Revenue Management in a Pandemic
Revenue management endured a strange year in 2020. Last March, the terabytes of historical demand data that drove many revenue management system (RMS) algorithms became useless overnight. Someday it will return, but hotels can’t wait. In the meantime, many hotels have rethought their approach to revenue management in ways that will result in permanent as well as temporary changes.
For this week’s article, I spoke with several of the revenue management companies, both established and new, to get a better sense of what’s changing and how they are responding. Perhaps surprisingly, their mood was comparatively upbeat, with several saying they ended up having a “decent” year. In past downturns, revenue management often fell by the wayside when occupancies dropped, as it was seen mostly as a tool for managing high-demand periods. That view appears to have changed, in part because RMSs have gotten smarter in the way they measure and analyze competitive pricing decisions. In periods of oversupply, demand forecasting may become less important, but understanding price elasticity and competitive pricing become more critical.
This article is about the trends and changes I found. It is not meant to rank or compare the providers. It is quite possible that some of the interesting things I heard from one vendor are also available from others, but didn’t come up in conversation, or are offered by one of the many providers with which I didn’t connect. So, if something captures your interest, please ask your preferred vendor(s) – if they’re not already doing it, it may be on their roadmap.
One trend that several providers mentioned was the large variation in COVID’s market impact. While hotels in downtown core and convention markets have suffered greatly, other markets have seen normal, high, or even extreme demand. Destination markets within a 2-3 hour drive of crowded metro areas, such as Key West, the Monterey Peninsula, and the Catskills have proved extremely popular with well-heeled customers who want to escape the city for a weekend (or longer). Some RMSs found that templates for some of these situations were helpful in getting each customer onto the right path.
Hotel companies have made significant organizational adjustments around the revenue management function as they have adjusted to lower staffing levels. In groups with centralized or cluster operations, each revenue manager is likely now handling many more properties. For individual hotels that may have had dedicated revenue managers in the past, revenue management might now be a part-time task for a general manager or rooms division executive.
These organizational trends have led to demands for innovation by the RMS providers: for more automatic decision making, for more flexible reporting, for exception reporting to highlight issues requiring human attention, for the simpler processes and user interfaces needed by non-experts, and for mobile, on-the-go functionality for managers who may only have a few minutes per day to focus on revenue management.
Larger hotel groups are increasingly separating the reporting and analytics aspects of revenue management from the process of setting or recommending rates and inventory controls, using best-of-breed solutions for each. The sophisticated analytics platforms that many of them use can ingest raw revenue management data, combine it with other data sources, produce periodic and on-demand reports, highlight exceptions, and enable managers to drill down to investigate unusual situations. Some of the revenue management providers said that with larger hotel groups migrating toward more use of such platforms, they have been able to focus more of their product development resources on optimizing rates and inventory.
Some of the RMS providers said they were getting requests to extend the time horizon for rate setting because hotels are experiencing many more “cancel and rebook for next year” requests, which extend beyond the traditional one-year forecasting window. Others reported requests to change their pace comparison reporting capabilities to skip 2020 as a point of prior-year comparison, allowing the more-normal 2019 to be used instead. Hooks have also been added to allow hotels to better respond to hotel closures and reopenings within their competitive set; to adjust for changing government travel restrictions; and to differentiate between unsold rooms that are actually available for sale vs. ones in hotel wings or floors that have been shut down entirely.
A couple of companies noted that room-type pricing has become more important. I suspect this may be driven by the shift to a much heavier leisure component as business and group volumes have dropped. The specific room configuration (e.g. bedding) can matter much more to a family group, friends traveling together, or even a couple, than they will to a lone business traveler. It makes sense that a hotel that can guarantee a specific room type should be able to charge a premium over competitors that cannot. This seems like a capability that will have value post-COVID as well; I know that when I have traveled with family there were many times over the years when I would happily have paid a premium to get exactly what I wanted. Cendyn recently introduced pricing optimization at the room-type level, an extension of the more common room-category approach. It will be interesting to see metrics for its impact on revenue.
I asked several of the providers how their products had changed due to COVID and also how some of their preexisting capabilities had helped them to cope with the changed world. One common theme that played out in different ways was the reduced value of historical demand data and the greater need to rely on current market data. Some systems, such as IDeaS, adjusted their models specifically to reflect this, while others, like Cendyn, had existing algorithms that detected the changed situation and automatically adjusted. RoomPriceGenie, which focuses on boutique hotels, does not forecast demand explicitly at all, instead using competitors’ prices as a primary indicator of demand – and is thus able to “piggyback” on the demand forecasts produced by systems used by the hotel’s larger competitors. That approach can work reasonably well for a small hotel whose rooms amount to no more than rounding error in the total local market.
A common theme was that situations change much faster now; a good weekend weather forecast in a destination market may generate a lot of sudden demand just one or two days out, or government travel restrictions may cause a sudden wave of cancellations. Cendyn and LodgIQ both reported moving to more frequent optimization in at least select situations, and Atomize felt that its real-time approach had proven very useful.
Several product capabilities have helped to address changes in how hotels organize the revenue management function. Cendyn noted that its recently launched RevIntel product had been useful in addressing the resulting changes in reporting requirements, as it enables each user to specify the frequency, timing, and content of their reports. A cluster revenue manager, for example, can easily schedule a high-level daily review of each of their properties and an in-depth analysis of individual properties on a rotating schedule. In contrast, a hotel general manager who needs to condense revenue management into a few minutes a day could schedule a summary report every morning and a detailed look weekly. Most mainstream business intelligence/analytics tools have similar capabilities, but many RMSs do not.
Atomize reported that its near-real-time on-demand reporting capability had proven popular with revenue managers who were responsible for multiple properties. Like many systems, it can produce a report for an entire brand, and then drill down to a region, a cluster of hotels, or a specific hotel. Unlike most, however, it has implemented an optimization algorithm that enables each new query to be answered in about two seconds, usable even on a mobile device. This could have a lot of appeal for a revenue manager who is suddenly handling three times as many hotels and needs to quickly identify and analyze anomalies.
LodgIQ recently introduced “Droid,” an artificial-intelligence driven bot designed to identify and highlight situations that merit management attention. This seems useful particularly during a period where everything has changed, as canned reports and exception reporting are usually based on historical behavior, and today’s anomalies may appear on dimensions that were never important before.
Perhaps one of the biggest areas of innovation has been the incorporation of more data in the revenue optimization process. Some of this was not specifically driven by COVID, but may have more value because of it. For example, Cendyn’s continuing integration of its extensive customer database with its Rainmaker revenue management platform is intended to enable personalized product offerings that should help drive more sales as travel recovers. Several vendors reported incorporation of competitive hotel closures, either manually or based on external sources, although I was left unimpressed with what seemed to be overly simplistic approaches or use of stale data. One vendor reported incorporating epidemiological data into their forecasts.
IDeaS noted that hotels were finding more value in their total revenue forecasting capabilities. Many costs are associated with food and beverage operations, and the rapid fluctuations in demand create greater need to forecast well so that staffing and purchasing can be optimized.
I was very intrigued by Atomize’s approach to utilizing real-time demand-pressure data that it gets from OTA Insight. This enables a highly localized view of online travel agency (OTA) searches for hotels for specific destinations and future dates, as well as for alternative lodging, event and flight data; they also get a dynamically generated competitive set based on which competitors are being shortlisted by the OTAs. Because it is a real-time query, Atomize is able to almost instantly identify any change in demand, and demand pressure is expressed as a “heat map” that is specific to an area the size of a few city blocks. This was announced last month and is expected to be commercially available soon.
Many RMSs acquire market data directly from providers via partnerships and/or support interfaces that allow hotels to import it from their preferred providers. LodgIQ does this but in response to COVID it also developed an approach to obtaining data to meet short-term situations where very specific data is needed, perhaps for just one market for a few days. Using scrapers, they collect relevant data from the web on demand, avoiding the need for expensive data subscriptions. Unlike subscriptions with periodic data feeds, the scrapers can request updates several times a day (for example in the runup to a high-demand weekend) and can therefore respond very quickly to changing demand patterns.
Not specifically COVID-related, and not really even revenue management, but another interesting company caught my eye at Cyber HITEC and it probably fits here as well as anywhere. Aiosell could well fill a need for a certain kind of hotel. It doesn’t meet my definition of an RMS, but nevertheless could be very useful for a hotel that for whatever reason isn’t able or ready to use one. It addresses the challenge of the hotel manager who needs to manage rates more effectively but in a minimum amount of time.
Developed by a multi-unit hotel operator for his own hotels but now sold to several hundred others as well, Aiosell effectively codifies all the “rules of thumb” for setting rates so that day-to-day attention isn’t needed (or at least, isn’t as important). It has an extensive set of factors that work together to determine each rate, including day-of-week, seasonal, and similar factors; rules for adjusting to the booking pace which can vary as you get closer to the arrival date; rules for calculating specific rates for room types; adjustments for market demand factors from external sources; competitive factors; minimum and maximum rates; and overbooking of lower room categories. Rates can be auto-published to a PMS or channel manager. The basic idea is that a hotel manager can enter parameters he or she is comfortable with, and largely leave the system on autopilot, looking at it once or twice a month and dealing with exceptions as reported.
Aiosell isn’t an RMS and would benefit from a few enhancements like automatic length of stay controls around high-demand periods, but it’s useful (especially for a smaller hotel with a seat-of-the-pants GM – we all know a few of those!) and has three inexpensive, transparent pricing options.