00:00:01
we're here today to discuss some new Wharton research about customer lifetime
00:00:04
value and company valuation here to talk with us is wharton marketing professor peter fader Pete thanks for being here
00:00:10
always a pleasure Rachel and Wharton doctoral student Dan McCarthy Dan welcome back yeah thank you very much
00:00:16
for having me so Pete first of all could you give us a quick summary of a research now this is a bit of a follow
00:00:20
up from something you've done previously where you looked at company evaluation
00:00:24
of companies that have customers that buy subscriptions and this is a little bit different so could you tell us a
00:00:30
little about that old research and then how you're building upon it in this way
00:00:32
sure well there's no such thing as old research here this is all brand new there's this idea of customer based
00:00:38
corporate valuation it's actually a concept that's been floating around but
00:00:43
Dan and I are taking it very very seriously and we want to bring just just a tremendous amount of rigor and breadth
00:00:48
to it in the previous work we looked up euerle on the contractual or the subscription-based side so if a company
00:00:56
knows when a customer's leaving it's kind of easy to put to project what the
00:01:02
rest of their life is gonna be and and the payments tend to be kind of steady but most businesses are what we call
00:01:09
non-contractual which is the you're just buying things on occasion and then there's a long hiatus between them and
00:01:16
and if you think about this is the kind of behave you have with with a retailer or with a travel firm or even
00:01:23
pharmaceuticals or media consumption most businesses are non contractual and that makes it much more difficult to
00:01:31
understand who's doing what and to project it into the future so a big part of what we're doing here is trying to
00:01:37
take this this broader concept of customer based corporate valuation but make it just as palatable in the
00:01:44
non-contractual setting as it has been in the contractual subscription-based one and and everything add yeah that's a
00:01:51
it's exactly right I'd say one of the key challenges in this work is really
00:01:55
the the data that we need to be able to perform this estimation and then project
00:01:59
forward you know how many purchases customers are gonna make and how much they're gonna spend on each of those
00:02:04
purchases so again we want these methodologies to be as broadly applicable as possible the vantage point
00:02:10
of this work is really someone who's on the outside looking in and so maybe it's
00:02:14
a shareholder a hedge fund private equity firm any number of financial institutions that are kind of everyday
00:02:21
looking at the public markets and trying to say so I want to buy this company do
00:02:25
I want to sell this company and those those those people they don't have access to the internal transaction logs
00:02:31
to the company and what they have access to are usually the sort of information that a company would put and its
00:02:37
quarterly or its annual filings and so basically we went at this problem by saying imagine that a company was to
00:02:44
disclose a certain very small set of common customer metrics what are the ones that would really enable us to be
00:02:51
able to perform this first estimation procedure for the various model components and use those to project
00:02:57
forward eventually what total revenues are going to be and Dan tell me a little bit now
00:03:02
once you've how did you figure out what those were and then what did you do with
00:03:06
them yes we figured it out by doing a large-scale simulation analysis so kind of the step one was saying this is a set
00:03:14
of metrics that we think could be usable and really they needed to kind of satisfy two main criteria the first is
00:03:21
that they're actually used by companies so we found many metrics that essentially we have found public
00:03:26
companies that are using them in disclosing them regularly and then the second criteria is really that they
00:03:32
actually help inform our model and so the way that we answered that was we said imagine that I was to generate data
00:03:39
from all these different possible types of worlds basically any different type of non
00:03:44
contractual company and just let these metrics kind of Duke it out so we took every possible collection of this set
00:03:51
and said imagine that I only got to see those metrics let me try to predict the future and then essentially see which
00:03:58
combinations of metrics kind of percolate to the top absolutely I think it makes sense to talk about the metrics
00:04:04
themselves so as Dan said we've done a lot of scraping of financial statements
00:04:10
listening to CEO conference calls seeing what third-party firms are saying about
00:04:16
the size and age of customer bases and we came up with six metrics so metric number one the most common one that a
00:04:23
lot of companies will report is what we call active users so how many people have made
00:04:28
transaction used our product or service sometime within the trailing 12 months number two is what we call heavy active
00:04:35
users so how many people have made a repeat purchased have engaged with us at least twice over that trailing 12 months
00:04:42
number three would be this idea of a forward repeat rate of all the people who made a transaction with us back in
00:04:50
2015 how many came back and did it again sometime in 2016 the the fourth and fifth would be kind of the flip side of
00:04:58
that so of all the purchases that we had today what percent of them are from customers who did something with us in
00:05:07
the previous year and that one we can either do on a customer by customer basis of all the customers who bought
00:05:12
with us what percent were with us previously or all the orders that were placed with us this year what percent of
00:05:18
them are by customers who bought previously and then last and least reported is the idea of frequency of all
00:05:25
the customers who have done anything with us in the past year how many things did they do
00:05:30
how many purchases did they make or so on so that's the big six and as Dan said
00:05:34
we went into it without a really strong sense of which one or ones would win out
00:05:39
but it was really interesting to do this grand bake-off across lots of different
00:05:43
worlds as you said and to see there actually were some very strong consistent results so when you did when
00:05:49
you did this bake-off what was the winning what was the winning metric here what was what are some of the
00:05:54
conclusions that you were able to make from this sure of course it's always easy in hindsight and I'm kind of
00:06:01
stalling right now because I want our listeners and viewers to think that among the six metrics that I just listed
00:06:07
if you had to pick say two of them which ones would bubble up to the top and would give you the ability as Dan said
00:06:15
to to kind of uncover the repeat buying pattern as if you had the granular data at your fingertips
00:06:22
well it turns out that far and away the number one metric is the least commonly reported one that I mentioned which is
00:06:28
frequency so if you can tell us among the people who did anything how many things did they do on average that now
00:06:37
by itself any one metric isn't going to get you far but if you take that frequency metric and you come
00:06:42
with any of the other five it doesn't even matter frequency plus anything gives you really good predictive
00:06:48
capability just winning by a nose among the five was the first one I mentioned just active users so if you can say
00:06:56
here's the number of people who have done anything with with us and here's
00:07:00
the average number of things that they've done you take that one-two punch and it's remarkable how well you can
00:07:07
uncover future purchasing and then benefit from all the the the other good things that you can do with those
00:07:13
insights yeah I think that really sums it up I'd say one thing that was a bit
00:07:19
surprising was the fact that when we went from two metrics to three there was virtually no improvement in our ability
00:07:25
to predict future purchases that really the bulk of the the benefit that we got was by moving from kind of the best
00:07:32
single metric to the best pair and what was almost equally interesting was you can't have too much of a good thing so
00:07:39
as we kind of kept that process going you know what's the best quartet what's
00:07:43
the best combination of five that we went when we went from the best combination of five to the best
00:07:48
combination of six performance actually got slightly worse and it got a bit more
00:07:54
variable from scenario to scenario and so actually you know sometimes less is better yes I think it was kind of an
00:08:00
interesting example of Occam's razor with the data so Dan tell me if now I'm
00:08:06
a company or I'm a customer I'm anyone that might be interested in this sort of
00:08:10
data what's the best way to take your findings and apply them in the real world really I would just demand that
00:08:17
companies disclose these metrics I think emphasizing both to the company itself and perhaps even to regulators whose
00:08:25
responsibility it is to ensure that investors have the information that they need to you kind of make informed
00:08:31
investment decisions we're basically saying that these metrics are really informative so I think it would be
00:08:37
wonderful to just kind of speak up and get people to disclose the metrics on the back end we have this you know
00:08:44
fairly complex estimation technique that's needed to really kind of take that scattered bunch of customer metrics
00:08:51
and be able to kind of map it down to the underlying parameters of these models that we have for how
00:08:55
customers are acquired over time how many purchases they make and the spend associated with each of the purchases so
00:09:02
there is some you know some math involved and some have heavier duty computation than we saw you say in the
00:09:07
subscription-based paper but it is certainly quite doable and I think as these method methodologies kind of take
00:09:13
hold we'll see them you know be kind of made more readily available do you have
00:09:19
anything to add it also I guess I'm curious like what do companies have reasons not to release this data I mean
00:09:24
is there a reason why they wouldn't want to and how do you get over that so first
00:09:28
let's talk about some of the other benefits then we could talk about some of some of those strategic aspects as
00:09:32
well so obviously the motivation here is customer based corporate valuation so can the investor make more informed
00:09:39
decisions about the current and future health of a company's customer base and
00:09:44
therefore the value of the enterprise but it can be useful in other ways as well so this could be a great source of
00:09:49
competitive intelligence so given that these metrics are not that hard to get say from a third-party firm or from a
00:09:57
company's first-party disclosures you can start to see of rival companies playing this game about other
00:10:04
competitors in the same sector to understand where they stand not just in terms of overall sales but in terms of
00:10:10
the nature of the customer base so so we think that there'll be quite an imperative to be reporting these things
00:10:17
not only out of some kind of fiduciary responsibility but for some of these kind of competitive activities as well
00:10:23
which goes directly to the other part of your question this idea of should there
00:10:27
be some sense of when you disclose which kinds of metrics right now it seems kind
00:10:34
of haphazard there are some companies that disclose some metrics all the time others that do it never there seems to
00:10:40
be no logic as to why certain companies do or don't and which ones they disclose
00:10:45
there's some companies out there that just just boast about disclosing metrics
00:10:50
may because it makes them look a little bit more rigorous or sciency but no one ever looks at these things and can
00:10:55
really make heads or tails out of them or map them back to actual or projected future revenues so we want to see much
00:11:02
more Smart's about who discloses what when and and maybe companies would then start
00:11:10
to say so are they revealing these metrics now because that they're trying to signal
00:11:16
something or or maybe they're gonna stop revealing them because they're trying to
00:11:19
hide something so there's gonna be this whole chess match about who discloses
00:11:23
what when that'll be kind of interesting when we get to that point right now it's
00:11:27
just so early on that if companies are doing this is all it really seems that gets more kind of the ceo's ego as much
00:11:34
as any kind of real information value we want to see just again much more a rigor
00:11:38
and discipline about who's disclosing what when and why do you feel like this
00:11:44
is gonna become people are gonna become more interested in doing this and more interesting talking about it as their
00:11:49
becomes more of this realization that maybe you're not saying everything there
00:11:52
is to say about a company by just disclosing its financials that there is this other element there that may have
00:11:57
nothing to do with earnings or revenue at least not for that quarter that that's right so I so for many companies
00:12:04
that they they go to Wall Street on a quarterly basis and Wall Street's kind
00:12:08
of look at them saying your earnings aren't up to snuff but they say you know
00:12:11
what we're investing in the customers well so far it's just been a trust me
00:12:16
but now we can provide definitive proof so now investors can say is that if that's really true
00:12:21
then show us these two metrics so that we can make that assessment for ourselves so I want that the companies
00:12:27
to be more forthright about it I want investors to demand it I want competitors to be kind of curious
00:12:32
regulators as Dan said I just want just a more active conversation about this this concept of customer based corporate
00:12:40
valuation I think it's in everyone's best interest and of course I'm a
00:12:44
marketing professor I think that if we can establish that there's something here then it will also trickle down
00:12:49
through the rest of the organization as well and start to impact other kinds of decisions that the company is going to
00:12:54
be making I say the one other thing I would add to that it's a very legitimate
00:12:58
concern that some companies would have that as soon as they start disclosing these metrics that you know there's some
00:13:04
sort of negative you know repercussion from that and I think there's been some
00:13:07
very interesting recent work that was just put out and one of the top marketing journals by a colleague of
00:13:12
ours Bern Ciara that basically came to the conclusion that when companies disclose forward-looking metric
00:13:19
it really helps investors make more accurate projections of revenues thus lowering their kind of perception of the
00:13:25
uncertainty about what's what's to come in the future and that actually can help
00:13:29
increase the valuation of firms and so all else being equal disclosing forward-looking metrics can actually be
00:13:37
beneficial to the stock price so it can help just kind of overcome some of those
00:13:41
concerns that companies might have so dan do you feel like that there are other misperceptions that might this
00:13:46
research might dispel I've I'd say yeah for one just the competitive concerns
00:13:51
you know we're showing that these metrics can be actually very helpful for projecting what will happen in the
00:13:57
future I'd mentioned the one about you kind of too much of a good thing but it's a another kind of spin on that
00:14:03
would be just the overwhelming kind of influence of quality over quantity that it's an extremely important it's
00:14:12
extremely important to have the right combination of metrics and oftentimes having a very small set of very good
00:14:19
metrics can be much more beneficial than having kind of the wrong set of very numerous metrics so in that large-scale
00:14:27
simulation analysis we actually found that if you gave us the right set of two metrics it could do much better than
00:14:33
another set of five metrics so that kind of surprised me in terms of misconceptions I'd say the final one
00:14:42
that comes to mind is up until this point most people have been using customer metrics as kind of a dashboard
00:14:48
that every month or every week the company will have a spreadsheet that sent out to the organization and it's
00:14:54
got this long list of all these different customer metrics and how each of them has gotten better or worse over
00:14:59
time and usually people are looking at each of these different metrics is kind of you know in its own world yeah that
00:15:06
essentially if it went up it must be good if it went down it must be bad and let's figure out why here we're kind of
00:15:11
saying we don't just need to look at these metrics kind of on a standalone basis they're not ends unto themselves
00:15:17
essentially we can be able to kind of tie them up and summarize the you're kind of effective you know what happened
00:15:24
over the past week or month on the overall valuation of the firm by kind of putting them all together and weaving
00:15:29
them into this integrated model now Pete I guess anything from you but also do you feel like cus the companies
00:15:35
that don't operate on a subscription model do you feel like there was the percentage out there that maybe they
00:15:39
couldn't do something like this that that was just for the subscription-based
00:15:43
businesses of the world it's not for us do you think this dispels that if it
00:15:47
well it does dispel that but actually was a worse problem that there were a lot of these kinds of non-contractual
00:15:53
non-subscription firms out there and they would just kind of make metrics up to make it seem like they could use some
00:16:00
of these subscription oriented approaches so for instance in the subscription world you have a formal
00:16:05
retention rate you know of all the people who had a subscription with this last period how many of them renewed you
00:16:12
know whether they did it or not there is no equivalent metric in the non-contractual world there's no way
00:16:17
that you that you kind of actively raise your hand or a worse yet if you're not
00:16:22
present if you don't make a purchase if I didn't buy anything from Amazon in the
00:16:26
past twelve months doesn't mean that I'm gone as a customer it just means that
00:16:30
I'm a light buyer so what a lot of non-contractual firms were doing is they would take some of these metrics like
00:16:36
the one I mentioned before the for word retention from the for word repeat rate among all people bought in 2015 how many
00:16:42
also bought in 2016 this you know what that's kind of like a retention rate let's just treat it that way and then
00:16:49
pretend that we're a subscription company well that's just wrong and and
00:16:53
not only is it conceptually incorrect but if you were to follow it all the way through and some of the earlier
00:16:59
published research showed this that your ability to assess the overall value of the customer base is going to be way way
00:17:05
way off and the Diagnostics that you get around it would be off as well which takes me back to the other part of your
00:17:12
question one of the other benefits of doing all this is to create credibility for the marketing organization if we can
00:17:19
say here are the really meaningful metrics and here's the way that we can tie them together to make statements
00:17:24
about valuation instead of having these things being little trophies on a shelf where the rest of the organization say
00:17:30
the people in the CFO's office would say that's nice let the marketers play with
00:17:34
their toys if we can show the CFO and again other folks to the organization that these metrics can actually help
00:17:40
with day-to-day operational decisions as well as bonafide long strategy it's going to cast the entire
00:17:47
marketing organization under a very different light and to me personally getting that kind of respect getting the
00:17:54
kind of common language is more important than taking some of these valuation models and trying to find a
00:18:00
little bargain here or there as a hedge fund might do they're both interesting
00:18:05
they're both important but for me it's that the common language and respect for
00:18:08
marketing that that's number one now P can you tell me what's next for this research where you go where are you
00:18:13
gonna go from here well right now we're just establishing these the basic methods like here are the metrics you'd
00:18:19
want to use here's the way you tie them together to get the overall valuation
00:18:23
but it starts to lead to all kinds of other questions including some that we've already touched on this idea of so
00:18:28
strategically what kinds of metrics should I disclose when but I really should defer to Dan McCarthy because
00:18:34
this is his dissertation work and he's been thinking of a lot of other different directions to take it that
00:18:40
will be part of what he does when he graduates from here and becomes a professor down at Emory so so Dan you
00:18:46
tell us what's next yes in terms of the boxes that have already been checked we
00:18:51
kind of think of it along a few different dimensions you know what's the sort of data that we have you who's
00:18:56
performing the exercise what is the goal and what's the type of firm and right
00:19:01
now we've primarily focused on external stakeholders you know like the shareholders or the private equity firms
00:19:07
or the hedge funds who are their goal is really to make an accurate estimate how
00:19:12
much the value the firm should be and we've done that for both contractual and
00:19:16
for non contractual firms so in terms of where we go from here I'd say one of the
00:19:20
big questions is can we move beyond measurement imagine that I'm inside the company is there anything else that I
00:19:26
might want to do so maybe we can be able to kind of manage value not just measure
00:19:31
it and kind of part and parcel with that we'll be using internal data as opposed
00:19:37
to external data instead be one example another is in both the subscription based valuation paper and a non
00:19:44
subscription valuation paper we were kind of only looking at one company at a time yeah so whether it's Dish Network
00:19:50
Sirius XM or this apparel retailer and in the current paper we didn't really
00:19:56
focus on any any come Pettit --iv effects or anything like that so I think an interesting area of
00:20:01
future work would be imagine that we had multiple companies in the same industry
00:20:05
could we learn something you know that we wouldn't have learned about if we were looking at each of the companies
00:20:11
individually make better predictions or just learn about competitive effects so I think the sky's the limit
00:20:17
you know we're just really scratching the surface and I'm very excited to see
00:20:21
where we can go from here dan thanks so much for being here with us again always
00:20:25
great to talk with you Rachel thanks yeah thank you very much for having us [Music]