For thousands of years, humans used the movement of the sun and other celestial bodies to get a general sense of the passage of time. This method of timekeeping was reliable and accessible, but it was very inefficient. As early economies became more competitive, humans built contraptions that counted time without relying on the sun. By allowing us to optimize our days, these devices forever changed the way we measure and control our world.
Today, a similar revolution is happening in the hotel industry as hoteliers with the highest profit expectations are beginning to follow the customer instead of the product. They are using advanced analytics to dump their dependence on the crude “Per Available Rooms” (PAR) averages and have begun the move to measure their world by predictive indicators based on “Per Available Customers” (PAC). This is the Big RM Reset. It is a seismic shift that is reframing the Hotel Revenue Manager’s world view from PAR to PAC. Here are the steps required to profit from this transformation.
Step 1. From Property to Purchases
The conventional PAR measurements were inherited from a time when data collection and data access was limited. Since a property’s number of rooms rarely changed, they became the simplest gauge by which to measure changes in the world. Thus, “after-the-fact” averages like room capture rates, RevPAR, and TRevPAR became the benchmarks for evaluating the success of “what” happened. Ironically, while PAR is great for investors looking at how one property compares to another, it is almost irrelevant for hotel managers trying to make the great decisions that create value for those same investors. In fact, for most managerial functions, PAR does more to obscure problems and opportunities than to reveal them, often allowing unprofitable trends to go undetected. As an example, while an outlet’s capture rate may have remained unchanged over several years, the mix of guests actually making purchases may have completely shifted from a higher value to a lower value guest – a clear indicator that the product or service of that outlet needs attention. To truly understand your hotel’s unique demand rhythm, you have to listen to data at the most granular levels. Diving down to individual purchases is the first step in exposing the important patterns in the guests’ actions that truly drive or destroy profit. This is the spark that ignites all successful Total Revenue Management initiatives.
Step 2: From Myths to Measures
PAR is easy to calculate, but its simplicity often forces managers to make assumptions about the real drivers of profits. In time, these assumptions become powerful myths which then become the basis by which many critical business decisions are made. PAC, on the other hand, cuts through biases by illuminating what matters most – the customer. For this reason, it is critical to establish a set of PAC measures that will be used to establish “guest-centric” goals, objectives, budgets, bonuses and other performance indicators. From RM to Marketing to Human Resources, every department will now be evaluated by one version of the truth. For example, instead of just looking at the traditional PAR capture rate to evaluate an outlet’s performance, you can instead track the number and type of guests that actually used an outlet. Empowered with this PAC intelligence, managers can then identify the true sources of good and bad trends, move quickly to take immediate action, and track the impact of their decisions.
Step 3: From Flow to Footsteps
Flow through analysis, or how much profit changes with changes in revenue, is a time-honored tool in the PAR world. Similar to PAR, however, Flow is irrelevant when trying to identify and evaluate the variables that actually create additional profit. To gain the clarity of the PAC world, you need to trace how profit is affected by every guest choice. Footsteps analysis is the process of mapping the series of choices and actions made by each guest and the impact they produced. Identifying each guest’s unique booking pattern, marketing response, outlet choices, and even their food preference will reveal which of your hotel’s touchpoints is delivering real value, which need tweaking, and which are unnecessary, thereby allowing you to enhance the interactions that drive loyalty and profit and eliminate those that are obsolete. Here is where untapped opportunities begin to reveal themselves. Naturally, the sheer scale and complexity of following each guest action is beyond the ability of any human. Luckily, the mathematics of Decision Rules, Network Analysis, and Market Basket Analysis can be employed to monitor the significant choices made by each of your guests and expose the areas where making changes would yield the most immediate and long term impact.
Step 4: From Classes to Clusters
The traditional classes used for guest segmentation in the hotel industry are centered around billing instead of behavior. This is because, in a PAR world, RM only needs to focus on the mechanics of booking instead of what motivates it. Hence the term “heads in beds”. For instance, an “OTA” label says a lot about how the guest booked but nothing about how much they are willing to spend. However, a “Relaxation Weekender” tells you a lot about both. The insights gained from Step 3 are used to scrap the traditional hotel classes for smarter Behavioral Clusters. PAC is then measured at the behavioral cluster level. Here is where you begin to see constellations where you had only seen stars. Your new guest segments will open up limitless options for designing rates, packages, and incentives that drive single stay and lifetime value, forever redefining your approach to RM and Marketing.
Step 5: From Sorting to Scoring
Until now, Marketing has been limited to working with segments derived from using the basic methodology of traditional demographic analysis. This has forced hotels to confine their CRM tactics to simple acts of creating mailing lists by sorting guest records. Usually these “sorts” are based on arbitrary filters such as location, total spend, or even number of stays. The guest clusters created in Step 4 will give Marketing invaluable insight into the true mechanics and motivations behind conversions, loyalty, and lifetime spend. With behavioral clusters, Marketing can design campaigns that carry more relevant branding messages and promotional offers. Then, they can target the repeat and potential guests who are most inclined to commit to your brand. PAC-focused Marketing simultaneously lowers costs while increasing engagement of the types of guests that actually create profit. Mathematical Guest Scoring helps you identify the guests that have the highest likelihood of responding to specific campaigns, allowing your hotel to reach new levels of customer conversion and retention.
Step 6: From Rates to Relationships
Total Revenue Management begins when the power of RM is unleashed on all revenue streams. Now, your new guest clusters will allow RM to look beyond room revenue for profit and growth. By designing rate strategies that optimize total PAC performance, RM can simultaneously drive the success of F&B, Spa, Golf, and all other ancillaries. Best of all, RM will now be able to tie the impact of yielding decisions across all revenue streams. This improves the quality of short and long-term forecasting and therefore leads to a more predictive basis for operational and financial planning. In addition, RM can signal to Marketing when and where incentives are needed in order to improve PAC results. By shining the spotlight on the customer instead of the rooms, RM can now concentrate on the all the variables that help create meaningful relationships with valued guests.
Step 7: From Silos to Symphonies
In the PAR world, Marketing and Revenue Management are typically disconnected, pursuing objectives that are often at odds. While the first is focused on the motivations for booking, the latter is single-handedly trying to manipulate the mechanics of booking. This “silos” behavior implodes in a PAC world as RM and Marketing become focused on the same target. For this reason, this step is not a step at all, but rather the natural synchronicity that emerges between RM and the rest of the hotel. This is where the impact of the PAC transformation is actualized. As all departments begin to realize that the guest sees them as one interconnected package of perceptions, managers begin to coordinate their activities around optimizing their impact on PAC. That is the synergy that the best hotel companies in the world are chasing.
Step 8. From RM Staff to RM Star
The shift from PAR to PAC requires that hotels seek talent that understands the math that maximizes results in forecasting and decision making. This transformation has constrained the vast majority of hotel RM professionals, who do not have a mathematics background, to a position where they can not offer the most value to their property. Subsequently, this talent gap will only increase in the years to come as technology zooms ahead into Big Data platforms — all, while hospitality schools are only now, slowly, beginning to modernize their RM curriculum. In order to quickly capitalize on the Big RM Reset, it is critical to use Hotel Analytics experts to build mathematical models that will allow your RM Staff to successfully navigate the transformation to PAC. To this end, many top hotel companies have centralized RM and put rate decisions in the hands of mathematicians and technology experts.
The Big RM Reset is well underway in many forward thinking hotel companies. Embrace this new world view and you will forever change the way you see and do business.
Posted by Robert Hernandez
in Decision Science,
Hotel Analytics,
Strategy on July 11, 2012
In the absence of mathematical analysis, decisions done under uncertainty become highly subjective activities. The most relevant example in hotel RM is the process of selecting the rate that has the best chance of optimizing revenue for any given night. In the vast majority of properties, this task remains a highly judgemental exercise. Even in properties that claim to use a “scientific” approach, the analysis applied is mostly algebraic or dependent on some “rule-of-thumb” set, which means that the effects of variance or “chance” are not being properly quantified to make a truly informed decision.
This is not to say that RM professionals without a math background are not smart enough or experienced enough to guesstimate the best rate. However, they must be keenly aware that any rate selection done without a thorough mathematical interpretation of what is actually happening is likely to be subject to the unintentional application of some very human, but very dangerous biases. I thought it would be useful to highlight my list of the 10 most common biases applied in rate setting.
1. Confirmation bias (i.e. wishful thinking) is the tendency of people to favor information that confirms their beliefs. For example, we focus on spikes in the booking pattern that seem to verify our guesstimate of where the final occupancy will end up or how the month will close. These spikes may, in fact, be statistically insignificant.
2. Ambiguity bias is the tendency to avoid options for which we have little experience. This is often seen in the reluctance of some hotels to charge a rate that is outside of a defined “comfort” range. For some, this range may be defined by always staying below a certain competitor or within a percentage of the rates charged the previous year.
3. Bandwagon effect is the tendency to do or believe things just because others do. When everyone in the strategy meeting believes that “this will be a tough summer”, your rates will inevitably go low and stay low.
4. Anchoring bias is the tendency to rely too heavily, or “anchor,” on one piece of information. For example, you may perceive all “event” dates as strong demand days and should therefore be priced differently from all non-event dates when, in fact, some non-event dates may show similar demand patterns and can be priced just as aggressively as event dates. An additional hazard of this bias is that you run the risk of overestimating the performance of event dates.
5. Overconfidence is the excessive faith in one’s own abilities to make judgements. Enough said.
6. Semmelweis Reflex is the act of automatically rejecting facts without thought or real consideration. Often managers “stick to their guns” because they have of a gut feeling about the outcome when all the evidence points to the contrary. This is especially true among senior managers who want others to respect their natural intuition.
7. Illusion of Control is the tendency to underestimate the magnitude of uncertainty because we believe that we have more control over events than we actually do. This is experienced by most in RM when they make aggressive rate changes but nothing significant materializes in occupancy.
8. Averages bias is the over reliance on averages as a metric because we feel it capturea the “essence” of what is going on. In most cases you should be taking into consideration the significant changes in variations and deviations. A common example is practice of rate setting by adjusting rates by a percentage of the average comp set prices. Remember, if the lowest rate in the comp set drops by $50 and the highest increases by $50, the average is still the same , but the market dynamics have completely changed.
9. Illusory correlation bias is when we perceive a close relationship among two variable that are in fact uncorrelated or only loosely correlated. A classic example is the myth that higher total occupancy drives profit and outlet spend. Today, most sophisticated hotels have realized that outlet spend is driven by attracting the right customer, not just by increasing total occupancy. Furthermore, the “right customer” base may, in fact, be only a portion of the total customer population.
10. Recency bias is the tendency to focus on the most recent data and regard older data as irrelevant. For example, most Pace reports are based on comparisons to previous year information when, in some instances, it may make more sense to compare the current year to a peak performance year, a three year average, or a similar performing year. In fact, if you are consistently ahead on Pace, you may be high fiving each other while simultaneously undermining your performance standards.
This is not meant to be a comprehensive list, but more a review of the biases that I have encountered when working with clients. Keep in mind that I did not count the frequency of each bias as I witnessed them, therefore the creation of the above list was subject to my own biases.
There are thousands of sources on the internet to help you study decision biases. You can start with Wikipedia’s comprehensive list at http://en.wikipedia.org/wiki/List_of_cognitive_biases. I would love to see your top 10 list.
Posted by Robert Hernandez
in Analytics,
Marketing Analytics on April 2, 2012
Welcome to our blog! Today is a great day. It’s great because I unexpectedly got a free cup of coffee from a multinational corporation and, at least during this transaction, I gave them nothing in return, except maybe, some discouraging data.
Here’s the story of my $2 coup in a cup:
My father and I decided to take my two young boys to the park to drain some of their energy (if that’s possible). On the way, I asked my father to stop at Starbucks for coffee. Without hesitation, he went into his glove box and pulled out a coupon that he received in the mail for a free McDonald’s coffee – any size, and there happened to be a McD’s at the next light whereas a Starbucks trip would have taken us away from the park. I scanned the coupon to find out what I would have to buy to get the free coffee and to my surprise it did not have any restrictions. When I alerted my father to this fact he was also very confused.
We pulled up to the drive thru speaker where I proudly announced that we wanted to exercise our moral obligation to redeem this “gift” which we are legally entitled to obtain. After successfully fielding a series of questions that were aimed at discovering whether I was a thief or stupid, such as ”Are you suuuure it does not ask you to buy another coffee at full price?”, the attendant was so perplexed that he asked us to drive up to the window. He took the coupon, called his manager, entered a few dozen override codes into the register, and then he asked me how many creams and sugar I wanted. It seemed too easy. Was this really happening? I told him that I would like three creams with my large coffee, no sugar. The manager predictably told me that I would get a regular coffee. I reminded them that the coupon said “any size”. They agreed, gave me my coffee, and I rode into the sunlight.
Now my analytics hat went on and I tried to make sense of what just happened. I discarded the notion that this coupon could have been part of an actual promotion to drive sales. I mean, McDonald’s had to know that the bearer of this “gift” was most likely going to take the one coffee and run. Then again, it’s been well documented how fast food franchisors often launch discount programs aimed at increasing traffic at the the expense of the franchisees’ profit. But by now we have to assume that McDonald’s has enough data to successfully separate the marketing actions that work from the fruitless.
Let’s retrace our steps. The intended audience, I think, was my father, who never eats McDonald’s or drinks coffee. The coupon was part of a flyer full of coupons with buy-one-get-one offers. But I redeemed the coupon – a morning coffee drinker who never drinks McDonald’s coffee. I bought nothing else. We made the purchase at a store about 10 miles from his house. What did McDonald’s get?
Of all the myths ever perpetuated in B2C marketing, the most dangerous is the belief that giveaways create customers. While this tactic may work for the town’s new baker who is trying to earn the trust of the local villagers, in today’s non-captive and saturated marketing environment, giveaways achieve the opposite of their intended effect. In trying to promote an “incentive to try” via a “foot-in-the-door” tactic, giveaways actually open up an unintentional “floodgate to pillage”. The result is that you WILL attract new users, but you will not convert enough of them into profitable customers.
To be clear, by giveaways, I am not referring to limited demos, betas, stripped down samples or freemiums – standard versions of premium products. Those are all inferior versions of the actual product. My cup of coffee was the same as the one that the customer behind me paid for.
For the past 50 years, psychologists have studied the effects of promotions and giveaways on purchase behavior. In a 1977 paper, Shoemaker and Shoaf found that the probability of repurchasing a brand actually dropped if the previous purchase was on promotion. Since then, numerous papers explained this phenomena using self-perception theory which postulates that you are most likely to value a product by the promotion that you received rather than by the benefit the product gives you. This cognitive condition makes it unlikely for you to ever pay full price for that product in the future. In other words, once a person has received something on discount or for free they will subconsciously ask themselves, “Am I the type of person that would pay for something that I just enjoyed for free?”. For most, the answer is no, and Groupon is now feeling this well documented fact.
Examples of our reluctance to go from free to paid are everywhere. Most recently we had the Netflix service split, rate hike fiasco. But most notable is the strategic minefield crossed by Google and Facebook in trying to monetize their platforms. In the end, both had to conform to the “ad impressions” model wherein their role is to engage potential customers while sellers try to bate them to spend. Unfortunately, once you go free, you can’t go back.
More importantly, the McDonald’s coupon was ill conceived in another category – they did not get any useful data. Had the coupon stipulated that to redeem it, I had to fill out a short (3-5 questions), well-crafted behavioral survey, then I could definitely understand why they were willing to give away the immaterial cost of a cup of coffee. I could have easily completed the survey at the drive-thru window or I could have activated the coupon on my phone via a survey app that required my facebook login. In the analytics age, all business activities should be measured by either a direct revenue minus cost figure or by its indirect “data capture” value. By focusing on the potential sale instead of the potential data, this company missed an opportunity to gain some insight into discovering the process that could potentially make its competitors’ customers switch to their brand.
Still, some of you may be thinking that since I have tried McDonald’s coffee, I would be more likely to buy one if a Starbuck’s was not within reach (foot-in-door strategy). It’s true, the fast food business thrives on accessability. However I am a loyal subject of the “experience” culture. I go out of my way for Starbucks and pay a premium, not because their product is better (I’ve had plenty of bad ventis), but because they make coffee drinking sexy. Marketers must use behavioral data in order to figure out new ways to affect the motivations behind brand loyalty instead of blowing their money on trying to manipulate the mechanics of the purchasing process.
Therefore, my free coffee was a missed opportunity, not only to impact immediate sales, but also to capture valuable data that could have impacted future sales. In this regard, this giveaway was poorly brewed.
