Notes

Here is what I suspect most of you did.
You probably took a weighted average.
You said, look we understand the breakdown
of our customer base,
70%, 20%, 10% into the low, medium, and
high risk groups.
And then, they have their associated
attrition rates.
And so, you took this simple weighted
average.
You multiply 0.7 by 0.06, and so on, and
so on, and
so on.
You did the calculation and, as it says
right
here, the overall average attrition rate
would be 17.7%.
And, what does that average attrition rate
tell
us about the overall length of the
customers life?
And therefore, the overall financial value
of the customer base.
Well once again, 1 over 17.7%, do that
calculation, you get 5.6 years.
Well, that sounds kind of familiar, right?
Isn't that just about the same number,
that we got, when we didn't celebrate
heterogeneity.
So, what difference does it make?
Why is taking heterogeneity into account,
going to help us at all here?
I'll tell you why, because you did the
wrong calculation.
You did calculate an overall average but,
like,
we've said before, there is no average
customer.
Telling me that the average attrition rate
is 17.7%, doesn't do me any good.
Because, there is no customer out there,
at least, according
to the Vodafone analysis, who has that
kind of attrition rate.
So, we've just calculated an expected
lifetime, an expected
financial value, for a customer who does
not exist.
So, the question is, what's wrong with
this calculation?
How do we do it the right way?
How do we, truly celebrate heterogeneity?
because what we did, here, is we
eliminated
heterogeneity, we just squashed it all
together and said...
heterogeneity is gone, and that's why we
ended up getting the same result.
That we would have gotten, if we didn't
look at heterogeneity
[LAUGH]
in the first place.
So, what's wrong with this calculation,
and what's the right way to do it.
So, think about that for a moment.
And then, I'll give you a super big hint
from one
of the world's leading thinkers about
customer loyalty, and so on.
So, let's turn to the words of Frederick
Reichheld, who, a
number of years ago, wrote a book called,
The Loyalty Effect.
And, he laid out so many, many good
ideas about what loyalty is, how we
measure it, how we capture it.
He's a consultant for Bain, so, so not an
academic.
So, he's really seen loyalty in action,
and has
helped a lot of companies create and
monetize it.
And, here's the quote from Reicheld, and
you can, you can see
it here, and I hope the logic makes a lot
of sense.
The average makes no sense at all.
We need to do the calculations
separately, group by group.
Okay, so, that's where the celebration of
heterogeneity is going to come in.
Let's understand the separate value for
the high, medium, and low,
and then combine them together, instead of
combining them together first.
I hope that makes sense.
Perhaps, your inclination is to say, seems
kind of similar.
You know, let's take an average, and then,
calculate the lifetime.
Or, let's calculate a lifetime,
then take an average, seems kind of
similar.
Well, you know what?
It's not.
So, let's revisit this example, but do the
lifetime calculation, first.
Okay, so for our low risk customers, their
average attrition rate is 0.06.
So, what does that mean, just for those
customers,
how long are they going to stick around,
on average?
So what's 1 over .06?
Well, as you can see here, its about 16.7
years.
And, if we repeat that calculation for the
other two
groups, we can see what their expected
life time would be.
And here, you see the, really, dramatic
differences.
This is the celebration of heterogeneity.
We see an, order of magnitude difference
between the best and the worst customers.
We don't want to ignore that, we don't
want to eliminate that,
we don't want to average over that, we
want to celebrate that.
Once we see the expected life times,
for each of these different customer
groups.
Then, let's take the weighted average.
Then, we'll multiply by 70, 20, and 10%,
and when we do that, what do we get?
Our overall expected lifetime for this
customer base, 12.4 years.
That's a big difference, I'm sure you'll
agree.
So, instead of doing the calculation the
wrong
way, where we, where we weren't
celebrating heterogeneity.
Once, we acknowledge and, and explicitly
take into account heterogeneity.
We have, more than doubled, the value of
our customer base.
Just like that.
Not by doing anything, just by doing the
correct calculation.
This is the celebration of heterogeneity.
Now, you might be wondering, alright, so,
this is this Vodafone example.
How does this work in general?
Answer?
Always works the same way.
Whenever there's any heterogeneity at all,
any spread
among the customers, in terms of their
attrition rates.
There will always be, money left on the
table, if we ignore heterogeneity.
We will always understate the value of the
customer base, by ignoring heterogeneity.
The only question is, how much?
So, in this case, it's a, more than,
twofold increase.
And, as you might imagine
the, the magnitude of that increase
depends entirely
on, the magnitude, and nature of the
heterogeneity.
So, if we, if the customers are more
spread out.
Then, ignoring heterogeneity is going to
cause us
to even steeper, understatement of the
customer value.
So, again, it's not a question of whether
it will occur.
It's not a question of whether it will be
an overstatement, or an understatement.
It always
works this way.
It's just a question of how much.
Now, if that's not a celebration of
heterogeneity, then, I don't know what it
is.
But, this shows you, that by explicitly
accounting for heterogeneity.
By really understanding the differences
among our
good customers, and our not so good
customers.
We have just doubled the value in our
company.
Now, you might wonder, so, what are the
implications of that, for managing our
customer base, and
this is really, really important.
Especially, for all of you who voted for
customer retention, as
the activity that we should be putting our
incremental dollars on.
So, let's summarize our thinking about
customer retention, and there, there's
two really important points that I want to
make, over here.
First and foremost, there is no "average"
customer, and
you can't do calculations based on an
average customer.
Yes, it's easy, yes, it's convenient, but
its wrong.
You will understate the value of your
customer base, and
that difference can be huge, if you ignore
the heterogeneity.
And here's a second point, that's much
more subtle,
and I'll spare you the painful math to get
there.
But, if we ever want to calculate an
elasticity, if we
ever want to find out, what's the
incremental gain that we get.
For say, a 1%
reduction in the attrition rate.
It turns out, we can do that calculation
two ways.
If we do it by ignoring heterogeneity, as
many firms do.
Or, if we do it by taking heterogeneity
into account,
as we just did, we get a very different
conclusion.
So, in some of my research, we've actually
done that calculation separately.
And, we show that the, the, the retention
elasticity, the gain that we get by
lowering the attrition rate by
one percent, is much, much less.
When you account for heterogeneity.
Which means that, efforts to boost
retention, or decrease attrition,
or churn are, are much more modest, then
you think they are.
Once you've exclusively account for
heterogeneity.
Now, I want to
be, really, clear about this.
I am not saying that you should stop
spending
on retention, no, I am not saying that at
all.
Okay, retention is one of our three major
pillars of customer centricity.
It is very, very important.
You want to figure out who the good
customers are and make,
and do whatever it takes to keep them
around, for a long time.
You must do that, but at the margin,
companies seem
to want to spend more, and more, and more
on retention.
That seems to be the constraint
for them.
I contend, that some companies, might,
actually, be overspending on retention.
And, taking some of those retention
dollars, which are
often being allocated to customers who
aren't really that great.
To customers, who have a fairly high
attrition rate,
and are always going to have a high
attrition rate.
And, even if we kind of bribe them, or
incent them, to stay around for another
year or two.
They're going to leave at the first
opportunity, after that.
We're better off takin' some of those
dollars and spending them where?
Customer acquisition.
Let's spend some of those dollars, finding
new customers, who might be really good.
So, so, right away you're starting to see
some of the implications.
Some of the trade-offs between,
acquisition and retention, that
arise, when we have an explicit focus on
heterogeneity.
So, that's the retention story.
Let's take a short break, and come back,
and talk about customer development.
[MUSIC]