August 28, 2025 • 27 mins
Is AI-powered personalized pricing the next frontier—or the next fiasco—for customer loyalty and corporate reputation?
That’s the burning question I dive into with my guest, Dr. Peter Fader, on this episode of the Delighted Customers podcast. With Delta Air Lines rolling out AI-driven pricing and the world watching closely, the conversation takes a provocative turn: Should the C-suite pursue every possible drop of revenue if it risks alienating the very customers whose business it depends on? Dr. Fader walks us through the razor’s edge between innovative revenue optimization and the real dangers of going “too granular” with pricing, shining a light on the long-term risks few companies are thinking about today.
Why should you tune in? Dr. Peter Fader is not just an academic—he’s a trailblazer whose work bridges the gap between data science, behavioral psychology, and pragmatic business leadership. His pioneering research on customer lifetime value and predictive analytics has redefined how top firms measure, value, and grow their customer bases for sustainable profitability—not just quick wins. If you work in pricing, revenue management, customer experience, or strategy, you need to hear his take on dynamic pricing, the emotional minefields of AI, and how to make decisions that customers—and shareholders—can live with tomorrow, not just today.
Here are three key questions Dr. Fader answers in this episode:
-
What’s the real difference between smart dynamic pricing and dangerous personalized pricing—and why should every business care?
-
How can companies avoid the temptation of “slicing the bologna too thin” with AI, and what are the hidden long-term costs of over-optimizing for short-term profit?
-
Where are the true opportunities and ethical boundaries in customer-based valuation, and what lessons can firms learn from industries like airlines and pro sports?
Listen now and subscribe so you never miss an episode—find us on Apple Podcasts and Spotify, or on any of your favorite podcast platforms.
Meet Dr. Peter Fader
Dr. Peter Fader is a Professor of Marketing at the Wharton School at the University of Pennsylvania, renowned for his transformative contributions to customer analytics, customer lifetime value, and predictive modeling in business. His influential books include “Customer Centricity: Focus on the Right Customers for Strategic Advantage,” “The Customer Centricity Playbook,” and “Customer-Base Audit.” Beyond academia, Peter made waves as co-founder of Zodiac, a predictive analytics firm acquired by Nike in 2018, and continues pushing the boundaries at Theta, a company focused on customer-based corporate valuation for private equity and enterprise clients.
Peter’s work stands out for blending rigorous quantitative modeling with real-world application, helping executives across industries balance the drive for immediate revenue with strategies that nurture lasting customer relationships and true organizational value. His recognitions include teaching and research awards, and he is frequently sought after for his insights by global brands, sports franchises, and media.
Connect with Peter on LinkedIn: Dr. Peter Fader
Show Notes/References
-
Delta Airlines’ AI-driven strategic pricing news: Delta, Fetcherr, and AI Pricing Example
-
Customer-Based Valuation (Theta): Theta
-
Prospect Theory – Daniel Kahneman & Amos Tversky: Nobel Prize in Economic Sciences 2002 – Daniel Kahneman
-
Dr. Fader’s books:
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Today on the Delighted Customers podcast, I have a very special
returning guest. Peter Fader is a professor of marketing at the
Wharton School of the University of Pennsylvania. He
came on the show, he was one of my first guests. He was episode
eight and nine, a two part special episode. We featured
his book Customer Based Audit where we talked a little bit about
(00:22):
customer lifetime value and the valuation of firms. His influence
in the realm of customer analytics is far reaching, marked by
impactful books, three of them, and hosting awards
celebrating his research and teaching. He stands out
for his exceptional ability to bridge the gap between
academic insights and pragmatic business applications.
(00:44):
And a prime illustration of that is Zodiac, his first venture
into predictive customer analytics, which Nike acquired
back in 2018. He remains passionately
dedicated to pushing the boundaries in this field
at Theta and I'm so excited, Pete, to have you back on the
show. Mark, always a pleasure talking to you. I appreciate the opportunity.
(01:06):
Well, you put a post out a little while back when
Delta Airlines announced a use of
AI in their strategic pricing. And let me
just share a little bit about the quote from an article from Delta.
They're rolling out an AI driven pricing system that aims to set
fares for for up to 20% of their flights by the end of
(01:28):
2025, a sharp increase from the 3%
affairs determined by AI algorithms in the past. And this
is part of a partnership. They're doing a Fetcher company, a
company that specialized in this. And so you, you warned
people that this might be not such a good idea,
especially that we're going the way they're going about it. So let me just
(01:50):
outline a couple of things related to how they plan to do it and
how it works. Talking about real time adjustments reminds me of
congestion pricing and also what you see on the Beltway and in
the Capitol Beltway in Northern Virginia where pricing goes up if you want to ride
in the express lane to like an exorbitant amount, they do fair
recommendations. As sort of a super analyst working around the clock to
(02:12):
update price points and personalized offers over time.
They hope each fair will be a custom offer
specific to each traveler. So with laying out those
things, have I have I did a good, done a good job of laying out
what exactly we're talking about here? Well, actually yes and no,
because the original article I saw on it exactly as you
(02:35):
described. But Delta has walked it back a
bit and even Fetcher itself says, wait a minute, we're not doing
anything that's personalized per se. This is
a really, really important issue. Still, it might Be a little bit too
fine grained for my taste, even if it's not one to one
pricing. But there's a bit of
(02:58):
disagreement or lack of clarity about just how personalized
it really is. Okay, so before we get into some of the real
danger zones, why should this be a big deal to people in
the C suite? Well, there's the opportunity and then
there's the cost and the risk. The opportunity to
set better prices, to make them more flexible,
(03:20):
to have them better reflect demand and supply conditions.
That's kind of, you know, golden opportunity. And not just for
airlines, but really for any firm. But then there's the risks of
either overdoing it and being maybe a little bit too
smart, as well as the potential PR backlash.
There's all kinds of ethical concerns and even potential legal concerns.
(03:42):
So the downside is huge, the upside is huge.
How to find the right balance? Well, we're going to find out the hard way.
Well, and so now we have this opportunity, if you want
to call it that, through AI and it, it's very tempting, right,
for to just squeeze as much profit as you can out of
your customer base. What do you see? You mentioned some of the dangers
(04:05):
relative to customers. Talk about it from a pragmatic standpoint
and also sort of a macro standpoint and also from
an emotional standpoint. Sure. So at a macro standpoint, there's
no doubt that airlines want to maximize revenue.
Personally, I think that's a mistake. I don't think that's the right objective. Going back
to our previous conversation, they should be trying to maximize
(04:27):
the profitability of their customer base, the long run
profitability of their customer base. Because when you talk
about, you know, squeezing all that profitability, what people are going to think,
maybe rightly so, is they're going to be gouged. They're going
to kind of take all of that from me now. And you know what? That's
going to hurt. I'm not sure I'm coming back. I'm not sure I'm going to
(04:48):
keep doing this. So if you think about it from a lifetime value
standpoint, gouging your customers would be a terrible idea.
You're not a charity, you're not giving stuff away here. But if we want to
maximize the net present value of the customer base, which is what
every firm should do, that really is the fiduciary obligation
that they have to their stakeholders. That's going to be different than trying
(05:11):
to maximize short term revenue. It's going to lead to very different
outcomes, very different processes. And very different
reactions from the customer base and other stakeholders. So if
I could rewind my own. Going back many, many years, the
four Ps of marketing price made it into one of the
four Ps. And one of the, one of the aspects of it is, yes,
(05:32):
it's connected directly to revenue that an organization can bring in, which
connects to profit, but also it's somehow connected to the
perceived value of the offering. Right, and so how does that
go haywire with AI type pricing? Or could it. That's
exactly right. It can and it will. And this is not necessarily a
knock against any AI per se. It's easy to kind of paint that as
(05:55):
the boogeyman. But it's the kind of lack of control, the
lack of accountability, and maybe just a bit too much flexibility,
a bit too much precision on the part of
the part of the price setters. That AI just enables us
to. Even if we're not pricing at an individual level,
let's just do something that you alluded to, which would be changing pricing at a
(06:17):
very granular level, let's say from hour to hour. And sometimes when you
start slicing the bologna a little too thin, it can
backfire both from an economic standpoint as well as an
emotional one. And again, that's where you're going with this. The price is
more than just an economic lever. It is an emotional
one. And people just have very strong feelings about
(06:39):
the price levels. I mean, even, you know, specific numbers,
you know, if it ends in 0, 0 versus 9, 9, it's amazing
how impactful these things can be. People are also very responsive to
price changes. So it's not just a matter of what we're charging
now, it's what we're charging now to what we were charging yesterday
or the last time I bought your products or services.
(07:02):
My belief, I could be wrong about this, I really don't know Fetcher's
algorithms, but I really believe they're leaning heavy duty into the
optimization and either ignoring or at
least downplaying some of these psychological issues that
again can lead to backlash and can lead to, you know,
underperformance of the algorithms. So one example as
(07:24):
you just described is if I'm going from point A to point B on a
flight, and I booked that flight today, and then I have another
flight, I don't know, a month from now, not even that far
away, and it's double the price, I'm going to start saying, hey, what,
what happened? What's going on with this particular airlines or offering?
Right? And that creates Some angst for me, no question. In fact,
(07:46):
the single most impactful paper I ever wrote, it was
actually 30 years ago, is on this idea of reference price
effects, that we all have reference prices and when we look at the
current price of an item, we code it as gains
or losses relative to that reference price. Now it might
be the reference price, might be the price I last paid, it might be the
(08:09):
price that a competitor's offering, it might be the price even of unrelated
goods, perhaps. But the idea of reference pricing,
the idea, as some folks listening to this might be familiar with
the notion of prospect theory, that's the thing that Danny
Kahneman won the Nobel Prize for. Well, one of the many things
that he and his late co author Amos
(08:30):
Tversky did. This is the way that people react
to pricing decisions and pricing changes. And a lot of these
optimization algorithms do not account
for it. They just look at price in this very objective way
as if we are just bots making these trade offs
between price and quality, but not taking these more
(08:52):
nuanced and these more impactful dynamics into account.
So it affects loyalty. It could have a big
potential impact on loyalty where I feel like, hey,
either you're gouging me or you're playing games with the pricing
and I don't like that. Exactly. And that consequence,
yeah, maybe you'll get me to pay more for that fare or
(09:15):
whatever the cup of coffee or whatever it might be today, because I have no
choice and I have to do it. But I'm not going to stay with you
as long, I'm not going to interact with you as often. I'm not going to
say good things about you. It's easy for us or it's easy for
economically minded people and firms to dismiss some
of that stuff, say, nah, that's just fluff, that's just cheap talk.
(09:37):
But A, we know that it's not and B, in this day and
age, again, going back to previous conversations we've had, we
can quantify the qualitative and measure
the impact of it and say, how much are we gaining or
losing? And more importantly, how could we make these
policies more effective for the long run value
(09:59):
of the customer? So I think that's just so important because you, you keep
reframing it back to the long run when there's so much
pressure, I call it Corduris, where public
companies are trying to meet shareholder and the market's needs.
And yet what we're asking here is to say someone's going to come to
you, whether it's from the finance department or
(10:21):
product or somewhere else saying, look, here's on the black and
white. You know, you can, you can optimize pricing at
1 2% improvement in pricing, could,
you know, add millions to the bottom line. What would you, what would you say
to them? That by using both the kind of
psychological perspectives that I've been alluding to, as well
(10:43):
as the statistical models around them that really not only validate that
work, but let us operate on it, we can do a better job of
projecting what long run revenue profitability is going to
be through a more appropriate customer oriented lens.
And we can say, yep, you're going to make more money today, but you're going
to make less money in the long run. And we can do that in a
(11:04):
way that's not just, you know, hey, pay attention to us marketers,
but let's do it in financial terms. Let's win
over the people in finance and accounting and
FP&A to show them that we can help them do their
job more effectively by taking some of these metrics,
models and patterns into account. We've had great success with it.
(11:26):
As you mentioned, you sold one company, started this new one
Theta. And a big focus of that company is this idea of customer
based corporate valuation. And we have the largest private equity firms in the
world asking us what's the long run value of that company we're
thinking of buying, what's the value of his customers? How much
more value can we create and then squeeze out of it? But
(11:49):
the more and more recognition that the long run can be and
should be quantified. And we've seen it even on the dynamic pricing
side, airlines, they still haven't woken up yet. We have some
wonderful case studies working, for instance, with a number of major league
baseball teams where a lot of them would start with a more rational
economic short term dynamic pricing optimization
(12:11):
approach. And we've shown them, you know what, by taking some of these underlying
customer behaviors into account, we can come up with better policies that
will serve better in the long run and also be more holistic,
not only looking at what you paid for that ticket, but trying to be more
customer centric. Let's have it spill over to the merchandise
that you might buy, whether or not you actually frequently
(12:34):
the advertisers who are advertising in the venue. So
trying to take both a broader and more psychologically appropriate
perspective. Big part of that is, like I said before,
don't slice the bologna too thin. A little bit more stability,
a little bit more predictability. Instead of having prices Bouncing all over.
It's better for everyone. So I'm curious to double tap
(12:56):
on two things. One in this order, if it's okay, is
if you could go one layer deeper, not all the way down in the
weeds, but one layer deeper into how. And
you, by the way, you do a fantastic job. I can only imagine the level
of granularity you get into in the classroom with your classes on
this, algorithms and so forth. But just one layer, like
(13:19):
sort of from a logic flow of how you
get to that point of short term
profitability and connection to the these price
for example. We're talking about price now as a driver of
loyalty or customer value and the long
term value for the business owners love it. And it is so
(13:40):
important. Goes back to my training as a math major
at the Massachusetts Institute of Technology.
Plus years ago I was on track to become an actuary.
You know the people who kind of set the insurance rates and they
know to not get too granular,
they're not going to say how old are you, Mark Slayton going to be when
(14:02):
you die. They know that while they can come up with the best
guess about that, it's probably wrong. They wouldn't want to be held accountable for
it. So instead they're going to say people who share the
same characteristics as you, the same relevant
behavioral characteristics as you. What percent of that
micro segment will live to be 85 years old? It makes all
(14:24):
the sense in the world. Right? Again, you don't want insurance policies
doing the one to one thing. Again, not only would it be kind of creepy
and unethical and perhaps illegal, it would be ineffective
because the variability in our estimate of any
one person is huge. So I build the
world's best, best lifetime value models, but I know my
(14:47):
limitations. Now. Yes, I can tell you what you're worth, but I don't
necessarily want to be held accountable for it nearly as much as I'd rather make
a statement about the micro segment of the Mark Slaytons. What
are they collectively going to be worth and what kinds of policies should we develop
for them? So I'm not into the one to one thing. And it's
not only because of the ethical, legal, all that sort of thing.
(15:09):
It's because statistically it's really, really hard to justify
that. I do want to get granular, but not too granular
because we know the limits of the statistics. Yeah,
that's a beautiful way of describing it. And maybe you've answered,
you tell me, answered the next question which is connected to that, which is
I Think offline. You described this to me as the third rail, but is
(15:32):
this personal? Is it personalized pricing? What is the danger of
that? Are we really talking about getting too granular? So
the big danger is. I don't know what the biggest danger is. There's a
lot of big dangers. One is just the kind of PR backlash.
So no one remembers this anymore when Amazon, which does a lot
with dynamic pricing, and we got to really clarify the distinction, but
(15:55):
sometime around 2000, they ran one little bitty experiment where they got
into personalized pricing, where you and I would see different prices for the same
book at the same time. Credible backlash. You know, the thing went viral
and they backed away from it. And as far as I know, they've never tried
it again. So people just don't like the idea
that you and I, at the same time for the same item could be paying
(16:17):
different prices. Dynamic pricing, fine. Dynamic pricing. Saying
that, you know you're paying a different price than the person next to
you on the airplane, but that's because she bought the ticket two weeks earlier
and so she got a better price. That's okay. You would have gotten that price
too if you had bought it back then. Dynamic pricing, cool.
Again, we don't want it to be too granular where we're changing the price every
(16:39):
second because again, we start getting in the weeds, but kind of
gradually, occasionally changing. Dynamic pricing, I'm
a big advocate of that. The other problem with the personalized pricing, and
especially when you start bringing the AI in, is that not only will
it get too granular, but the factors upon which it
might make those decisions might start to get into issues of
(17:01):
legality because we can't discriminate on the base of
gender or race and other kinds of personal characteristics of.
I'm not a lawyer, don't push me on it.
But once we start bringing in the AI and we start looking at
people's media habits or who they're connected to in the social
network or what other kinds of products they've purchased. In
(17:23):
theory, that stuff is kosher, but it does start
getting real close to that third rail. And if we find
that certain demographic groups tend to do more
things than others, then we could find ourselves in a whole lot of trouble.
First degree price discrimination is just a very
dangerous thing to be working with, both because of the perceptions and because of the
(17:46):
legalities. And to me, the statistical issues.
Love it. So we talked also a little bit offline about the
possibility of other industries adopting this.
And what do you see as maybe those
Industries that could be ripe for falling into this trap. And what
advice would you say to give them? Well, for one thing is
(18:10):
move slowly. That's really important. Again, we want to dip our toes in the
water with dynamic pricing. Like I said, I had the privilege
of working with Major League Baseball and it was really fun to work with a
number of clubs where I think I won't name names, but one club
said, you know what, we'll just try doing this one section in the stadium. Just
let us just try that. Other clubs would say, well, we'll just do a couple
(18:31):
of games and even then we'll just have kind of maybe two
tiers of prices. So it was starting slowly, it was
experimenting, it was looking at the results very, very
objectively and trying to do so. Not just which one made
us more money over the time that we did that campaign, but the long
run implications. Did it chase season ticket holders away?
(18:54):
Did it get people to engage more and spend more
on things in the, in the store, at the concession
stands? So having again, a broader long term
perspective on it, it's such a tempting drug. I
mean that article about Delta gained so much attention.
Again, some people saying dangerous, other people saying opportunity, that
(19:16):
with great power comes great responsibility. And we just
need to look slowly, carefully, thoughtfully look. The fact that
Delta said they're doing it just for 20% of, of
folks, that's great, that's terrific. That's a good sign. And again,
I think that the, the changes that they're talking about might not be
nearly as dramatic and concerning as what some of the original
(19:38):
press suggested. So, so they might be okay here. And for the most
part, Delta is not only a wonderful airline, but they're pretty
smart about things. So we might be jumping to
conclusions there, but it, but it is an interesting cautionary
tale for other firms that aren't as smart, don't, that don't
have quite the, the analytics sa that don't have the experience running
(20:00):
these kinds of experiments. Take it slowly, but at the same time
try it. Don't be afraid to try it. A lot of companies
say we're just never going to do that, never going to do that. We don't
want to even take that first baby step because it's a slippery slope. That's a
mistake as well. I think that thoughtful,
dynamic pricing is actually, I think not only
(20:20):
smart. I think it's getting to the point where it's going to be kind of
an obligation for companies to be doing it. Maybe disclosing
some of the limits of what they do,
and maybe even being held accountable by their investors and other
stakeholders for doing it appropriately. Yeah. One of the things
that comes to mind is you're sharing and I think about theta and if
(20:43):
I'm listening and I'm a business owner, so I think in Major League Baseball and
some of the larger organizations that I'm sure you guys do work with, but
there's a long tail of small, medium sized
businesses that really drive this country. And, and I'm sure
they're listening closely about. Well, you know,
part of the question is, well, I don't nearly have enough data to
(21:05):
make those kind of decisions. Is that true? Well, it's kind of interesting
and almost ironic. If you're really a teeny tiny
business and you're just selling, you know, bespoke handmade
crafts, then you are doing, I don't know, it
sounds so evil first degree price discrimination. You're
going to figure out what that person's willing to pay and you're going to charge
(21:26):
them appropriately. You don't necessarily have any list prices. When you're a teeny
tiny company and you're figuring it out, there will be a bit of
negotiation, there will be a bit of flexibility. You're just
trying to figure out what this stuff is worth and what people are willing to
pay for it. The issue starts as you start to scale the
business and instead of leaving those pricing decisions in the hands of the
(21:48):
founder and as you start having, you know, multiple stores, you're operating
in multiple business units. That's when you want to start to
standardize things. That's when you just want to start bringing in the algorithms.
And at first, when you have this sense that you don't have a lot of
data, let's just figure out what the price is or the pricing policy. Let's lock
it in and stay with it. We have bigger, more important things to worry about.
(22:09):
But at some point you kind of owe it to yourself to start saying, maybe
we could get a little bit more sophisticated with the pricing.
So yeah, a big part of it is data is having granular
customer level data to tag and track to know not
just how responsive the customer is to our
price changes, but how differentially responsive different
(22:31):
customers are and tying those differential price
responses back to their lifetime value. If we find out
that our best, best customers are the ones who are most price sensitive, which
by the way, usually isn't the case, then we got to be really careful about
that. Maybe we'll use just a much simpler, clearer pricing policy,
but if we find out that our best customers are relatively price
(22:53):
insensitive because they know we're going to give them a good deal. They know we're
going to treat them well. They know it's much more about than price. It's service,
selection, quality, convenience. Then maybe we have a little bit more wiggle room there.
So it's really important to understand how the customers differ
from each other in pricing, price sensitivity, in
value, and just broader engagement. Great advice. I
(23:15):
learn so much every time we are together.
And thank you so much. I want to end this conversation with the
same question I ask all my guests, which is what delights you as a customer?
What delights me as a customer is service that's. And it's not
just necessarily rolling out the red carpet. It's not just kind of giving me
free stuff that I might have paid for otherwise. I think that often is a
(23:37):
mistake. Is incredible insight that goes
beyond the kind of training play. But I'm going to tell you a story here.
My favorite customer service experience. I have a thing and
maybe this makes me a bad person. I don't know. I like Ferragamo
shoes. They're just great. And so I had a pair of Ferragamos.
I kind of wore them out. And I went to the Ferragamo store in fifth
(23:58):
Avenue, New York. I walked in there, I said, I need a new pair of
shoes. I mean, literally standing at the door and this, this woman looks down
at my feet and she says, you are wearing an 11D.
You need a 10 and a half. E. Oh, my gosh. Bought two pairs of
shoes. Wow. It's that kind of thing that, I mean, it
is surprise and delight as we often talk about it. Yeah. But it's
(24:20):
unscripted surprise and delight. And that just shows that
this person knows her stuff. This company knows its stuff.
That's amazing. Love it. And you know your stuff. And I thank you so
much for returning back as a guest on the delighted customers podcast.
Mark, anytime. I'm happy to talk to you and I hope that
people heed the message.
Today on the Delighted Customers podcast, I have a very special
returning guest. Peter Fader is a professor of marketing at the
Wharton School of the University of Pennsylvania. He
came on the show, he was one of my first guests. He was episode
eight and nine, a two part special episode. We featured
his book Customer Based Audit where we talked a little bit about
(00:22):
customer lifetime value and the valuation of firms. His influence
in the realm of customer analytics is far reaching, marked by
impactful books, three of them, and hosting awards
celebrating his research and teaching. He stands out
for his exceptional ability to bridge the gap between
academic insights and pragmatic business applications.
(00:44):
And a prime illustration of that is Zodiac, his first venture
into predictive customer analytics, which Nike acquired
back in 2018. He remains passionately
dedicated to pushing the boundaries in this field
at Theta and I'm so excited, Pete, to have you back on the
show. Mark, always a pleasure talking to you. I appreciate the opportunity.
(01:06):
Well, you put a post out a little while back when
Delta Airlines announced a use of
AI in their strategic pricing. And let me
just share a little bit about the quote from an article from Delta.
They're rolling out an AI driven pricing system that aims to set
fares for for up to 20% of their flights by the end of
(01:28):
2025, a sharp increase from the 3%
affairs determined by AI algorithms in the past. And this
is part of a partnership. They're doing a Fetcher company, a
company that specialized in this. And so you, you warned
people that this might be not such a good idea,
especially that we're going the way they're going about it. So let me just
(01:50):
outline a couple of things related to how they plan to do it and
how it works. Talking about real time adjustments reminds me of
congestion pricing and also what you see on the Beltway and in
the Capitol Beltway in Northern Virginia where pricing goes up if you want to ride
in the express lane to like an exorbitant amount, they do fair
recommendations. As sort of a super analyst working around the clock to
(02:12):
update price points and personalized offers over time.
They hope each fair will be a custom offer
specific to each traveler. So with laying out those
things, have I have I did a good, done a good job of laying out
what exactly we're talking about here? Well, actually yes and no,
because the original article I saw on it exactly as you
(02:35):
described. But Delta has walked it back a
bit and even Fetcher itself says, wait a minute, we're not doing
anything that's personalized per se. This is
a really, really important issue. Still, it might Be a little bit too
fine grained for my taste, even if it's not one to one
pricing. But there's a bit of
(02:58):
disagreement or lack of clarity about just how personalized
it really is. Okay, so before we get into some of the real
danger zones, why should this be a big deal to people in
the C suite? Well, there's the opportunity and then
there's the cost and the risk. The opportunity to
set better prices, to make them more flexible,
(03:20):
to have them better reflect demand and supply conditions.
That's kind of, you know, golden opportunity. And not just for
airlines, but really for any firm. But then there's the risks of
either overdoing it and being maybe a little bit too
smart, as well as the potential PR backlash.
There's all kinds of ethical concerns and even potential legal concerns.
(03:42):
So the downside is huge, the upside is huge.
How to find the right balance? Well, we're going to find out the hard way.
Well, and so now we have this opportunity, if you want
to call it that, through AI and it, it's very tempting, right,
for to just squeeze as much profit as you can out of
your customer base. What do you see? You mentioned some of the dangers
(04:05):
relative to customers. Talk about it from a pragmatic standpoint
and also sort of a macro standpoint and also from
an emotional standpoint. Sure. So at a macro standpoint, there's
no doubt that airlines want to maximize revenue.
Personally, I think that's a mistake. I don't think that's the right objective. Going back
to our previous conversation, they should be trying to maximize
(04:27):
the profitability of their customer base, the long run
profitability of their customer base. Because when you talk
about, you know, squeezing all that profitability, what people are going to think,
maybe rightly so, is they're going to be gouged. They're going
to kind of take all of that from me now. And you know what? That's
going to hurt. I'm not sure I'm coming back. I'm not sure I'm going to
(04:48):
keep doing this. So if you think about it from a lifetime value
standpoint, gouging your customers would be a terrible idea.
You're not a charity, you're not giving stuff away here. But if we want to
maximize the net present value of the customer base, which is what
every firm should do, that really is the fiduciary obligation
that they have to their stakeholders. That's going to be different than trying
(05:11):
to maximize short term revenue. It's going to lead to very different
outcomes, very different processes. And very different
reactions from the customer base and other stakeholders. So if
I could rewind my own. Going back many, many years, the
four Ps of marketing price made it into one of the
four Ps. And one of the, one of the aspects of it is, yes,
(05:32):
it's connected directly to revenue that an organization can bring in, which
connects to profit, but also it's somehow connected to the
perceived value of the offering. Right, and so how does that
go haywire with AI type pricing? Or could it. That's
exactly right. It can and it will. And this is not necessarily a
knock against any AI per se. It's easy to kind of paint that as
(05:55):
the boogeyman. But it's the kind of lack of control, the
lack of accountability, and maybe just a bit too much flexibility,
a bit too much precision on the part of
the part of the price setters. That AI just enables us
to. Even if we're not pricing at an individual level,
let's just do something that you alluded to, which would be changing pricing at a
(06:17):
very granular level, let's say from hour to hour. And sometimes when you
start slicing the bologna a little too thin, it can
backfire both from an economic standpoint as well as an
emotional one. And again, that's where you're going with this. The price is
more than just an economic lever. It is an emotional
one. And people just have very strong feelings about
(06:39):
the price levels. I mean, even, you know, specific numbers,
you know, if it ends in 0, 0 versus 9, 9, it's amazing
how impactful these things can be. People are also very responsive to
price changes. So it's not just a matter of what we're charging
now, it's what we're charging now to what we were charging yesterday
or the last time I bought your products or services.
(07:02):
My belief, I could be wrong about this, I really don't know Fetcher's
algorithms, but I really believe they're leaning heavy duty into the
optimization and either ignoring or at
least downplaying some of these psychological issues that
again can lead to backlash and can lead to, you know,
underperformance of the algorithms. So one example as
(07:24):
you just described is if I'm going from point A to point B on a
flight, and I booked that flight today, and then I have another
flight, I don't know, a month from now, not even that far
away, and it's double the price, I'm going to start saying, hey, what,
what happened? What's going on with this particular airlines or offering?
Right? And that creates Some angst for me, no question. In fact,
(07:46):
the single most impactful paper I ever wrote, it was
actually 30 years ago, is on this idea of reference price
effects, that we all have reference prices and when we look at the
current price of an item, we code it as gains
or losses relative to that reference price. Now it might
be the reference price, might be the price I last paid, it might be the
(08:09):
price that a competitor's offering, it might be the price even of unrelated
goods, perhaps. But the idea of reference pricing,
the idea, as some folks listening to this might be familiar with
the notion of prospect theory, that's the thing that Danny
Kahneman won the Nobel Prize for. Well, one of the many things
that he and his late co author Amos
(08:30):
Tversky did. This is the way that people react
to pricing decisions and pricing changes. And a lot of these
optimization algorithms do not account
for it. They just look at price in this very objective way
as if we are just bots making these trade offs
between price and quality, but not taking these more
(08:52):
nuanced and these more impactful dynamics into account.
So it affects loyalty. It could have a big
potential impact on loyalty where I feel like, hey,
either you're gouging me or you're playing games with the pricing
and I don't like that. Exactly. And that consequence,
yeah, maybe you'll get me to pay more for that fare or
(09:15):
whatever the cup of coffee or whatever it might be today, because I have no
choice and I have to do it. But I'm not going to stay with you
as long, I'm not going to interact with you as often. I'm not going to
say good things about you. It's easy for us or it's easy for
economically minded people and firms to dismiss some
of that stuff, say, nah, that's just fluff, that's just cheap talk.
(09:37):
But A, we know that it's not and B, in this day and
age, again, going back to previous conversations we've had, we
can quantify the qualitative and measure
the impact of it and say, how much are we gaining or
losing? And more importantly, how could we make these
policies more effective for the long run value
(09:59):
of the customer? So I think that's just so important because you, you keep
reframing it back to the long run when there's so much
pressure, I call it Corduris, where public
companies are trying to meet shareholder and the market's needs.
And yet what we're asking here is to say someone's going to come to
you, whether it's from the finance department or
(10:21):
product or somewhere else saying, look, here's on the black and
white. You know, you can, you can optimize pricing at
1 2% improvement in pricing, could,
you know, add millions to the bottom line. What would you, what would you say
to them? That by using both the kind of
psychological perspectives that I've been alluding to, as well
(10:43):
as the statistical models around them that really not only validate that
work, but let us operate on it, we can do a better job of
projecting what long run revenue profitability is going to
be through a more appropriate customer oriented lens.
And we can say, yep, you're going to make more money today, but you're going
to make less money in the long run. And we can do that in a
(11:04):
way that's not just, you know, hey, pay attention to us marketers,
but let's do it in financial terms. Let's win
over the people in finance and accounting and
FP&A to show them that we can help them do their
job more effectively by taking some of these metrics,
models and patterns into account. We've had great success with it.
(11:26):
As you mentioned, you sold one company, started this new one
Theta. And a big focus of that company is this idea of customer
based corporate valuation. And we have the largest private equity firms in the
world asking us what's the long run value of that company we're
thinking of buying, what's the value of his customers? How much
more value can we create and then squeeze out of it? But
(11:49):
the more and more recognition that the long run can be and
should be quantified. And we've seen it even on the dynamic pricing
side, airlines, they still haven't woken up yet. We have some
wonderful case studies working, for instance, with a number of major league
baseball teams where a lot of them would start with a more rational
economic short term dynamic pricing optimization
(12:11):
approach. And we've shown them, you know what, by taking some of these underlying
customer behaviors into account, we can come up with better policies that
will serve better in the long run and also be more holistic,
not only looking at what you paid for that ticket, but trying to be more
customer centric. Let's have it spill over to the merchandise
that you might buy, whether or not you actually frequently
(12:34):
the advertisers who are advertising in the venue. So
trying to take both a broader and more psychologically appropriate
perspective. Big part of that is, like I said before,
don't slice the bologna too thin. A little bit more stability,
a little bit more predictability. Instead of having prices Bouncing all over.
It's better for everyone. So I'm curious to double tap
(12:56):
on two things. One in this order, if it's okay, is
if you could go one layer deeper, not all the way down in the
weeds, but one layer deeper into how. And
you, by the way, you do a fantastic job. I can only imagine the level
of granularity you get into in the classroom with your classes on
this, algorithms and so forth. But just one layer, like
(13:19):
sort of from a logic flow of how you
get to that point of short term
profitability and connection to the these price
for example. We're talking about price now as a driver of
loyalty or customer value and the long
term value for the business owners love it. And it is so
(13:40):
important. Goes back to my training as a math major
at the Massachusetts Institute of Technology.
Plus years ago I was on track to become an actuary.
You know the people who kind of set the insurance rates and they
know to not get too granular,
they're not going to say how old are you, Mark Slayton going to be when
(14:02):
you die. They know that while they can come up with the best
guess about that, it's probably wrong. They wouldn't want to be held accountable for
it. So instead they're going to say people who share the
same characteristics as you, the same relevant
behavioral characteristics as you. What percent of that
micro segment will live to be 85 years old? It makes all
(14:24):
the sense in the world. Right? Again, you don't want insurance policies
doing the one to one thing. Again, not only would it be kind of creepy
and unethical and perhaps illegal, it would be ineffective
because the variability in our estimate of any
one person is huge. So I build the
world's best, best lifetime value models, but I know my
(14:47):
limitations. Now. Yes, I can tell you what you're worth, but I don't
necessarily want to be held accountable for it nearly as much as I'd rather make
a statement about the micro segment of the Mark Slaytons. What
are they collectively going to be worth and what kinds of policies should we develop
for them? So I'm not into the one to one thing. And it's
not only because of the ethical, legal, all that sort of thing.
(15:09):
It's because statistically it's really, really hard to justify
that. I do want to get granular, but not too granular
because we know the limits of the statistics. Yeah,
that's a beautiful way of describing it. And maybe you've answered,
you tell me, answered the next question which is connected to that, which is
I Think offline. You described this to me as the third rail, but is
(15:32):
this personal? Is it personalized pricing? What is the danger of
that? Are we really talking about getting too granular? So
the big danger is. I don't know what the biggest danger is. There's a
lot of big dangers. One is just the kind of PR backlash.
So no one remembers this anymore when Amazon, which does a lot
with dynamic pricing, and we got to really clarify the distinction, but
(15:55):
sometime around 2000, they ran one little bitty experiment where they got
into personalized pricing, where you and I would see different prices for the same
book at the same time. Credible backlash. You know, the thing went viral
and they backed away from it. And as far as I know, they've never tried
it again. So people just don't like the idea
that you and I, at the same time for the same item could be paying
(16:17):
different prices. Dynamic pricing, fine. Dynamic pricing. Saying
that, you know you're paying a different price than the person next to
you on the airplane, but that's because she bought the ticket two weeks earlier
and so she got a better price. That's okay. You would have gotten that price
too if you had bought it back then. Dynamic pricing, cool.
Again, we don't want it to be too granular where we're changing the price every
(16:39):
second because again, we start getting in the weeds, but kind of
gradually, occasionally changing. Dynamic pricing, I'm
a big advocate of that. The other problem with the personalized pricing, and
especially when you start bringing the AI in, is that not only will
it get too granular, but the factors upon which it
might make those decisions might start to get into issues of
(17:01):
legality because we can't discriminate on the base of
gender or race and other kinds of personal characteristics of.
I'm not a lawyer, don't push me on it.
But once we start bringing in the AI and we start looking at
people's media habits or who they're connected to in the social
network or what other kinds of products they've purchased. In
(17:23):
theory, that stuff is kosher, but it does start
getting real close to that third rail. And if we find
that certain demographic groups tend to do more
things than others, then we could find ourselves in a whole lot of trouble.
First degree price discrimination is just a very
dangerous thing to be working with, both because of the perceptions and because of the
(17:46):
legalities. And to me, the statistical issues.
Love it. So we talked also a little bit offline about the
possibility of other industries adopting this.
And what do you see as maybe those
Industries that could be ripe for falling into this trap. And what
advice would you say to give them? Well, for one thing is
(18:10):
move slowly. That's really important. Again, we want to dip our toes in the
water with dynamic pricing. Like I said, I had the privilege
of working with Major League Baseball and it was really fun to work with a
number of clubs where I think I won't name names, but one club
said, you know what, we'll just try doing this one section in the stadium. Just
let us just try that. Other clubs would say, well, we'll just do a couple
(18:31):
of games and even then we'll just have kind of maybe two
tiers of prices. So it was starting slowly, it was
experimenting, it was looking at the results very, very
objectively and trying to do so. Not just which one made
us more money over the time that we did that campaign, but the long
run implications. Did it chase season ticket holders away?
(18:54):
Did it get people to engage more and spend more
on things in the, in the store, at the concession
stands? So having again, a broader long term
perspective on it, it's such a tempting drug. I
mean that article about Delta gained so much attention.
Again, some people saying dangerous, other people saying opportunity, that
(19:16):
with great power comes great responsibility. And we just
need to look slowly, carefully, thoughtfully look. The fact that
Delta said they're doing it just for 20% of, of
folks, that's great, that's terrific. That's a good sign. And again,
I think that the, the changes that they're talking about might not be
nearly as dramatic and concerning as what some of the original
(19:38):
press suggested. So, so they might be okay here. And for the most
part, Delta is not only a wonderful airline, but they're pretty
smart about things. So we might be jumping to
conclusions there, but it, but it is an interesting cautionary
tale for other firms that aren't as smart, don't, that don't
have quite the, the analytics sa that don't have the experience running
(20:00):
these kinds of experiments. Take it slowly, but at the same time
try it. Don't be afraid to try it. A lot of companies
say we're just never going to do that, never going to do that. We don't
want to even take that first baby step because it's a slippery slope. That's a
mistake as well. I think that thoughtful,
dynamic pricing is actually, I think not only
(20:20):
smart. I think it's getting to the point where it's going to be kind of
an obligation for companies to be doing it. Maybe disclosing
some of the limits of what they do,
and maybe even being held accountable by their investors and other
stakeholders for doing it appropriately. Yeah. One of the things
that comes to mind is you're sharing and I think about theta and if
(20:43):
I'm listening and I'm a business owner, so I think in Major League Baseball and
some of the larger organizations that I'm sure you guys do work with, but
there's a long tail of small, medium sized
businesses that really drive this country. And, and I'm sure
they're listening closely about. Well, you know,
part of the question is, well, I don't nearly have enough data to
(21:05):
make those kind of decisions. Is that true? Well, it's kind of interesting
and almost ironic. If you're really a teeny tiny
business and you're just selling, you know, bespoke handmade
crafts, then you are doing, I don't know, it
sounds so evil first degree price discrimination. You're
going to figure out what that person's willing to pay and you're going to charge
(21:26):
them appropriately. You don't necessarily have any list prices. When you're a teeny
tiny company and you're figuring it out, there will be a bit of
negotiation, there will be a bit of flexibility. You're just
trying to figure out what this stuff is worth and what people are willing to
pay for it. The issue starts as you start to scale the
business and instead of leaving those pricing decisions in the hands of the
(21:48):
founder and as you start having, you know, multiple stores, you're operating
in multiple business units. That's when you want to start to
standardize things. That's when you just want to start bringing in the algorithms.
And at first, when you have this sense that you don't have a lot of
data, let's just figure out what the price is or the pricing policy. Let's lock
it in and stay with it. We have bigger, more important things to worry about.
(22:09):
But at some point you kind of owe it to yourself to start saying, maybe
we could get a little bit more sophisticated with the pricing.
So yeah, a big part of it is data is having granular
customer level data to tag and track to know not
just how responsive the customer is to our
price changes, but how differentially responsive different
(22:31):
customers are and tying those differential price
responses back to their lifetime value. If we find out
that our best, best customers are the ones who are most price sensitive, which
by the way, usually isn't the case, then we got to be really careful about
that. Maybe we'll use just a much simpler, clearer pricing policy,
but if we find out that our best customers are relatively price
(22:53):
insensitive because they know we're going to give them a good deal. They know we're
going to treat them well. They know it's much more about than price. It's service,
selection, quality, convenience. Then maybe we have a little bit more wiggle room there.
So it's really important to understand how the customers differ
from each other in pricing, price sensitivity, in
value, and just broader engagement. Great advice. I
(23:15):
learn so much every time we are together.
And thank you so much. I want to end this conversation with the
same question I ask all my guests, which is what delights you as a customer?
What delights me as a customer is service that's. And it's not
just necessarily rolling out the red carpet. It's not just kind of giving me
free stuff that I might have paid for otherwise. I think that often is a
(23:37):
mistake. Is incredible insight that goes
beyond the kind of training play. But I'm going to tell you a story here.
My favorite customer service experience. I have a thing and
maybe this makes me a bad person. I don't know. I like Ferragamo
shoes. They're just great. And so I had a pair of Ferragamos.
I kind of wore them out. And I went to the Ferragamo store in fifth
(23:58):
Avenue, New York. I walked in there, I said, I need a new pair of
shoes. I mean, literally standing at the door and this, this woman looks down
at my feet and she says, you are wearing an 11D.
You need a 10 and a half. E. Oh, my gosh. Bought two pairs of
shoes. Wow. It's that kind of thing that, I mean, it
is surprise and delight as we often talk about it. Yeah. But it's
(24:20):
unscripted surprise and delight. And that just shows that
this person knows her stuff. This company knows its stuff.
That's amazing. Love it. And you know your stuff. And I thank you so
much for returning back as a guest on the delighted customers podcast.
Mark, anytime. I'm happy to talk to you and I hope that
people heed the message.
Popular Podcasts
CrimeLess: Hillbilly Heist
It’s 1996 in rural North Carolina, and an oddball crew makes history when they pull off America’s third largest cash heist. But it’s all downhill from there. Join host Johnny Knoxville as he unspools a wild and woolly tale about a group of regular ‘ol folks who risked it all for a chance at a better life. CrimeLess: Hillbilly Heist answers the question: what would you do with 17.3 million dollars? The answer includes diamond rings, mansions, velvet Elvis paintings, plus a run for the border, murder-for-hire-plots, and FBI busts.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.