If you don’t know what hurdle points are, please see my video
A quality that I tend to notice in good revenue managers is a willingness to question the status quo, and actively experiment with the various tools at their exposure to try to exact extraordinary outcomes.
Back in the fall 2014 I attended a revenue management course hosted by one of the Major Big Brands. The course was mostly forgettable, but one part that did stick with me was the revenue management simulation that acted as a sort of ‘final exam’ after the four oddly-long 10-hour classroom sessions. For the simulation, all of the groups in the class (5-7 people each) were given control of their own hotel and were instructed to make a series of strategy changes for each simulated ‘season’. Each season had its own unique challenges: maybe a one-time group opportunity, or adverse macro-economic trend which required some analysis and decision making as to which course of action to take. The simulation was run for several rounds to simulate a few years of performance so we could begin to determine which hotels were making the best improvement in terms of Occupancy, ADR, and RevPAR growth, or in our class’s case: which hotels were left standing afterward.
I don’t remember the outcome of my team’s simulation (I think my team led for year one, but ended up dropping to 4th place by the end), but what I do remember was that the most fun was had by the teams that were actively trying to mess everything up. For example, one team outsmarted the algorithm and managed to sell a single room night for $1,000,000 which was a permanent ADR boost that kept them a perennial rate and revenue leader through the entire simulation. When I realized that more fun was to be had this way, I convinced my team to go toward the other extreme: non-stop 100% occupancy with a total disregard for profitability. We managed to get close by offering rooms pennies above cost-per-occupied room, much to the chagrin of the course leaders who were gradually watching a room full of adult children mutiny and run their simulated hotels into the ground using a myriad of apocalyptically bad strategy decisions.
The takeaway that I got from all of this was that even though the smartest guys in the room may have developed the algorithms that we use every day for demand forecasting, pricing optimizers, and doomed revenue management simulations, each of these automated algorithm-driven tools have their limitations and aren’t perfect. They should be questioned, monitored, and overridden where needed to ensure that the real expert, the revenue manager, has final say in strategy decisions.
Not to continue to beat up on the big brands, but one system that I notice which is very sensitive and should be questioned and monitored often is the Hurdle Point system used by certain price optimizers. Just as a disclaimer I do love hurdle points/last room values and the pricing model that uses them, I think that it gives length of stay pricing a run for its money in the ‘Most Clever Pricing Model’ competition, but system-generated hurdle points easily qualify as the most histrionic thing I have to deal with daily (you came in a close second place, sales people).
I started to notice that hurdle points could be dangerous after taking a few new branded properties onto our RM services last year; the initial red-flag for me was the fact that anytime I would start revenue management services with one of these properties, one of my first action items with each one of them was to machete a jungle of high hurdle points just to show reasonable rates and availability on Expedia and Booking.com.
For example, look at this Expedia rate shop of a property that I don’t manage. I see this kind of rate shop at my hurdle-point using competitors all the time: spotty availability and volatile rate strategies due to automated hurdle points.
This got me thinking. What if I start to track the price optimizer’s recommendations day to day? I thought that this experiment would take several weeks to begin spotting trends, but I actually found enough material to write this blog after just one over-night ‘optimization’.
Take for example the volatility of hurdle points after just one day of pickup at one of my properties shown below:
The above business on the books report perfectly illustrates a few things that I thought were very much representative of the overly-sensitive nature of hurdle points, and their counter intuitive and sometimes destructive nature in the short-term decision making window. Let’s start with making a few observations about the pickup. The ‘Pickup from Yesterday’ column shows rooms that were picked up for this property since February 8th, as I am writing this portion as of February 9th. Notice that there are a few strange decisions being made in the background:
First, the price optimizer has opted to slash our hurdle point by 51% for today after picking up 4 rooms. This luckily doesn’t affect OTA availability, as yesterday’s hurdle was only in the high 30’s, but seems rather volatile for a day-of change.
Second, and much more concerning, is the fact that the Hurdle point on Wednesday the 15th has skyrocketed by almost $31 dollars, and now restricts 1 night business from booking on Expedia even though the forecast algorithm only expects us to be at 71% occupancy!
Pause and let that sink in: the price optimizer doesn’t think you will sell out, but restricts your booking channels anyway. Why should a property be closed out on OTA’s for a night when it is at 40% occupancy and forecasted to hit 70%? As an owner, operator, GM, or RM this should make you very angry; even more annoying is the fact that Wednesday nights have the lowest negotiated margin with Expedia at 12.5% with this particular brand, meaning that the cost of turning away these lucrative OTA reservations is at their absolute highest.
Another night that is yet again forecasting a non-sell-out would be Monday night where only premium room types are selling on Expedia due to a high hurdle point. This isn’t as criminal as the example before, but again I’d like to see the algorithm start a bit more conservative and ease in to these hurdle points. We still have plenty of our most basic room type to sell and I’d like to remain competitive across all channels.
Adding an additional layer of complexity to all of this is the cumulative effect of advanced purchase discounts combines with OTA commissions, which make hotels with system generated hurdle points look very inconsistent in their availability and pricing strategy within the advanced purchase window. For example, if you have an advance purchase window of 9 days with a 12% discount, and a guest is shopping an arrival date at your hotel that meets that criteria where you are priced at $99, if your hurdle point is above $71.43 then you will not shop for 1 night availability on Expedia. This of course may be fine on nights where you are forecasting to sell out, but as we’ve seen above, price optimizers raise Hurdle Points at the drop of a hat, and I’m willing to bet that many, many prospective guests are being turned down by overly aggressive HP strategy at big brand properties everywhere. It’s no wonder that this property did an embarrassingly low 5% OTA market mix of total room nights sold with only 26% sellout efficiency.
Breaking down this market segment pie chart even further into stay pattern, the days of week where those OTA reservations did manage to limbo under hurdle points at my property in 2016 (before I took over RM duties) tended to be the weekends where this big brand’s negotiated margin with OTA’s is the highest. This is a bit strange when you think about it, especially considering the current climate of strong push back against OTA commissions by the big brands. I know many, many big brand properties in the midscale tract that are hurting for transient/leisure business who are now combating their own price optimization system to fight for OTA business.
Hurdle points are not an inherently bad thing, but they must be watched and controlled very tightly. Since many price-optimizer’s automated features are advertised as a revenue management panacea it makes me a bit worried that there may be a massive number of franchisee big brand owners and operators out there that simply don’t even know that this is an issue. If you have questions or concerns about your branded property, feel free to reach out!
– Daniel Foreman: Lead Machine Learning Guidance Guru at Kriya RevGEN