 1 shows the distribution of membership and the distribution of preferred brands by loyalty program.  We also asked whether they belonged to a “non-brand” loyalty program (like hotels.com or Leading Hotels), and you can see that 45% of them were members of such programs.

Figure 1: Loyalty Program Membership of Sample

Figure 1: Loyalty Program Membership of Sample

 

Figure 2 shows the demographic composition of the business travelers in this study.  You can see that about half of them take 6-10 trips a year, and the vast majority stay two or more nights per business trip.  Interestingly, the vast majority of these business travelers do read reviews and say they are influenced by them.

Number or cards held, which programs

Figure 2: Business Traveler Study Demographics

Study Results

Overall Attribute Importance

The first output of a choice modeling study is the overall importance of each attribute in driving customer’s decision making.  The following list of attributes had a significant impact on value perceptions, and they are presented in order of importance to the business traveler:

  1. Review Sentiment (positive, negative)
  2. Brand
  3. Aggregate Ratings
  4. Price
  5. Review Language (descriptive, emotional)

Contrast this with the list for the leisure traveler:

  1. Review Sentiment (positive, negative)
  2. Price
  3. Aggregate ratings
  4. TripAdvisor Rank
  5. Brand

As you can see, there are already some interetsing differences in how these two segments assess value and make decisions, yet review sentiment is  most influential for both segments.

Influence of attribute levels on value

The next step in a choice modeling analysis is to understand how value assessments vary with the different levels of the attributes.  In other words, how do negative versus positive reviews impact value?

The following results were obtained from the analysis of value impact by attribute level:

  • When reviews were negative as opposed to positive, not surprisingly, there was a large negative impact on value perceptions
  • There was a small positive impact of the known brand over the unknown brand, but a large positive impact on value of the preferred brand over the known brand
  • While the value impact associated with both a mid-range aggregate rating over a low rating and a high rating over a mid-range aggregate were significant and positive, the impact of mid over low was actually greater than the impact of high over mid
  • There was a negative impact on value when the price was raised to the mid-level over the low price.  Business users did not notice the difference between the mid and high price
  • Finally, there was a slight positive impact on value of a review with descriptive language (“The bed was very comfortable”, “The room was cold”) over emotional language (“I LOVED the bed”, “The temperature in the room was annoying”)

Looking at the aggregate rating results suggest that it is more important to business travelers that the hotel is OK, than great.  They will assess what the experience will be like (by reading the reviews, especially descriptive ones), and as long as they feel it will be OK (and they can get points), they will book.  This doesn’t mean they don’t appreciate a great hotel, but in the balance between staying with their preferred brand that is just "OK", and a hotel with “great” reputation, they are likely to choose the preferred brand.  While business travelers are relatively price-insensitive, they do respond to a deal.  Hotels with an “unknown” brand that are in a better reputation position than the “preferred” hotels in their market (the ones with well-known loyalty programs), might be able to encourage business travelers to stay by offering a deal.  However, if the price isn't among the lowest prices in the competitive set, the business traveler is not likely to notice.

Assessing overall value

Choice modeling analysis also allows you to look at a combination of attribute levels and assess the overall value to a consumer of an option with that set of attribute levels.   We are going to present a set of equations below representing an hotel option from the study. The important thing to focus on is not the value of result of the equation, but how the number changes as we change attribute levels.

The below equation represents the most valuable option for the business travelers, in order of attribute importance.

Positive Reviews + Preferred Brand + High Rating + Low Price + Descriptive Reviews = 1.52

With this set of attribute levels as a baseline, we can manipulate the attribute levels and track their impact on overall value.  For example, holding everything else constant, the below equation represents overall value when the price is raised from lowest to highest.  You can see that the overall value is impacted slightly, but not by very much.  This is not surprising since the price was only the fourth most important attribute to business travelers, and it was only the low price that the business travelers valued.

Positive Reviews + Preferred Brand + High Rating + High Price + Descriptive Reviews = 1.25

Contrast this result with the leisure traveler study, where price was the second most important attribute.  The first equation is the most valuable option for the leisure traveler, and the second shows the impact of raising the price from low to high.  Leisure travelers are clearly highly price sensitive.  The overall value drops significantly when the price goes from low to high.

Positive Reviews + Low Price+ High Rating + High TARank+ Known Brand = 1.95

Positive Reviews + High Price+ High Rating + High TARank+ Known Brand = 0.46

Now look at what happens when the most important attribute for each traveler changes.  Holding everything constant from the baseline, the first equation below demonstrates the impact of negative reviews on overall value for business travelers, and the second for leisure travelers.  You can see here that the impact on value for business travelers over the baseline was quite significant, nearly half of the value is lost when the reviews are negative as opposed to positive.  However, for the leisure travelers, all of the value is lost when reviews are negative.  Business travelers clearly pay attention to reviews, but they are willing to balance that against other attributes like brand and rating.  Leisure travelers simply wont consider the hotel if the reviews are negative.

 Business: Negative Reviews + Preferred Brand + High Rating + Low Price + Descriptive Reviews = 0.69

Leisure: Negative Reviews + Low Price+ High Rating + High TARank+ Known Brand = 0.01

 

Summing it all up

So, what does all this mean?

There are a couple of key takeaways from this study:

  • Reviews matter – business travelers look to the reviews to assess what their experience will be like.  If the review is positive or negative they want to know why.
  •  Loyalty matters – Business travelers will put up with “good enough” or “OK” if they can get their points
  • Price matters – Business travelers still recognize a deal, but it’s only the lowest price that entices them.  They are relatively insensitive beyond that.

Figure 3 summarizes the key takeaways from this study as compared to the leisure study.  These takeaways present several opportunities for hoteliers.  For example, a hotel that has a heavy mix of business travelers who are members of their loyalty program might be able to put off renovations for a while (assuming the declining quality of the product is reflected in declining reviews and ratings), but a leisure property would not.  A brand that is relatively unknown in the market, could attract business travelers to forgo their preferred brand if their reputation was better, by offering a “deal”.

The bottom line is that pricing in today’s social world is not getting any easier.  Not only do you have to understand your price relative to the market, and your reputation relative to the market, but you also need to understand your business mix.   All of these decisions require good knowledge of the market backed by solid revenue and reputation analytics.

Figure 5: Key Study Takeaways

Figure 3: Key Study Takeaways

 

 

 

 

Post a Comment

What can hotels learn from casinos? Focus on the service environment.

What can hotels learn from casinos? In my first post on this topic I explored how casinos apply predictive analytics to help them focus on their customers. In this post I will explore how casinos use analytics to help manage the service environment with the dual goals of improving profitability and customer experience. But first, let’s kick off with some of the crazier myths about how casinos manipulate their physical environments to make you, the patron, gamble more.

I’ve heard it claimed that casinos pump hyper-oxygenized air into casinos to keep your mood buoyant and attention focused on gaming. I have also heard claimed that there is a big red button in the basement that can dial up or dial down how much patrons can win! Neither of these rumors are true, but there are a lot of moving parts in a casino.  Casinos do manage their service environment, and they use analytics and data management to help them do it. They focus on things such as the location and placement of the slot machines and gaming tables, the denomination of those machines and minimum bets of the tables, as well making sure that they have the skilled team members to manage these experiences and also the patron themselves.

When it comes to planning the operations of a casino floor, bad decisions can mean significant losses in customer loyalty and potential revenue. The challenge is how to plan the right mix of gaming choices, denominations, and table or machine placement to optimize the patron’s interest. In Canada – 85% of gaming revenue is made up of slot machine revenue. As a result, slot operations are a primary focus of their analytics.

Saskatchewan Gaming Corporation, operator of the Casino Regina, has placed a lot energy and analytics into how they manage their slot operations. They pulled together all of the transaction data on their slot machines and created a best-case predictive forecast into how each game would perform in the year to come. In the process, they began collecting insights into leading predictors of patron preference for machines. This helps them optimize profitability. They also looked at how they can predict the impact of potential changes to slot performance based on ‘what if’ scenarios.

Finally, Saskatchewan Gaming Corporation used advanced optimization to determine the best approach to future business, considering factors such as physical space and budget. This information helps the casino company to optimize its slot machine purchase options, including the analysis of which machines to replace and when to replace them, while also ensuring the patron experience is not hampered through excessive machine down-time. Analytics allows this casino company to offer the right games, in the right locations to attract loyal and valuable patrons.

Optimization techniques are not isolated to machines for casinos but can also be used to ensure that the right service operations staff are available to serve patrons. It’s important that if you offer a patron a meal, that you have a seat available in the restaurant for that patron. Patrons who have to wait for service in a restaurant or a seat at a gaming table may be tempted to take their business elsewhere. Forecasting demand for each service area, and matching the right skill set to the area is very important for casinos. Just think of all of the different skill sets required to run a gaming floor, from the different dealers to the cashiers, pit boss and machine technician. Optimization approaches are needed in every aspect of casino operations.

Casinos also carefully manage the patron’s service experience while they are in the casino. Casinos use master data management techniques to ensure that they can use what they have learned about their patrons to enhance the patron experience. Master data management can deliver the results of analytics to the service operations team members, to ensure that each patron is recognized and treated appropriately and most importantly - consistently.

Managing the patron’s service experience is particularly important when something starts to occur that none of us really want to think about when we are in a casino, and that is losing. If a patron hits a losing streak their enjoyment may start to decrease.  The patron may decide to leave the casino and “try their luck” somewhere else. Casinos know if they offer patrons a free meal or a drink at the bar, allowing the patron to step away from the losing experience and focus on something else, the patron will start to enjoy themselves again. Casinos trigger alerts to their service teams to identify patrons who are losing. These alerts can be informed with analytic such as customer lifetime value and next best offer, which helps the service team take the appropriate action to correct the customers experience while at the same time remaining profitable.

While the loss of a casino patron may be much more “of the moment” than that of a hotel guest, the use of analytics and data management to help detect and avoid that act of attrition is something that all hotels could benefit from. How are you detecting behavior changes in your guests? Do you have steps in place to identify when the customer experience is going wrong, or when the customer is about to leave you?

Casinos use master data management (MDM) techniques to communicate important customer preference information to staff who sit at each interaction point. Master data management is the processes, governance, policies, standards and tools that consistently define and manage the critical data of an organization to provide a single point of reference.  One of the benefits to service organizations of using MDM is that when that single point of reference is a customer profile, the master data can ensure that the treatment of customers is consistent and that preference information reaches all customer points of contact.

How do you ensure the service experience is consistent across your operation and throughout your estate? Are you relying on information hidden in the comments field of your property management system? Are there small “moments of truth”, such as the preferred pillow or using the preferred method of address that mean the guest stays satisfied?

When you proactively manage the service environment, you can deliver on a great guest experience and avoid customer attrition. Casinos use solid data management techniques to help communicate their guest’s preferences and predictive analytics to identify moments when intervention is required.

No hyper-oxygenized air required.

Post a Comment