The Difficulty with Data
The entire world races to collect, analyze, and utilize data and hospitality is no different. From using weather trends to influence pricing and guest preferences to drive extremely personalized marketing the industry is ripe with information.
Here’s the thing though, data in the hospitality industry is extremely difficult to deal with. While OTAs use clicks, A/B testing, user profiles, and machine learning on massive data sets to drive bookings, that is not the data environment for most hotels and vacation rentals. To add gas to the fire, analysis is just plain tricky when it comes to hospitality.
Here we’ll touch on a few key reasons data is so difficult in the short-term rental industry and how someone trained in hospitality analytics can help you tackle those challenges head-on.
Seasonality
Seasonality is a simple concept that’s easy to understand. Markets perform better at sometimes than others. It may be the weather, a big event, or just the booking pattern of a certain demographic but there are times when rates and occupancy skyrocket.
This simple idea really complicates any sort of data analysis. Its relatively easy to seasonally adjust individual metrics like ADR, occupancy, or RevPAR to scorecard performance. Its relatively complicated to measure how ADR and occupancy interact to maximize RevPAR.
Due to seasonality, ADR and occupancy are extremely colinear. When one increases the other does too. If there’s a huge conference happening near by its likely that rates went up anticipating the added demand but it’s also highly possible that occupancy went up at the same time. If we looked at the rates and occupancy on a graph it would look extremely weird. We fundamentally know that you should sell more rooms at a lower price than a higher price. However, the graph would show as rates increase so does occupancy.
Technically speaking, there aren’t many great instrumental variables to sort out the endogeneity. Simply, its really hard to sort through the noise. Using seasonal adjustments and market performance isn’t enough to filter the noise out. A rising tide raises all ships and regardless of sophistication or execution the market compresses during high season and starkly falls during slow season. Seasonality makes data analysis difficult without advanced statistical treatments.
Complex Interactions
Much like seasonality, there are interactions that muddy the waters of the data lake in hospitality and very few of them come from independent variables. ADR make be the only truly independent variable we can consider. After all, no matter what we can always control our price.
But beyond that its much more difficult. Occupancy is influenced by ADR, but also by seasonal demand, how the booking curve was managed, cancellations, and marketing campaigns. Demographics and target segments alter booking patterns. Different pricing and marketing tactics are used simultaneously while competitors due the same. Meanwhile, seasonality is still lurking in the background.
With so much happening at once its hard to say what causes what exactly. With so many components influencing each other or moving in lock-step only expertly designed experiments can occasionally gain key insights. But even experiments are complicated.
Room nights is a perishable product, once the night is over it can never be sold again. Unlike other products that can be kept in inventory if they don’t sell, room nights are spoiled and gone forever making any risky experiment potentially costly. It places a high need to execute an experiment properly or to find a way to take advantage of natural experiments and big data.
Not Enough Data
Big data … it’s a buzzword that’s thrown around quite loosely and to the unfamiliar seems like the elixir to every business problem. The problem with short-term rentals is that there really isn’t a lot of data.
Every vacation rental only sells once per night. Every night only sells once per year at one price. It takes a long time to get multiple observations of similar scenarios to really begin to statistically dissect performance. Even if you have a large portfolio its unlikely that the units are identical. They probably have different pictures, review scores, numbers of reviews, booked at different times, or may be located in a different place. With so little data available its really hard to gain insights that are robust or insights that go beyond industry fundamentals.
Not the Right Data
Even if a property has a lot of data it often doesn’t have the right data. Most businesses have a record of what room night sold when and for how much. But that is only one small piece of the puzzle.
What was the price when it didn’t sell? Most people don’t record this information. Even if you use a pricing platform like Beyond Pricing or Price Labs they don’t record that information either. In fact, they don’t record what price they post to the consumer period. All that a manager can see is the base price and the rules they set around it. Analyzing only dates and prices that were actually booked is a huge survivor bias that limits data analysis.
Quick to Judge
Naturally, people are quick to jump to conclusions and create an explanative story to determine the cause of something. If a manager runs a promotion once and gets a booking they’re likely to assume that running promotions are effective. If they were sold out for Thanksgiving one year they’re likely to think they’ll always sell out for Thanksgiving. People grasp at anecdotes to explain the world around them (if you want to know more on this, check out “Thinking Fast and Slow” by Daniel Kahnemen).
Using the example from above, we can see how a lack of data and quick judgement can cause the wrong conclusions. If a listing typically sees a booking once a week but hasn’t sold anything in four days they may decide to lower the rate. If the very next day they book a night they’ll think it was due to the price decrease. However, its very possible that they would’ve booked that room at the higher price as well. After making an assumption on a very small sample set (four days), they took an action (dropped price) and observed a result (a booked night). However, that didn’t increase the rate at which they booked. They could be seeing low traffic to their listing and so they simply needed a few more people to see it.
Jumping to conclusions and making snap judgements in the absence of data is extremely dangerous. In the short-term rental industry, it often causes managers to be scared of loosing their perishable room night and dropping price prematurely thinking “some revenue is better than none” when really they could’ve gotten “more revenue instead of just some.”
Not Enough Scale for Skill
All of the things we discussed is not easily attainable. It takes some experienced or trained to not reach snap judgments on small sample sizes and to consider all of the complicating factors. It takes someone knowledgeable and skilled to carry out statistical treatments and direct analytic strategies. It takes with expertise to carry out.
These expertise aren’t cheap and they’re not easily attainable. They cost a lot of time, a lot of money, or both. Many short-term rental portfolios are quite small. They don’t generate enough revenue to throw around and convoluted projects or for hiring experts. They also see a lower ROI because they have a smaller data set and fewer properties at which to see a revenue increase. Its hard to address all of the difficulties and have the resources to do it. However, its often impossible to do on your own.
This is why RevPARTY is here to help. We have the experience and expertise needed to tackle these challenges and create a real revenue impact. We also have built tools, developed methodologies, and gathered insights that can more efficiently be leveraged by a portfolio without starting from square one. All of our clients have contributed to honing the perspective of our team and every engagement makes the next that much more successful. The complicated analysis, deep thought, and executional strategies have been tested time and time again so that one individual client doesn’t have to bear the burden to develop a tried and true strategy. RevPARTY makes data a little less difficult.
If data analytics consulting could help drive your business forward or you want to leverage RevPARTY’s industry partnerships and discounts, please reach out and we can develop a custom engagement that not only meets your needs but is a cost efficient value add to your organization.