COVID-19 has brought an end to the economic and hotel industry cycle, and while recessions are nothing new to revenue managers, this is the first downturn they’ll have advanced pricing software equipped with AI and machine-learning functionality to set rates on the road to recovery.
REPORT FROM THE U.S.—Hotel revenue managers have their work cut out for them now during the COVID-19 environment and once travel demand resumes.
Veteran revenue managers have experienced downturns before and know the challenges that accompany them, but this time around, they will have artificial intelligence and machine-learning software to assist them in setting pricing strategies. Still, lower travel demand due to the coronavirus and historic hotel performance declines are a testing ground for this technology.
Sources said the lessons learned in past downturns will apply to revenue strategy in the current environment.
Tina Meredith, VP of revenue management at PM Hotel Group, said in an email interview that she worked in the hotel industry through the past two recessions and worked in revenue management during the Great Recession years.
“What strikes my memory the most about both instances, but more so in ’08 to ’09, was how quickly rate strategies changed, how quickly downsizing measures were put in place and how we, as hoteliers, lost the bargaining power with our clients,” Meredith said. “The leverage swung to the side of our customers and clients almost as quickly as a pendulum swings.”
Priya Chandnani, VP of revenue management at Benchmark, said in an email interview that she has “never seen demand trends as low as they are” today, but in past downturns hotel revenue managers endured and developed creative pricing strategies.
“In the last decade, we have developed a stronger focus on net profitability, driving revenue through direct channels and higher reliance on technology making the work process more efficient,” she said.
Vikram Pradhan, group VP of global revenue management for Wyndham Hotels & Resorts, lived through the demand declines of the Great Recession and even the SARS epidemic of the early 2000s. He added he’s not worried about the long-term health of the hotel industry.
“While SARS was localized in its impact on cities like Hong Kong and Toronto, it wiped out demand for hotels in those cities, similarly to what much of the industry experienced here in the U.S. during mid-March this year,” Pradhan said in an email interview. “The first time I experienced a downturn in demand, I remember worrying how the industry would ever recover. But time and again we have seen this industry persevere. I have confidence it will recover this time as well as we see different travel demand start to resume.”
Revenue-managing during COVID-19
Meredith said AI and machine-learning systems have certainly benefited the revenue-management discipline.
“They are definitely effective, but only as effective as the information we feed into them to help them ‘learn,’” she said. “That said, we have seen instances where if the price points are where they need to be, the reservation system is clean and coding is good, and it has the information it needs, the revenue system can make very good decisions with little influencing. So, ultimately, the revenue systems today—depending on which one you have—can be very effective as long as those things just mentioned are all in the best shape they can be.”
Revenue-management software evolves to suggest better pricing strategies with the more data it has, and sometimes poor data can disrupt those patterns.
“We do use machine learning in our revenue-management systems,” Chandnani said. “They are effective in managing rate and inventory as well as forecast models if monitored and influenced appropriately. Garbage in, garbage out—a system is only as smart as the data that feeds it.”
Chandnani added that COVID-19 has created data trends that are difficult for the revenue-management software to predict.
“Even the most sophisticated advance learning system must rely on predictable data and rules,” she said. “In these unprecedented times, revenue-management tools will be highly dependent on human intervention. System learning will be quick if we use the right tools to influence rather than override.”
COVID-19 has already had an impact on automated pricing tools, Meredith said.
“I would expect—and have seen—them perform less effectively in this downturn,” she said. “It has taken more interaction with the systems for sure. Additionally, there are many different scenarios that are affecting the systems from hotels being temporarily closed to having the whole hotel bought out to having 3/4 of the rooms put out of order in order to consolidate floors. These unprecedented situations have wreaked a little bit of havoc on our revenue systems.”
Pradhan said AI and machine-learning revenue-management systems should adapt with changing demand patterns but added the micro trends in performance now outweigh the historical data.
“The job of a good revenue-management system is, by definition, to spot trend variations before anyone else and suggest or implement strategy changes to maximize those conditions,” he said. “However, systems are rarely designed to anticipate once-in-a-lifetime shocks like COVID-19. With historical data completely unusable, systems may struggle to adapt to the new reality.
“For instance, it may take longer for the machine learning algorithms to pick up a drastic change in market conditions. Or they may overreact to a sudden loss of demand. This is why we need trained revenue managers looking at the system to monitor it and step in when necessary.”
In general, a company’s revenue strategy shouldn’t dissolve during a downturn. Pradhan said it’s unwise to cut rates too much to chase demand.
“When the hotel industry sees a demand shock like this, the inclination of many is to react quickly, whether that’s by dropping rates or changing distribution channels, to hold on to falling demand,” Pradhan said. “While some tactics work, overreacting can lead to long-term changes in customer behavior and rate loss, from which it can be very hard to recover. It can be quite difficult to recover once a brand has gone down the path of selling a room cheaply to a customer who would’ve bought anyway in hopes of attracting new customers with a low price. Sometimes you just need to ride it out, as the past month has shown us.
“However, it’s important to react decisively when demand starts to pick up again. As the industry starts to see demand return, it’s essential that your revenue strategy identifies those segments that are returning and prices accordingly to win their business.”
Meredith also warned against “panic pricing” during a downturn and suggested reevaluating revenue channels to find business.
“Another key strategy is to ensure that your hotels are taking advantage of every revenue channel possible—no reasonable channel should be restricted or shut down,” she said. “If we’re doing our jobs in times of demand, we are setting restrictions appropriately to yield higher (average daily rates). When everything falls off the cliff, you must ensure via auditing your systems, that all channels are open (at appropriate rates) and not being restricted in any way.
“Lastly, be creative with your strategies and ensure they are tailored to the type of demand that is currently out there. One small example would be if you are a normal full-service, non-extended stay hotel, but there is now demand for longer-stay reservations or groups, build some extended-stay offers to attract that business where you might not have done that before.”
Chandnani said now is the time to do some housekeeping with systems and strategies to be in the best position once demand returns.
“As we all prepare for hotel reopenings, we must think differently on how to gain market share,” Chandnani said. “Pace to last year is relevant for tracking seasonal trends, but we need to be nimble and watch for changes to our demand trends. This is also a great time to audit our distribution channels. It is not glamourous but certainly effective as we prepare for an uptick in demand. Ensuring all our distribution channels are open, rate codes are mapped appropriately and property descriptions are updated—all of those are important but not urgent tasks we usually save for later.”