Episode Transcript
[00:00:00] Speaker A: That segmentation, I think is really important. And I also think the first thing that it brought up for me is what is actually possible.
[00:00:07] Speaker B: Segmentation in SaaS is so important when you think about packaging, right? That right seat, right person, right time still holds true, right? When I think about the bundles and I think about packaging and I think if you're going for a good, better, best, who is that Persona and how are we building a package that suits them?
[00:00:23] Speaker C: The AI is getting sharper about detecting and measuring, calculating demand than it is getting deeper into your personal stuff, right? And using the fact that I have two kids and that I fly certain classes and I work at certain companies and all those kind of things to determine how much I want to or am willing to pay.
[00:00:43] Speaker A: It's really important to think about those higher demand parts too, because as you indicated, that urban legend of when's the cheapest time to buy the ticket does actually line up quite logically for when is the cheapest time to fly.
[00:00:54] Speaker B: With AI, those combinations become almost infinite. So it might feel like the customer is getting a targeted price for them, but it's really just that the number of combinations have gone up at a much higher scale than they were.
[00:01:05] Speaker C: You know, when I read about folks worrying about price gouging, using private data to make decisions, a woman booking a flight and the airlines jacked up the price of the ticket because they somehow the AI knew she had a death in the family and she had to get to that location urgently and therefore raise the price on her.
[00:01:25] Speaker D: Yo, Mike check. What's up everybody? You're listening to the Street Pricing Podcast, the only show where proven SaaS leaders share their mindset and mistakes in pricing so we can all stop guessing and start growing. Enjoy, subscribe and tell a friend. Now let's break it down with your host and sought after slayer of bad pricing, Marcos Rivera.
[00:01:47] Speaker C: What's up and welcome to the Street Pricing Podcast. I'm Marcos Rivera, CEO of Pricing IO. And today we're going to play things a little bit differently. We're going to dig right into a big piece of pricing news. A controversial move by Delta Airlines to use AI in their ticket prices. It created quite a stir. I have two members of the Pricing IO team here to talk about it. Emily and Steve. Welcome to the show.
[00:02:11] Speaker B: I'm Emily. I'm the VP of Strategy and analytics here at Pricing IO I.
Prior to joining Pricing IO and a few other pricing roles, my first job out of college was in revenue management at American Airlines and spent some time at British Airways headquarters.
Also Doing revenue management there. So long history of revenue management.
I was the one who was doing the pricing. That sounds like AI is taking some control over and have a lot of perspective in that regard.
[00:02:39] Speaker A: Hi, I'm Steve. And my revenue management experience is just pre the American Airlines merger. And I spent over a half a decade in revenue management and in that department was engaging directly with the pricing team. Also did a lot of revenue analysis and worked in the operations research team and the ops research team. We were really looking at some of those individual characteristics of passengers as they were booking. And much of that predates the automation that would be really, really helpful in today's environment. And then I also oversaw an inventory management team.
[00:03:14] Speaker C: Excellent. So you got really behind the scenes of not only the, the inputs, but the accounting, the structuring and all that. So I'm, I'm guessing somewhere between a decade and a half of revenue management experience between you guys in the airline and the fact that you help companies with pricing all day long every day. I cannot wait to hear your perspectives. And so let's start, Emily, with you.
You saw the news. It broke out. It came out in an earnings call. The president at Delta said, hey, we're running all these AI pricing pilots. It's right now on 3% of their fares. It started with 1%. They moved it up to 3. They love the result so much that they're going to push it out to about 20% of all fares by the end of 2025. And get this, he said things are going amazingly favorable. He loved the unit economics. He was gushing about it. And I think it was the way, the way he said it and the words that he used, I think really set off some alarm bell, alarm bells out there. So I want to hear your perspective. Delta's move. How did you react? What are you thinking?
[00:04:22] Speaker B: So I think I read all the articles as they were coming out. I think one of the big ones was Forbes saying, you know, could trigger customer trust crisis. What does this mean? Here's why. And my first reaction was, okay, well, airlines have always been about the right seat at the right time at the right price. Right. It's always been revenue management that's been the objective. How do we optimize for that customer? Always been the goal. Historically, there's only been so many levers you could pull to get there. So when I first read this, my gut reaction was, well, this is the same optimization exercise airlines have been doing forever. It's why it's more expensive. If you go to Europe on a Monday and fly back on a Friday, you're more likely to be a business traveler. The fare is going to be higher. Nothing crazy there. So my thought was this is just a step further into that segmentation. So we're taking those really discrete elements we can segment on and taking it a step further to additional criteria and optimizing to allow for more fares. When I was there, there was still this desire. There were limits in how many fares you could have. There were a lot of exercises. And how can we actually increase that? With AI, those combinations become almost infinite. So it might feel like the customer is getting a targeted price for them, but it's really just that the number of combinations have gone up at a much higher scale than they were previously. The limits have kind of been taken off in terms of how much segmentation you can provide.
[00:05:43] Speaker C: So you're telling me that it's basically the same thing on steroids, Right. But with the AI's ability versus what you were doing back when you were working in revenue management.
[00:05:52] Speaker B: That's my hunch. I know Steve and I have chatted a little bit about the inventory side, so I'll let him touch on that. But from a pure pricing mechanics, all airline fares have to be filed somewhere, right? There's ATP code. Is the airline tariff publishing company regulatory, like from a regulation perspective, that has to be done. So they can't just be making up a fare because Marcos comes on delta.com today and it's, well, we know all this criteria about Marcos. We know he lives in San Diego, we know he has two kids. No, but there might be a fare for when someone's booking four tickets, what that might look like. Right. So it might be able to just have that deeper segmentation, but it's not going to be Marcos pops on delta.com and it's a Marcos Fair or an Emily Fair or Steve Fair. It's going to be that just deeper level segment.
[00:06:37] Speaker A: That segmentation, I think is really important. And I also think the first thing that it brought up for me is what is actually possible.
And so understanding those constraints. Emily, as you indicated, fares are filed through ATCO. So there is an opportunity for those fares to be filed, but then you have to make a determination as to whether or not those fares are going to be available for sale. And so it took me back into actually how do you manage revenue management and how do you actually manage those unit economics?
[00:07:05] Speaker B: What?
[00:07:06] Speaker A: When they look good, you're going to dive into them and you're going to be asking yourself the question, what are the key drivers. But one of the key drivers is likely related to how much of that inventory is available for sale. So a real simple example would be if we use a different industry where it's not actually expirable inventory, but you think you're selling two shirts, you've got one that's orange and one that's yellow. And the yellow shirt happens to be half price of the orange shirt.
[00:07:32] Speaker B: Half.
[00:07:33] Speaker A: The question is how many of each of those are you going to make available?
And that's what you're doing on the inventory side. So you may actually choose for that half price shirt to not have nearly as much inventory available. And that's where I can see AI assisting with determining how much demand there is for that particular segmented fare.
[00:07:53] Speaker C: Determining how much demand. So here's the thing, right? You're thinking about, there's a supply and demand equation in all of this. And almost like, I think use the term segmentation, which I love because it really is almost hyper segmentation at this, at this level. But you know, when I read about folks worrying about price gouging, using private data to make decisions, there was one in particular that caught my eye, which is a woman booking a flight. And the airlines jacked up the price of the ticket because they somehow the AI knew it had. She had a death in a family and she had to get to that location urgently and therefore raised the price on her. Folklore and fear is what that smells like to me. But those are the, those are the things that go viral and get people worried and riled up.
[00:08:36] Speaker B: That folklore and fear has always been there. Why I still to this day have people say, oh, is it true I need to clear my cookies before I search a fair ticket? Absolutely not. The airlines wish they had that data, right? They wish they had that. And maybe with this. I know, I think it was. Was it Fetcher or Fletcher fetcher?
[00:08:56] Speaker C: Yeah, with two Rs. That's right.
[00:08:57] Speaker B: Yeah. Fetcher is probably going to get better at that, right? Maybe they will have some of that data, but there is absolutely no way that happened. What probably did happen is she booked a ticket very close to departure. They are always more expensive. And maybe the death in the family happened and she waited three days to book her ticket. Well, guess what, that fare hit an advance purchase cliff and now it became more expensive. So if you're booking a ticket two days in advance, it's always more expensive.
So it maybe feels like it jumps up. And that piece Steve was talking about is really critical on the inventory side. So there could have been one seat available at that cheaper price. She waited an hour to book, now she goes to book it again and that inventory is closed. And that jump could have been substantial because you're so close to departure. Right. And so now this flight's more full again, supply and demand. Airlines love it, but there's so many components to, to it. And so I think that factor. Fiction. Definitely fiction. I feel very good saying that. I think there's a lot as AI does change the airline industry that I don't know about and I won't pretend to be the expert on, but a scenario like that feel very confident saying it's fiction. I think another kind of myth there's always been with airlines is like I said, there's the cookies myth, but there's also what's the cheapest day to book? And everyone always says it's Tuesday, it's Tuesday, it's Tuesday. That always bothered me. I wanted to understand why it was Tuesday. As someone who was again manually physically filing these fares. But you know, it's not too, I'm not filing a sale every Tuesday. What's causing this? So statistically it's Tuesday because that's when Southwest historically has filed sales. So they have enough of the market that it is, yes, statistically, technically cheaper on Tuesday. If it's a market Southwest doesn't fly, it's not necessarily going to be cheaper on a Tuesday. I can't say that's a hard and fast rule, but that's where that came from is they just have enough to actually move the industry because it is, it can feel so random. And I think that's the thing is airline pricing often does feel random. So this perception of it's targeted to me again, that cookie series, been there for years. People thought that before. And so I think the way that Delta communicated the AI shift was a bit damaging because it's, oh, we're able to see this. Look at all these returns we're seeing. Well, great. It's probably just means that you are optimizing those higher demand flights better and that's not going to affect the average consumer. When I'm booking my vacation eight months.
[00:11:22] Speaker A: I think it's really important to think about those higher demand parts too because as you indicated, that urban legend of when's the cheapest time to buy the ticket does actually line up quite logically for when is the cheapest time to fly. And so if we think about days when we might want to fly, we certainly a lot of folks on the business travel side they want to get back Friday afternoon, Friday evening, and those fares are going to be significantly more expensive than if you were to fly on a Saturday. And that is related to that segment of demand. But then also understanding events. And this is where I think AI could really play a role. You're looking for what is the demand for a particular city and how might that change? And so if we use specific events, maybe something like a Mardi Gras New Orleans that follows a lunar calendar. So you're going to need to make an adjustment in the demand forecasting system to account for the lunar calendar shift from one year to the next.
AI would be able to help identify and do that automatically, but it could also help when there's a new special event in town. So if we think about the super bowl, it rotates around from city to city and it would be able to adjust for that. There are also other events that come up and I could actually see AI being able to sweep for that and tune up on the inventory demand side.
[00:12:40] Speaker C: Hey, team, I want to take a quick pause here to ask for a small favor. This show is about helping entrepreneurs remove the guesswork and price with confidence. And it will be a huge help if you can rate the show and share it with a friend who you think is struggling with pricing. Takes about 10 seconds of your time, but it will mean the world to me. Thanks in advance. Now back to the show. So I think the big thing here is the reframing of it really is that the AI is getting sharper about detecting and measuring, calculating demand than it is getting deeper into your personal stuff. Right. And using the fact that I have two kids and that I fly certain classes and I work at certain companies and all those kind of things to determine how much I'd want to or am willing to pay.
But it's more getting sharper around, like you said, inventory. What's driving the seed demand, the cities, the times, the legs, the weather, current events, all that stuff. Getting much, much better at pulling that in. I can buy that story. I can buy that story mainly because I think that if we were to see examples of gouging of, of, you know, AI kind of taking personal information, that stuff will make its way to Congress so damn fast and start increasing all sorts of, I would say, watchdogs on this types of pricing and packaging, which is just creates more burden on the industry as well. So they want to tread very carefully and not to, you know, lead anyone down that path is what I'm thinking.
[00:14:07] Speaker B: And the way Delta communicated it kind of Sent shockwaves through.
[00:14:12] Speaker C: I think it was clumsy. I think, I think what they're doing is smart. But the way they said it was clumsy.
[00:14:16] Speaker B: Every other airline jumped on saying, don't worry, we would never do that.
[00:14:19] Speaker C: Oh, Southwest said it, Americans said it. Yep.
[00:14:23] Speaker B: And never do that is never target with a Marcos versus a Steve Fair. They are, they are segmenting. And so it was very easy for them to get good press on the way Delta said it, because it's very. Again, they will never get that detailed into your, you know, maybe family tragedy, because Delta's not yet either. Right. So it gave them an easy PR move. But I think that price gouging thing you mentioned, Marcos, I remember a situation when I was there and there was a hurricane. Well, the mechanic Steve was mentioning, there's a filed fare and there's an inventory. Inventory system recognized, oh, there's a spike in demand because everyone's trying to get out for the hurricane. So then that kind of went viral. People were saying, oh, I'm trying to leave Miami, there's a hurricane coming and the fares have gone up 4x. So what happened was they actually had to go in and say, here are emergency fares where we're going to cap the price for this emergency time point time band so that we don't get accused of price gouging. Because that's not the intention. But those are just the systems at work. You have to think about how many seats there are flying every single day. I stare at the airport from my desk, right? That thing's busy all day. There are millions of combinations here. And so the inventory system did what it was supposed to. There's a spike in demand. I'm going to raise the price, but I think AI would be able to, you know, you can start to program and have something that's not a human needing to go in and intervene, because when I was there, that's what that was. Someone would have to recognize there's a hurricane in Miami. I need to turn off the inventory system here and make sure people can get out. The airlines fundamentally wanted the people to get out.
And so, yeah, AI will make those type of decisions easier. It'll make for fewer mistake fares. So I do think the travel hackers who love to be able to get that mistake fare, I do think those days are numbered.
AI is going to be a lot quicker at recognizing those than humans are and shutting things off and canceling than previous systems.
You'd be surprised how long some mistake fairs stayed in there until they hit enough bookings to Go viral. I know I benefited from one to fly to Australia in business class for about $1,000 or something crazy like that.
[00:16:34] Speaker C: So I've had a few. Yeah, 200 bucks.
Yep.
[00:16:38] Speaker B: Yeah, we've all benefited from those. I think that those will go away and those are very costly for the airline. So Delta saying they're seeing this benefit, this revenue lift, it's not hard. Those big events Steve was talking about, you get those really right. Really, really right. That's huge. Think of how many types of events, even a conference. There was another situation, I remember when I was there, that a conference got missed and it was a medical conference. It's not something that you would think about in your day to day, but the entire business class was sold out eight months in advance because something small like that got missed. So making all of those tightening of the bolt kind of and not letting that slip through the cracks, you're going to see huge revenue benefit.
[00:17:20] Speaker C: Yeah, I think seeing the other angle too, which is not just trying to take as much money from an individual consumer, but also plugging those leaks, those revenue leaks. There's lots of them out there in the airlines, like you said, millions of variables, lots of stuff changing and happening. And also the potential for price to be lower. Going back to the inventory thing. Right. Of a family of six cancels or a soccer team cancels and all these seats open up, they may want to be more aggressive to get folks in those seats and fill them and drop the price even. Right. So it could happen in multiple ways. Guys, what can SaaS leaders, SaaS operators learn as they're considering? I know there's some out there considering, like how do I use or leverage AI to help me with my revenue management, help me with optimizing and making sure I'm charging for the right stuff. What do they have to learn from this?
[00:18:09] Speaker A: Well, one of the first things that I think we see quite obvious here is that the airlines know what their unitized measures and metrics are and they're measuring that. So I think it is really important for SaaS companies to understand their North Star KPIs and be able to evaluate those and be able to dig in and ask themselves a basic series of questions of why is it performing this way, how did it get there and what are we going to do and change? I think that's one key lesson that is even outside of the AI, but because they so clearly know how to measure their performance, they're able to do that. I think the second thing related to AI is that you should be looking for opportunities to as you just said, plug leaks. Looking for opportunities to track the data down into where might there be an opportunity to more closely aligned value to price?
[00:18:58] Speaker C: Absolutely. What do you think, Emily?
[00:18:59] Speaker B: I think segmentation, something we talk about internally here a lot. Something we talk about with our clients a lot. It's something that when I reflect on my career and see how did I end up here advising SaaS clients all the time segmentation that airlines do is second to none. I will argue segmentation in SaaS is so important. When you think about packaging, right. That right seat, right person, right time component still holds true. Right. When I think about the bundles and I think about packaging and I think if you're going for a good, better best, who is that Persona and how are we building a package that suits them with AI and the ability to use AI tools. Right. Because I don't think that behaviors as transparent to AI. Right There there's no event that's going to jump see SaaS prices jump. Right. But using AI tools to get better and deeper knowledge into what your customers are doing allows for that deeper segmentation, allows for that optimization. The question of should I have three packages or four? Well, what are the segments telling me and how are those days distinct and do they actually have distinct growth patterns, distinct willingness to pay, going deeper with who your who is. We talk about our 5Q a lot. Right. Who is the who being able to go deeper through additional data, I think is a big one. And then the last one, we touched on this earlier, but it's communication. We talk about that with our clients all the time. How you communicate price changes, showing that there's value there. You know, if Delta had done a lot of talking around their new aircraft or seats or what that looks like the experience at the same time as talking about AI and not leading with something. I think they preferred what the investor perception was going to be. I don't know if they thought about the customer perception. And that's what's been going wild. Right. I think investors were happy that the customers are not and they're quite fearful. Same thing with SaaS. How you communicate matters so much. And this isn't as scary as everyone thinks it is. But you don't know what you don't know. And airline pricing has always been a black box that translates to sas. Do your customers don't know the mechanics of why you're deciding where they land or what's happening? Exactly. Or you know, they know you're adding new features. They don't know what they are and they don't know what the value prop is. Right. You need to be able to communicate that value to them very clearly. And so I think the communication piece is really key.
[00:21:15] Speaker C: Absolutely. Totally agree. And I think that if you're not going to help with creating that transparency and make folks understand what you're trying to do, they will make assumptions for you and oftentimes it's not in your favor. Right. So, Steve, fully agree, man, you got to know what that North Star is, know who you are, what you're after. And Emily, I think we do now have a better opportunity more than ever to get deeper into segmentation, understanding our customers. What's the same, what's different? To really be able to exchange value in the best way possible. So high five to both you guys. Thank you for the, for dispelling the myth of the Tuesday for me because I, I'm a Tuesday booker and, and all the things that go behind the scenes of airline pricing. So lot more myth, lot more fluff, lot more fear than what's actually going on that's bad for consumers.
But we're still going to watch really closely here. I mean, I think other industries like hotels and cruises and concert tickets also will be experimenting with AI as part of their variables and their prices. And I think this is just the wave of the future, and I think we're going to start getting used to it, but we should really think it through on all angles, not just the, you know, jumping to the conclusion that I'm screwed. Right. In many cases where I think most folks jump to. So thank you guys for jumping on the show today. Appreciate you and team. That was Emily Sanz and Steve Inman coming in, giving us some inside scoop and perspective on Delta Airlines latest move in using AI to price their fares. All right, take that lesson. Think about some of the things we talked about, the North Star, the segmentation, the transparency and trust, and bring that back home and take one step away from that guesswork. And remember, stop guessing and start growing.
[00:22:58] Speaker D: Until next time, thank you and much love for listening to the Street Pricing podcast with Marcos Rivera. We hope you enjoyed this episode. And don't forget to like and subscribe. If you want to learn more about capturing value, pick up a copy of Street Pricing on Amazon. Until next time.
