Data Science Strategies for Short-Term Rental Revenue Management with Chris Garcia
Hello everyone and welcome back to the tech and real estate podcast. My name is Ariel Herrera, your fellow data scientists and I'm your host today. For this episode, we have Christopher Garcia with us. Chris is a professor and the founder of Polyak Decision Sciences. From our conversation, you will learn about revenue management, including price optimization with use cases, like how hotels, and short term rentals like Airbnb, dynamically change prices based on demand.
Ariel Herrera 0:00
Hello everyone and welcome back to the tech and real estate podcast. My name is Ariel Herrera, your fellow data scientists and I'm your host today. For this episode, we have Christopher Garcia with us. Chris is a professor and the founder of Polyak Decision Sciences. From our conversation, you will learn about revenue management, including price optimization with use cases, like how hotels, and short term rentals like Airbnb, dynamically change prices based on demand. Chris is a professor at the University of Mary Washington and quantitative methods. He has extensive experience in data science, cloud computing, and big data. Chris recently founded a company called Polya decision scientists that develops AI solutions for decision making. Let's welcome Chris to the show. Today, we have a special guest, Chris Garcia. Welcome to the channel, Chris.
Unknown Speaker 0:57
Great to be here. Thanks for having me.
Ariel Herrera 0:59
Yeah, super excited, not only to hear from someone who has a lot of experience, but particularly in the tech space. And I'd like to just maximize the time that we have and let the audience hear a little bit about your background, and how you've ended up on your career and developing your company today.
Unknown Speaker 1:18
Yeah, yeah, just to tell a little bit about myself, I am a professor, a decision sciences professor here at the University of Mary Washington in Fredericksburg, Virginia. And I've been doing this for about 12 years. So the areas I do a lot of work in are really decision sciences. So it's data science, on the one hand, a lot of predictive modeling, you know, and then, of course, the application of different types of cognitive technologies, natural language processing, computer vision. And then a lot of the other work that I do related is in the area of operations research. So this is really the mathematical modeling of large complex systems, and really the optimization of them, you find them in things like logistics and supply chain, you know, designing supply chain networks, or, you know, large scale vehicle routing algorithms, these kinds of things. So I kind of find a confluence between a lot of the work I do is really the confluence between those two areas, both in my research and my teaching. I also have a small consulting company, it's a relatively young consulting company that was started recently, out of a university consulting organization that I was part of, and my company is called Polyak Decision Sciences. And what we do is we focus on really bringing advanced analytics and optimization and artificial intelligence to support decision making and automation to a number of different industries. So some industries that we work in transportation, logistics, and supply chain, insurance, defense, are some of the ones that we are working on right now. So yeah, those two areas, those are the two things that are really, you know, kind of split my time between those two, between teaching and research on the one hand, and building a consulting practice on the other.
Ariel Herrera 2:52
That's excellent staying busy, you always know that you wanted to get into the space of teaching as well as in technology.
Unknown Speaker 3:01
Yeah, you know, I didn't really intend to become a teacher, per se. But when I was in my Ph D program, I had the opportunity to to teach, I loved it, it was just wonderful, you know, I realized, you know, this is really a fulfilling thing to do. And having done this, now, I'm in my 12th year, at the University of Mary Washington, it's really, it's very, it's wonderful to see students. You know, I remember some that were here when I was starting, and now they're doing just amazing things, you know, and they come back to me, tell me what they're what they're working on, and all the amazing things and you know, that you see, they got their start through this. And so, you know, I find it very rewarding, you know, the other side of that also is being able to do research and to work on problems in my discipline, and advanced the field, you know, and, you know, working on various problems, these kinds of things. And, of course, you know, you developed your expertise in different areas, as well. So I find those areas, yeah, really a great fit for me, in addition to my consulting work,
Ariel Herrera 3:55
that's great to hear, and diving into one of those areas. So I know, you've had experience in revenue management. That's one of the parts that is in your website, or the consulting that you provide, if you could just give an overview for those who don't know what revenue management is, what is it? And what are some of the problems that companies may face?
Unknown Speaker 4:15
Yeah, that's an interesting question. So revenue management is we actually don't hear a whole lot about it, you know, in the popular media, it's really kind of a discipline that is in the background, you know, but it's also a very important discipline for a lot of different industries. So, you know, the way we interact with revenue management, maybe, you know, you've had experience booking a flight or booking a hotel room at some point, right. So one of the things that you might know is, you know, if you book a flight six months in advance, you tend to pay a very low fare a few $100. You know, it can get you you know, up and down the East Coast. By contrast, you know, if you need a flight the next day, you might pay a few $1,000 for your fare, you know, for you know, so on a flight, you know, one person seat might have cost him $200 You know, another person seat might have cost them $2,000 right next to him. And so why the difference and that's that's kind of what revenue management does is it really applies advanced analytics and data science to help organizations to maximize the revenues that they can bring with their existing capacity. So the idea here is to increase revenues without increasing costs through smart, applying smart technologies, like, you know, predictive analytics, and optimization, really, it's about predicting real time customer demand at the micro recommend micro economic level, and then doing things like optimizing the price and availability of products or services.
Ariel Herrera 5:34
Not it's a great explanation. I already mentioned the flight example, it reminded me that recently, fly friends here, they have a new program where you pay flat rates, I think it's like 600 $700. And then you could book the day before, in order to take a flight, which I assume they have, they must have some model in the background to understand that these flights would likely have no seats or one day before an order to maximize what's available on the plane, providing a bundle package like this could help. Do you have any insight as to how some kind of program like that would work?
Unknown Speaker 6:11
Yeah, that's a, that's a great illustration, actually, and I'm sure you know, I'm not familiar with that particular program, but I'm sure that they're incorporating, you know, advanced revenue management practices. So really revenue management, you know, it has its roots back in the 1980s, really where it got start was in the airline industry, it was called yield management back then. And what happened was, in the early 1980s, a small airline called the people express came up, which is offering really low fare, low cost fares, you know, it was really a big threat to some of the big airlines, you know, they were actually competing on their core routes, and taking business away. And so American Airlines was really one of the pioneers in revenue management, what they did was they developed, they realized that there are two different types of fliers, among others, you know, there's the leisure fliers who'd like to book far in advance, and they'd like to, you know, they're very price sensitive that group and then there's the business fliers who you know, need to put, you know, oftentimes much later, and they're a lot less price sensitive than the leisure fliers. And so you can, you know, really, if you think about those are two, even though they're the same product, you know, physical product, you know, to the to different groups, they mean two different things, or their value is two different things to do two different groups here. So they're really offering two different serving two different customer segments like that with the same capacity. And so what they did is the techniques that they started doing was to forecast Okay, on each flight, or each flight leg, you know, what are, you know, what's the expected number of business fliers, you know, that are going to come within three days, you know, book within the three days before the flight departs, you know, versus those leisure fliers, how many of those going to be, so what they would do is they would realize that, you know, they got their forecasting models accurate. And then they would, you know, use optimization algorithms to determine how many seats do you reserve you hold ahead of time, and just open up at the last minute for those business classifiers. And you can see how that actually does serve a population is, as you mentioned, you know, if you open those seats up all of them early, you know, there's a big risk that all the seats fill up, and there's nothing available for the business class fliers who don't know that they need the seat until, you know, a couple days before and then all of a sudden, they can't have their flight. So, you know, by differentiating prices, you know, based on the different customer segments like this and and using optimization algorithms, American was able to restore the profitability, in fact, it was so successful, it really it actually wound up resulting in the closure of people express and as a result, you know, a lot of other industries over time started to adopt the same principles. One of the places is another place that we see it commonly today's and hospitality, you know, and hotel, the hotel business. Car Rentals is yet another where we see it, you know, even things like golf courses are starting to use it to schedule tee times and to charge accordingly.
Ariel Herrera 8:47
Oh, wow. Oh, boyfriend about that one, since he likes golf. But segwaying back to what you just mentioned, hotels as well. And since a lot of us, including myself in the audience, were either investors or interested in real estate in some aspect. How does price optimization have play into a factor for hotels and possibly for areas like Airbnb ease?
Unknown Speaker 9:13
Yeah, that's a great question. Um, you know, applying revenue management, you know, there's really, it's applied, there's, there's four key pieces to revenue management, the first thing is, you know, is involves customer segmentation, we have to really know the different customer segments, you know, and the value that we get what we give to them, you know, how does the value differ based on the different customer segments. So, as I mentioned, the airlines, you know, two seats on the plane mean different things to different people, you know, and so understanding that value, the, you know, what we have to offer means to the different customer segments, is really, really important. Another thing, another aspect here is the, determining the price response. So, you know, as we increase prices, you know, there's a connection between the demand and the price and generally as the price increases, we're going to start to see that the demand decrease. Now, in some cases, you know, if you imagine a curve where price is on the x axis and demand is on the y axis, you know, you can, you know, it's gonna be the demand is gonna be highest at the price of zero. And then as you increase the demand, you know, typically the increase the price, the demand is going to start to drop, you'll find these points, you know, where it's relatively flat, as you increase the prices or places where, you know, there's not a lot of price sensitivity, and you can increase the prices. So part of that is, you know, for each each segment, you have to understand, you know, how does each segment respond to price changes, you know, and obviously, you know, if you price too high, you know, you're gonna price people in that segment out of the market, you know, and, of course, you're gonna lower your sales, you're gonna hurt your revenues to that, if you price too low for a given segment, you have no problem with sales. But of course, you know, you're leaving some revenues on the table. So it's really, it's a science about, you know, pricing, you know, at the right price for the right people, you know, the right product at the right time, you know, through the right channel, this is really what it's about. Another Another factor is, you know, using demand forecasting, because what we have to do, you know, we're gonna price and we're going to sell, you know, we have to know, you know, what the expected demand for the different products are, you know, and again, you know, a plane seats on a plane, there's demand for by leisure fliers, and there's demand by, you know, the business flyers. And so knowing those, you know, each groups demands on a given flight is important, because we have to know how much to reserve for each group accordingly. And then, you know, there's the other piece of that is, you know, the last piece is really capacity allocation. So we have to use optimization to figure out, you know, obviously, you know, if you imagine hotel rooms, you know, one room can be sold for, you know, a three night stay, or a five night stay, you know, so depending on that, you know, that, you know, if it's sold for three nights stayed in those last few nights can be used for another customer, but it's four or five nights stay, they can't. So you have all these conflicts between how you allocate your capacity. So you want to figure out how to do it intelligently. So that you're, you're making the best use of your available capacity or so, you know, those are the kind of the four main legs of revenue management. And as we apply those into, you know, a new situation, you know, what revenue management looks like, in the airline industry looks quite different from what it looks like in the hospitality industry. You know, that being said, I think, you know, in the real estate industry, you know, there's a lot of potential uses for it, you know, one that we see is and certainly landlords are using this company called Real page builds some revenue software, revenue management software for landlords, I think another place where we can really start to see this to, you know, I don't know if there's a lot in this space yet, it's actually relatively new area, is on short term rentals. You know, this is a growing area where real estate investors are starting to work in they're finding, there's, you know, this is meeting a need of a lot of people out there to rent for, you know, not a long term, but a relatively short term. And I think, you know, revenue management techniques can be adapted, and that sphere was probably in a very promising way, I think we'll start to see that in the future.
Ariel Herrera 12:51
Agreed. And that's one point that a lot investors are one riches in the niches so they want to see what segments can they target for specific, I guess, areas in real estate. So when it comes to short term, medium term rentals, a lot of investors have targeted say, like, nurses that are working different hospitals, and need to stay for maybe three months or six months stay, as an investor, say, was looking at a different area? How would they first start to break out segments? Is there a data driven approach that they could use to understand who is in a certain area may want to use their product? And how often? And how immediate?
Unknown Speaker 13:35
They certainly, yeah, certainly, you want to use a data driven approach, you know, that being said, you know, that segmentation process, you know, you have to get to that from a few different angles. So part of that is just using some intuition and knowledge about the customer base. Now, you know, the best thing that, you know, if you're gonna take a data driven approach, you want to try to collect as much data on the people that are inquiring, and the people that are actually renting your properties, as many as as much as you can, because then that starts to allow you to build different profiles. And you can start to see characteristics, you know, maybe demographic characteristics, you mentioned careers, like a nurse, you know, and see how those different characteristics you know, correlate with, you know, different usage patterns, you know, different lengths of stay that are typical, and different wants and needs, you know, that you find, and once you understand those, you know, that gives you the ability to start differentiating your offerings, you know, among the different segments.
Ariel Herrera 14:24
That's useful. And for the investor that may say, what is correlation? How do I observe something like this about knowing how to say program? Can you give me like a basic overview of what correlation is and how it applies?
Unknown Speaker 14:39
Sure. So, correlation is basically, it's a measure of how strongly two variables relate to each other. So, you know, we can have positive correlation. So correlation ranges from between negative one and positive one. If you have if it's positive one, that means that the two variables are perfectly gray. So for instance, you know, you know, height and weight would be examples of variables that are correlated, not for graphically, but generally, if somebody gets taller their weights gonna go up, you know, would be one, you know, another one a well known one is, you know, an inverse correlation is where as one variable increases, the other variable decreases. So, you know, happiness and debt, you know, those are inversely correlated. And generally, as your debt goes up, you know, your degree of happiness goes down here. So, we look at correlations in the data, you know, as I mentioned, you know, looking at nurses, for instance, you know, we might say, okay, you know, three months stays, you know, there's a very strong pattern here that people who stay, you know, three to four months, in a four term rental are nurses here. So that'd be an instance of a correlation between a profession and the length of stay. And so, you know, using data driven approaches, you know, for that segmentation piece, it's very advisable, you start to build profiles, you start to see, okay, notice patterns, okay? When you have these characteristics, it generally translates for, you know, to a property of this size, you know, or price range of this, you know, for a length of stay like that, you know, and this is how you start to determine your customer base and start to differentiate your different customer segments.
Ariel Herrera 16:02
Excellent. And then, in terms of some of the datasets that you worked with, when looking at price optimization problems, could you briefly explain, like how to use mostly public data sources, private, How's that look,
Unknown Speaker 16:17
I mean, it varies really, you know, when we get into price optimization itself, so that's really one component of revenue management, you know, so sometimes, you know, the pricing is set, you know, ahead of the time period, and in the old days, you know, airline revenue management early on, they would set the prices once before the sale period, and then, you know, it would be fixed for the different fare classes, and then, you know, there was about capacity allocation optimization. You know, when we get into more modern revenue management systems, one of the things that you see is very dynamic pricing, and the prices are constantly changing, as markets change, its market conditions are changing. So it's more reactive, and, you know, it's more of a feedback cycle, than it is just, you know, the step and then this step, and then the step. You know, when we talk about datasets, you know, there's a lot of things you have to take into account to, to really do pricing? Well, one of the one of the main ways that we get data for pricing is through pricing experience. So you actually have to try this out in real life, you know, in the real market conditions to see, you know, how are the customer segments gonna respond to different prices. So, you know, the classic A B testing that we see, you know, in web based data science, you know, to figure out, you know, what's the, how do we design our website in a way that's going to make it, you know, most usable and easy for customers lead to conversions, we can take that same approach, you know, and apply it to pricing experiments, you know, where we try some customers at one price point within a segment, and then change it up or lower, start to see how that impacts demand. And that will allow us to begin to estimate our demand curve, you know, between price and demand, what does that relationship here, but really, there's a lot of other things that we need to take into account as well. You know, one of the things, you know, is industry data, you know, if there's a shortage in the industry, obviously, you know, shortages, you know, lower supply is going to be linked to higher prices, you know, higher willingness to pay for the, for the goods or services here. So, that's another thing, you know, using that kind of economic data based on industry, another very important piece of the puzzle is to incorporate competitor data. So if you really understand your competitors, you really need to bring that into your pricing model somehow, you know, either explicitly through the, you know, knowing what your competitors prices are, and how they're changing, or, you know, by doing your pricing experiments, you know, in real market conditions, where the customers are exposed to the competitors, they know, the competitors, what they're offering, and they're reacting to you based on, you know, incorporating their interaction with the competitors, taking that into account here. So those are a lot of different datasets really, you know, again, there's no one way to do the price optimization, you have to tailor it, really, to the different circumstances here. But you know, that the main sources of data are, you know, through pricing experiments, you know, bringing in economic data, you know, especially, you know, things related to supply and demand, information, you know, about your customer segments, you know, characteristics, and, of course, competitive data, all those really kind of come together to help you determine those best price points.
Ariel Herrera 19:00
Excellent, very good synopsis and then go into what you said, like the macro economics and such. So, we've been seeing a shift because of interest rates rising, some Americans don't have as much buying power as they used to have, how do you incorporate for that in your models to count for those types of segments that maybe were purchasing more, you know, post COVID, they were trying to spend a lot of their funds, but now kind of reeling back in on those leisure activities, how does that get incorporate into their model?
Unknown Speaker 19:32
Yeah, that's a that's a really important question, actually. And so, you know, this illustrates something also another important principle of all this, you know, especially, you know, in terms of the price estimation, and these kinds of things, you know, determining that price response has is changing, right. It's not a static thing. So we need to we need to periodically re update, you know, our demand curve estimates here. And so these pricing experiments have to be fairly ongoing. This incorporation of new data up to date data, you know, into our pricing models has to be really ongoing. You know, so as things change, you have to be able to respond to those here. And if you don't, obviously, you know, you have model drift, and, you know, basically your models no longer correspond to reality anymore. So one of the things that revenue management can help with, you know, that's a good use case you mentioned here, you know, it kind of illustrates the power of it, you know, so obviously, you know, you find organizations using revenue management to, you know, to help them deal with over demand, you know, so when you have more demand that you can really supply, you can increase your prices to get the demand to a level you can handle, you know, and that also helps you that's beneficial, because it helps you, you know, also increase your revenues. Another thing, though, you know, just as we can see that demand, and price are related. You know, as people, you know, price sensitivity becomes more poignant, and like, you know, increasing interest rates in these kinds of things, you know, and there's off peak hours, you know, off peak times in different industries, you know, I'm sure this happens in real estate as well. You know, maybe even in short term rentals, you know, in those times where the demand starts to go down, where there's generally less demand, one of the things, one of the ways you can actually use to stimulate demand is price by offering lower prices than typical, you know, maybe it's at a time, that's not the most convenient for somebody to take a vacation, but there's a, Hey, I couldn't take one at all, if you know, if things are as they continue, but hey, here's this deal, you know, it's in March, maybe it's not, you know, June at the beach, but it's this nice little vacation place in March, instead of really low cost, you know, by doing that, you can actually stimulate demand, and you can start to serve customers who, you know, have a more constrained budget, you know, who are much more price sensitive. So, you can actually, you know, this is another principle of revenue management is you can actually stimulate demand, you know, an altered demand by, you know, and supply by changing your price around. So you can move, you know, those, those ones from our higher segments, you know, from a more more in demand, you know, high usage segment, you know, or time period, you can have those start to shift their usage to, you know, times that are less peak, and we're lowering your prices and stimulating demand like that.
Ariel Herrera 22:00
Excellent. So say, if I'm a large investor, I have a whole team behind me, we have 1000, built to rent properties that we bought a couple of these large developments. And now we want a pricing model to help us in terms of short term rentals. So we're going to convert this into three months or six months rentals. We're seeing that because our area, maybe isn't one of the major cities that what's available to buy tools aren't really that accurate, or aren't helping us as much. So we want to come to you to give us a custom solution. And to help us in our price management, what would be some of the first steps that I would need to provide you? And what are some things that would like pique your interest to look into first?
Unknown Speaker 22:50
Yeah, that's a very interesting question. There's really a you know, as you start something like this, as I mentioned, you know, revenue management really is a framework more than anything else, it's not a strict rigid recipe to be applied the same way to every industry, you know, so, you know, how we apply it, you know, is really important to being able to ensure that we get, you know, the best return on our investment here, you know, so to begin, you know, you really, you have to really be able to understand the distinct customer segments, you know, the distinct usage patterns here, you know, so there's not a real differentiation, you know, there, you know, and again, you can collect data science techniques, you know, really help you when there's signal in the data, as opposed to just noise, if there's no underlying signal, you know, then it really can't be that much of help. So we have to start taking some data, we have to start looking at the customer usage patterns, you know, and start understanding, you know, what are the distinct segments here, you know, and what are their different price sensitivities, you know, so we kind of work these different pieces, we first, you know, we want to start off thinking about, you know, segmenting the market properly. And then we want to think about, you know, of course, you know, estimating the price responses here. And so typically, that would, that would look like, trying out some different pricing models, you know, trying out some price experiments, incorporating what data is available, obviously, as you mentioned, if there's less data available, you know, we have to find ways to start generating that, you know, and the way you can start generating that is by trying something out and recording what's happened, and then start incorporating that, and, you know, obviously, it takes a little bit of time to build up data, enough data that you can make good decisions on, so you have to do some experimenting for a time, if there's not a lot of pre existing data, you know, and then, you know, you know, once we understand, you know, the, you know, the segments, you know, the usage patterns, we can start to develop a reasonable forecast of usage, across the different segments, and across the different time periods, you know, we can start figuring out how to optimize and so what's something like, this would look like, you know, if there was a consulting engagement, you know, we would take some time, you know, to start off, you know, just with the understanding that, hey, you know, we have to make sure that there's some real signal and the noise in the data as opposed to just noise, you know, that we can actually, you know, differentiate, you know, clear customer segments. And, and that we can actually forecast Well, once we can do that, you know, then we can start applying optimization, it works optimization works extremely effectively. It's deterministic, really, you know, if you Have good inputs, you know. And so it would be, you know, kind of prototyping these ideas, collecting some data, you know, and trying it out a little bit, you know, and kind of, you know, recap, calibrating the models, as we go along through an iterative process, you know, you wouldn't want to stand up a big software platform from the start, you would really want to kind of start calibrating your models, you know, your segmentation, process your models, and then you know, doing it somewhat manually, you know, for, you know, a time period, a fixed time period, you know, and seeing the results and adjusting, you know, for a number of cycles. And once you start to see a repeatable results, that's, that's what tells you, you know, it's time now now we kind of see how this is working, we can see that it's, it's increasing our revenues, you know, substantially over how we were doing it before. And that's the time when you start to build a platform that can automate that process, it can automate the sales, you know, it can incorporate all the data together, you know, and do your pricing for you automatically. You know, and your demand forecasting, you know, and allow you to optimize your capacity like that in a really a streamlined and automated approach.
Ariel Herrera 25:55
That's incredible. A lot better being data driven, and utilizing models and just guessing different pricing numbers and hoping that your revenue climbs up, do you have any hard numbers that you could share of how some of your company's work has increased revenue from x to y? For previous clients?
Unknown Speaker 26:18
Yeah, I can, I'll tell you about a research paper I'm working on right now. Actually, it's applying revenue management in a new a fairly new way. So it's the research is done. I'm writing the paper right now. But basically, what it uses is a reinforcement learning on machine learning techniques to price and offer allow companies who have long lines, basically a way to, for customers to skip the line, and go right on customers who are interested in that. And so we see, we see hints of this being used as a theme parks like Disney in these kinds of things, this works at a much more granular, you know, microscopic level where it's constantly changing pricing, capacity allocation, obviously, you get longer lines, you know, people are willing to pay more people, you know, want to not spend the time in the line, I remember, recently, you know, my teenage daughter, my oldest daughter got her learner's permit. And I remember being a DMV looking around saying, you know, as I was working on this resource thing, man, if only I could pay some money to, you know, to not have to spend half my day here at the DMV. And so, you know, through, you know, through these experiments I've worked on, you know, it really, you know, it increases can bring substantial increases in revenue, you know, 15 to 16%, increases in revenue, without increasing without really increasing your your operational costs. So that's one of the things that revenue management really does is it allows you to increase your revenues without increasing operational costs by smart use of capacity. So that extra revenue really translates mainly into profits, you know, the older revenue management systems, you know, would get, you know, four to 7% increases, you know, typically, you know, of pure revenue, airlines and these kinds of things. And now, the ones that are really focusing on those were the ones that focus more on capacity allocation. Some of the more modern systems that are very focused on the dynamic pricing can increase it, you know, by more than 10% 12% or more, in many cases. So it's, you know, it really does, it really does work, you know, it really does have compelling results, when it's applied properly.
Ariel Herrera 28:16
Yeah, so I'm excited to read your paper once completed. And for the person who's maybe an aspiring data scientist, and wants to learn more about reinforcement learning, price optimization, what are some tips or advice into like, what programming languages, they should look into any specific books that they should read up on that you would suggest?
Unknown Speaker 28:41
Yeah, that's a great question. Yeah, I mean, a lot, you know, obviously, data science is very much a, you know, a programming driven discipline. So, you know, really, you know, the best thing is to work with a programming language, I think, if you're getting started, one that's going to help you connect with all the API's and datasets, Python is really the most popular language out there for data science, it connects to all the API's, it's really easy to work with, you know, so if you don't have programming background already, you know, and you're looking to start, I would, I would definitely recommend starting with Python, and then a couple of other skills to that are important, you know, I think it's important to be able to work with databases, knowing knowing how to work with SQL is going to be an important thing, because that's, you know, doing real data science, you know, you have to get your data from different sources. Still, you know, I mean, there's a lot of text data and these kinds of things, but really, I personally find, still, there's so much tabular data, that's really how most businesses are making their decisions is still working with table based data. So being comfortable with SQL and these kinds of things is really also, you know, in addition to your Python is gonna get you get you far, obviously, you know, it's important to learn some of the basic techniques of data science, you know, so there's a few different facets of it, you know, on the predictive learning side, you know, you certainly want to be, you know, spend some time learning, you know, the different forms of regression models, different forms of classification models, and then different forms of clustering models. Those will give you really a broad base once you understand those. So, Actually a lot of impactful data science work you can start to do. If you're interested in revenue management in particular, you know, another area that you have to get into, which is a little bit more complicated, is mathematical optimization. That's the discipline of operations research. So you know, that we're talking about things like linear and integer programming, and like, things that are used a lot in supply chain, and, you know, different types of resource allocation problems. So that's another topic that you can you really want to start learning as you're getting into revenue management. In terms of revenue management, there's a, there's a great book called pricing and revenue management by Robert Phillips, it's a it's kind of a book that's tailored to, you know, a lot of MBA programs use it for, you know, to teach the discipline of revenue man. So it's kind of a broad base, it looks a different industries, and kind of gives you the core principles, you know, and actually gets you down, not just, you know, not just the broad principles, you know, it does a good job of, I think of giving you the overview, but also helps you drill down into some of the mathematics of how it all works here. So, if you're interested in that particular field, you know, that's a great resource, I would recommend.
Ariel Herrera 31:02
Great suggestions, and I'll link that book to the bottom as well. So I know none of us have a crystal ball, Chris. But if you I guess we're to look back at the last five years and advances of technology in price and revenue management space, and where do you see it going in next five to 10 years? How would you describe that?
Unknown Speaker 31:21
I think one of the big moves, I mean, maybe this goes back a little bit longer than five years, but really, you know, kind of where things are at right now is, things have moved from a predominantly capacity allocation driven approach, which is just making smart use of capacity, that's still a crucial ingredient. But now it added to that is really the more dynamic pricing, you know, getting into real time pricing. And that is really something that has, you know, taken the revenue management software out there, it's really increased the returns, you get on it here. And I think we're gonna start to see that more, I think we're gonna see increasingly dynamic pricing components to the revenue management software. Another thing I think we're gonna see, you know, is, you know, as I mentioned, it started off in the airlines, and then move to hospitality and rental cars, and these other things. I think, now, you know, in E commerce, we're going to start to see, you know, facets of revenue management, you know, not just in services, but you know, in all kinds of products, you know, all kinds of, you know, pricing optimization is probably going to start coming. In many cases where we sell things in general, not just, you know, not just services, you know, revenue management traditionally has worked with cases where you have, like a fixed capacity. So airlines only have, for instance, a certain number of seats on a flight, or hotels only have a certain number of rooms that they can work with, you know, and, you know, dealing with fixed capacity, you know, the perishable goods, you know, so when a seat doesn't sell on a flight, your ability to capitalize on that is forever lost, you know, same with a hotel room with the expiry date of columns, you know, if the day comes and you haven't sold a room, you just don't make money that night. You know, so that's another important thing, you know, it's traditionally been done, you know, and then, you know, the ability to book in advance. And so these are the traditional assumptions. But I think one of the things I see especially like an ecommerce, I think we're going to start to see, you know, that those assumptions are not really required anymore. Revenue Management is adapting beyond those traditional assumptions and being applied in new new ways. You know, and, and adding a lot more dynamicism. To, to how it works.
Ariel Herrera 33:13
Wow, really insightful, and not just going to be constrained to those that have certain capacity for a given period of time, but in other areas as well.
Unknown Speaker 33:24
Not the same assumptions. Yeah. Working beyond the traditional assumptions, I think, yeah.
Ariel Herrera 33:29
Great. Well, I'd love to segue into the audience getting to understand you a little bit better. First off, starting with, what is one habit that you have every day? Hmm.
Unknown Speaker 33:42
Let's see my daily habits. So I have a couple you know what one is, usually when I get up, something that's important to do for me is to go for a five to six mile run each day. And it just helps me clear my mind, you know, and get ready for the day. Another thing that I also find important to do, you know, I like to take some time during each day. And usually on my way to my office, I'll go stop at Starbucks and read for about an hour read the Bible or so for an hour. Just to kind of keep, it's always good to I think take it to keep the big picture in mind. Yeah.
Ariel Herrera 34:14
I think that's excellent. And have you What's the longest you've ever ran for? Just curious?
Unknown Speaker 34:20
You know, it's close to that. Usually, I've done about seven miles, I think. So I'm, I'm not a marathon runner or anything like this, but I do like to try to run, you know, maybe 25 or 30 miles a week. So usually about seven miles I think is probably about as far as I've gone.
Ariel Herrera 34:35
Okay, that's great. I want to try to half marathon and by mile eight, I was kind of limping my way through. Yeah, some more like the two or three miles. But another question I'd love to ask is do you have so much knowledge of resources and such What is your favorite book?
Unknown Speaker 34:56
Hmm. General book or business book Poor?
Ariel Herrera 35:00
Let's go with business book.
Unknown Speaker 35:02
Hmm, let me think about that for a second. So, I mean, I, there's so many wonderful books out there, I think that that help in various ways you know. So I think a lot of the business books that I've read are more tactical, and they help with specific things, wonderful things, I think the most impactful ones that I can think of are ones that are more of the big picture kinds of books. So favorite, I'll just I'll mention a book I'm reading right now, I don't know this. I can say, one favorite or over another, but this one that I find, I find it really interesting, and just really helpful is a book called every good work by Tim Keller. And, you know, I think what it does is it really kind of puts work in perspective, you know, it's, you know, and helps us realize, you know, it's not just about us, but really, you know, the work that we do is intended, it should help others in society, you know, we should we should do good for others, you know, brings things like failure into into a lot more perspective, I think, you know, and gives you a different perspective on these things. So I found it, as I'm going through it and finding it really a refreshing and very helpful book, just for thinking about work in business.
Ariel Herrera 36:04
Yeah, that's great. And that's one thing that you investors may be afraid of is failing, if tenant doesn't pay on time, or if I didn't do a certain measure correctly, does that mean I should just end my career and not start at all. So it's good to have these principles in these books as well, just to keep you going on that train. And what is another like piece of advice, generally, that you would provide for those who are either seeking to dive more into career to technology or for those that are looking to expand and scale their businesses through it.
Unknown Speaker 36:41
Um, there's a couple things I would say, you know, I think one on one hand, you know, is you've got to, the way to grow is to dive right in and start doing, you know, I think it's easy to get stuck in analysis paralysis, but I think I look back on the things that, you know, the big changes in my life, you know, the things, the areas of growth, it's come, you know, at times when I really didn't have a lot of certainty about things, I just jumped in and tried, you know, and you start to figure out your way, I think that's really, really important. You know, being able to do that. Another thing is to constantly keep learning, you know, the world, especially in the technology, and the real estate industry, these these are very, very rapidly changing fields. And you know, what works and what's being used, you know, a short time later, has become passe, and really is no longer that helpful. Of course, there's timeless knowledge, you know, there are some timeless principles, of course, but, you know, you have to really stay on top of the changes, especially in technology, so you have to be prepared, I think, to, you know, it's part of your build into your job, you know, a learning process, you know, a building to your job, you know, some time, you know, each week, to go and learn what's happening, you know, where are the new technology developments? And how can you learn how to, you know, how to capitalize on. And one other thing I'll say, that I think is also really, really important is, you know, as you're doing business, especially, you know, you're thinking about, you know, being on the cusp of breakthroughs, you know, working hard, and you know, it get really intense, and it's always important, I think, to remember to, you know, always prioritize ethics, you know, over profits, you know, it doesn't do you any good to gain the whole world and forfeit your soul to cut corners, you know, so you always, I think, you know, always, you know, even those who you're working with, you know, even those, you know, if you're a landlord, these kinds of things, you want to think about, you know, putting yourself in the shoes of the people who are using your services, right, these kinds of things and, you know, always treat others as you'd like to be treated. I think those kinds of things, you know, even if it doesn't always lead to the the absolute most profits, so make a healthy profit, and you'll enjoy it in doing so
Ariel Herrera 38:36
completely agree with that. And with that, Chris, how would the audience able to reach out to you whether it's for your services, or even just to get some advice within the space that you're specialized in,
Unknown Speaker 38:51
you can reach out to me my university email or my Pollyanna Decision Sciences email. So I'm Chris at you, and I'm sorry, I'm Chris apollyon.ai. If you're interested in you know, some consulting inquiries or just finding out a little bit more about how this works. If you have general questions, you know, you can certainly reach me at C Garcia U M. w.edu, which is my my u and w email address.
Ariel Herrera 39:13
Excellent. Thank you so much, Chris. I really appreciate your time.
Unknown Speaker 39:16
My pleasure. Thank you. Thanks for having me.
Transcribed by https://otter.ai