Chapter 3.

5 : Calculating the lifetime value of a customer

Chapter 3.5

Calculating the lifetime value of

a customer
This chapter includes:

J What is customer lifetime value?

J How long is lifetime?
J The factors affecting customer lifetime value
J How we use LTV analysis
J Two ways to calculate customer lifetime value
J Using LTV analysis to compare the effectiveness of various
marketing strategies

About this chapter

I
n this chapter we will deal with one of the most important measures in
direct marketing – customer lifetime value.

We will look at what we mean by it, how it is calculated and how management can
use it for decision making.

We will further look at different ways it can be calculated, both historic and
projected with examples.

Next we will look at discounted cash flow and net present value with several
worked examples.

Author/Consultant: Brian Thomas 3.5 – 1

 Chapter 3.5 : Calculating the lifetime value of a customer

Brian Thomas F IDM

thomas_b@btconnect.com After a sabbatical Brian became Chairman and
Managing Director of Saatchi and Saatchi Direct
Brian has been in in 1988, continuing in this role until 1991 when
marketing and he left to become an independent Direct
management for almost Marketing Consultant.
40 years. He held senior
positions with GUS, ICI, Fine Art Developments Brian is one of the Course Directors for the UK
and Early Learning before switching to the agency residential courses for the IDM Diploma in Direct
side in the late 70s. Marketing. He also runs a number of public
courses and seminars for the IDM. Between 1981
He was Managing Director of THBW, helping this and 1998 he ran all the Direct Marketing Courses
agency grow into the largest independent direct and Seminars presented by the Chartered
marketing agency in Britain. THBW was merged Institute of Marketing.
into Ogilvy & Mather Direct (now Ogilvy One)
with Brian continuing as Managing Director. In 1999 Brian was twice honoured by the Institute
During the next three years he helped build O & of Direct Marketing, receiving their award of
MD into the largest direct marketing agency in Educator of theYear in June and being elected a
Europe. Fellow of the Institute in November.

Chapter 3.5

Calculating the lifetime value of

a customer
Introduction

A
lthough direct marketers do need to measure responses, orders, sales
revenues and profits in the short term, i.e. as each individual promotion or
campaign is completed, such measurements do not take into account one
of the most important aspects of direct marketing – the long-term quality of a new
customer.

Direct marketing is rarely a low-cost option and the upfront investment in setting
up a database or Customer Relationship Management (CRM) system can be very
substantial. It can take several years to achieve payback from such an investment.
Furthermore, measuring the revenue from a campaign against the costs incurred
in that campaign, while of course helping to ensure that we do not lose money in
the short term, is a very tactical way of establishing an acquisition budget. What is
needed is a method that can help us at a more strategic level by answering the
question:

“How much can I afford to invest in acquiring a new customer?”

If we try to answer this from a short-term perspective, i.e. by judging it based on

the profit from a single sale, there are two problems:

3.5 – 2
 Chapter 3.5 : Calculating the lifetime value of a customer

1. We become very focused on short-term Return on Investment (ROI) and thus

may not invest enough to grow our business as effectively as we might, and

2. We assume that each new customer is worth precisely the same to us and
we ignore differentials in quality, longevity and so on between differing types
of customer.

Clearly, we cannot answer the above question until we know the answer to this
one:

“What will a new customer be worth to us over the time (s)he continues to
buy from us?”

We answer this question by using ‘Lifetime Value Analysis’ and the remainder of
this chapter will explain precisely what this is, and how it works.

Before we start let us look at a real life example:

A UK charity was evaluating recruitment methods and its main measurement

criterion was cost per new donor.

One of the ways in which charities recruit new donors is by use of face-to-
face recruiters who approach people in high streets, shopping centres and
railway stations. This is usually a very low-cost medium.

The media being evaluated were Direct Response Television (DRTV), Door-to-
Door Distribution, Direct Mail and Face-to-Face recruiters.

The evaluation showed the following:

Method of recruitment Cost per new donor – index

DRTV 30
Door-to-Door 40
Direct Mail 38
Face-to-Face 10

On this basis Face-to-Face was far and away the cheapest with Direct Mail
and Door-to-Door costing almost four times as much.

If they had used this evaluation alone to decide on acquisition strategy they
would have made a major mistake – the key to successful donor recruitment
is quality – how long will new donors continue to support the charity and
how receptive will they be to future fundraising requests?

The following chart shows the attrition rate by each recruitment medium –
attrition rate means how quickly they stop supporting the charity (by
cancelling their direct debit mandate).

3.5 – 3
 Chapter 3.5 : Calculating the lifetime value of a customer

Figure 3.5.1

$WWULWLRQE\5HFUXLWPHQW0HGLD

PRQWKV

 PRQWKV
UHPDLQLQJDFWLYH

 PRQWKV
 PRQWKV






)DFHWR 0DLOLQJV 'RRU 0DJD]LQH '579
IDFH WRFROG 'URSV ,QVHUWV
OLVWV

Although Face-to-Face recruits were the cheapest to obtain, we can see from this
chart that their future performance was very poor – after six months their
attrition rate was as high as that for Direct Mail recruits after 40 months.

A further study looked at the propensity to upgrade support according to the

recruitment channel and this too was very revealing:

Figure 3.5.2

8SJUDGHSURSHQVLW\E\PHGLDFKDQQHO










'RRU'URS 'LUHFW0DLO '579 )DFHWR)DFH

This study showed that Direct Mail recruits were most likely to increase their
level of support while Face-to-Face recruits were least likely.

Finally, the next figure compares ROI at recruitment stage with lifetime value
measured over five years:

3.5 – 4
 Chapter 3.5 : Calculating the lifetime value of a customer

Figure 3.5.3

0HGLDSHUIRUPDQFHDWUHFUXLWPHQW
DQG/79
52, /79






52,






'RRU'URS 'LUHFW0DLO '579 )DFHWR)DFH

This case demonstrates that the cost of recruiting a new donor is not really
relevant – long-term performance is the only true measure.

As this example shows, measuring at the wrong time (which usually means too
soon) can be seriously detrimental to business growth and profitability.

It is tempting to believe that simply understanding that we have to measure

customer quality as well as quantity is enough. However, if we do not take this to
its conclusion we will never know whether we are over or underinvesting in
customer acquisition, CRM and so on.

Careful use of lifetime value analysis enables us to:

G Plan and measure investment in customer acquisition programmes

G Compare the performance of customers recruited through different media

or offers

G Identify and compare different customer segments

G Measure the effectiveness of alternative customer retention strategies

G Establish the true value of a company’s customer base

G Make better informed decisions about products and offers

Using LTV analysis marketing managers can address the above issues in a
financially justified, long-term orientated way. LTV analysis demonstrates why
customer management should be right at the heart of a company’s marketing
activities.

Here are some examples showing the value of LTV analysis to businesses in
various fields:

3.5 – 5
 Chapter 3.5 : Calculating the lifetime value of a customer

Car buyers stay loyal to their marque

A major motor manufacturer wants to attract young adult customers to its
brand, but is uncertain how much it can afford to invest to induce a first
purchase. Studies of long-term car buying behaviour reveal that consumers
are very likely to repeat purchase within a brand. This year’s marketing
spending, therefore, may influence very large purchase decisions for
decades. This longer term perspective justifies larger investments in new
customer acquisition. And if this manufacturer does not take a long-term
perspective, it is very likely that one or more of its competitors will. In other
words, the manufacturer needs to estimate the long-term, or lifetime value of
each expensively acquired new buyer in order to know how much to invest in
recruiting him or her. Using these techniques, Volvo estimated that it costs
them only one-fifth as much to achieve a repeat purchase as it does to gain a
conquest sale. Conversely they can afford to spend five times as much to
achieve a first time purchase as a repeat one.

Charity spends money to acquire future donors

A leading UK charity found that the cost of recruiting new donors
significantly exceeded their first-year contributions. Analysis of its donor file
showed, however, that over a ten-year period the value of a new supporter –
in terms of donations, purchases and introductions to friends etc. –
amounted to many times the initial cost of recruitment. Thus it could afford
to invest in new donors. The question was by how much? Calculating the
lifetime value of each donor acquisition source provided the answer.

Mortgage customers do not become profitable for more

than ten years
A few years ago, a well-known UK bank published the fact that, on its most
generous ‘cashback’ mortgage deals (e.g. “Mortgage your new house through
us and we will give you £10,000 to spend on furnishings of your choice”) it
would not recover its acquisition cost until the eleventh year of the mortgage.

Without long-term LTV analysis they could not contemplate making such an
offer.

In each of the above cases marketers who looked only for short-term ROI would
make decisions that ignored long-term opportunities.

What is customer lifetime value?

It is the total value of all future contributions to profit and overheads we expect from that
customer during their ‘lifetime’, i.e. the period they remain a customer.
But future contributions to profit have to be considered in a special way. Why?
Because of the cost of money.

3.5 – 6
 Chapter 3.5 : Calculating the lifetime value of a customer

Let us imagine I lend my friend £100 and he unfortunately cannot afford to pay
me back for a whole year. In a year’s time he gives me my £100 and we are both
happy. Except that my generosity has cost me money. If instead of lending it to
him I had invested it in a savings account it would have earned interest and my
£100 would be worth, say, £105. So I am £5 down on the deal.

Of course when dealing with friends and relatives we would not normally think of
charging interest on a £100 loan, but, instead of lending my friend £100, what if I
had just persuaded my Financial Director to let me invest £100,000 in a new CRM
system?

The FD would naturally ask: “when will we recover our investment and start
showing a profit?”

I might reply: “I have calculated that this new system will enable me to gain
additional profits of £20,000 per year for the next five years, so you will have your
£100,000 back in full in five years.”

At this point my FD will explain to me about the cost of money and two concepts
that we are now going to examine:

G Discounted cash flow (or DCF), and

G Net present value (NPV)

My Financial Director will point out that if, instead of financing my new system,
he invested the £100,000 in an interest-bearing account it would appreciate
significantly in value over the five years.

In fact if it only achieved an interest rate of five per cent, in five years the
£100,000 would be worth more than £127,000. In other words, if I only pay back
£100,000 over the next five years the company will be £27,000 worse off. This is
the cost of providing that funding – the cost of money.

To compensate for this we use the process called discounted cash flow (DCF)
and discount each future payment. How much discount should we allow? The
simple answer is by whatever we think the interest rate will be over the period
in question. But this may not be enough. Let us examine some of the reasons
why we may want to discount at a higher rate:

G We may underestimate the interest rate – just a few years ago the UK
interest rate was well over 10 per cent, rising to 15 per cent for a brief
period. The longer the period of your projections the greater the
opportunity for error here.

G We may have been optimistic in projecting our sales targets – marketing

people are optimists at heart and this happens more often than you may
think.

G Our product may be superseded by a revolutionary development we have

not even dreamed of, so again we may fail to reach our sales targets.

G We may make some mistakes on the way, which will add to our costs or
further reduce our sales revenue.

For these and perhaps some other reasons you can think of, prudent direct
marketers increase the discount rate. The most prudent even double their
estimate of the interest rate. This may be over-pessimistic but you should at least
add around 50 per cent. So, if you think the interest rate over the period of your

3.5 – 7
 Chapter 3.5 : Calculating the lifetime value of a customer

projections will average 6.5 per cent, it would be prudent to work to a discount
rate of 10 per cent.

In real life, most company financial controllers will already have a DCF rate that
they have calculated to be appropriate for their market and trading conditions.
The simplest thing to do in your case therefore is to find out what DCF rate is
used in your company and apply that. This will give your purse-holders
confidence in your projections and you will probably also find that there is an
existing spreadsheet on your system that will do some of the calculations for you.

To demonstrate the basic process let us now work out just how much of the
£100,000 investment we would have paid back in today’s value (net present value)
if we produce an additional net profit of £20,000 each year for five years. We will
work with a discount rate of 10 per cent to allow for those factors mentioned
above.

The usual convention is that in period one we do not discount – for this
reason some companies prefer to work in seasons of six months or even
apply the discount quarterly. We will use periods of one year to keep it
simple.

There are several ways of applying the discount but the simplest method is to
deduct the percentage discount from one and multiply by the remainder. Thus to
discount by five per cent we would multiply the starting figure by 0.95; for 10 per
cent discount by 0.9; for 15 per cent discount by 0.85 and so on. As we are
working to 10 per cent in this example we multiply by 0.9 each time. The
discount compounds of course, so in year two (the first discounted year) we
multiply by 0.9; in year three (two discounted years) we multiply by 0.9 x 0.9
(0.81) and so on.

Table 3.5.1 Net present value of £20,000 profit per year over
five years
Year 1 Year 2 Year 3 Year 4 Year 5

Net profit in year £20,000 £20,000 £20,000 £20,000 £20,000

Discount factor 1 0.9 0.81 0.729 0.6561

Net present value £20,000 £18,000 £16,200 £14,580 £13,122

profit

Cumulative net £20,000 £38,000 £54,200 £68,780 £81,902

present value profit

As we can see in table 3.5.1, even if we do make an additional £20,000 profit each
year, its true worth at today’s values will be considerably less: at the end of the
fifth year we will still owe our accountant more than £18,000. At this rate it will
be almost two more years before we have paid back the investment.

This process is called discounted cash flow (DCF) and the discounted figures are
the net present value (NPV) of each £20,000 amount.

3.5 – 8
 Chapter 3.5 : Calculating the lifetime value of a customer

How long is ‘lifetime’?

There is no single answer to this question – it will vary according to several
factors:

G The type of products you sell – clearly there will be a longer lifetime
associated with a 20-year mortgage than a microwave oven. A product that
only appeals to parents of young children (e.g. disposable nappies) has a
defined customer life cycle that will only be extended if the customers have
more children.

G Service levels and delivery on promises made at first purchase – how

well you service your customers and how well your product lives up to the
promises you made about it.

G Market stability – in rapidly developing markets like IT, telecoms and so on

it can be difficult to make sound LTV projections. The high level of ‘churn’
in these and markets such as credit cards makes it very difficult to make
sound projections of the value of long-term customer relationships.

G Company objectives – financially strong, strategically orientated companies

have more confidence in taking a long-term view of customer lifetimes.

As we can see there is no norm – however, unless you are selling long-term
mortgages or life insurance, it is usual for companies to project LTV for periods of
between three and ten years. Five years would be a reasonable period when using
LTV analysis to answer that question posed a few pages ago: “How much can we
afford to pay to acquire a new customer?”

When dealing with products that are not normally replaced each year a
slightly different approach is required. With new cars or computers we might
find that the repurchase period is two or three years. So each period of
evaluation would be two or three years. It is important to realise that in a
three-year repurchase cycle, we still have to apply the discount factor for
each year. A car main dealer may look at LTV over a ten-year period
as follows:

3.5 – 9
 Chapter 3.5 : Calculating the lifetime value of a customer

Table 3.5.2
Year 1 Year 4 Year 7 Year 10

A Customers 500 350 262.5 210

B Retention 70% 75% 80% 85%

C Sales p.a. £20,000 £22,000 £25,000 £27,500

D Total Sales £10,000,000 £7,700,000 £6,562,500 £5,775,000

E Net Profit £1,500,000 £1,155,000 £984,375 £866,250

15%

F Discount Rate 1 0.729 0.531441 0.387

G NPV £1,500,000 £841,995 £523,137 £335,603

Contribution

H Cum NPV £1,500,000 £2,341,995 £2,865,132 £3,200,735

Contribution

I Lifetime Value £3,000.00 £4,683.99 £5,730.26 £6,401.47

at Net Present
Value

Note – table 3.5.2 looks at the predicted 10 year repurchase cycle of 500
customers who bought for the first time in year one. We can see from this table
how vitally important it is to discount future revenues. In year 10 an apparent
profit of £866,250 is reduced to only £335,603 if we discount it back to its net
present value. This is a very conservative view given that we have discounted by
10 per cent each year but it does give the dealer a clear view of how much can be
safely invested in achieving a first-time sale.

Factors affecting customer lifetime value

There are several factors that will affect the future buying performance of a
customer, including:

Method of recruitment
If we recruit customers through the use of heavy discounts or incentives their
future buying performance will not be as good. Our offer has raised certain
expectations and if subsequent offers do not match this, fewer orders will result.
Alternatively if we want to maintain order size and frequency we may need to
sacrifice profit margin in order to do so – either way our net profit will reduce.

Medium
As we saw earlier, respondents from certain media may cost more to acquire
but perform better in the long term than those recruited from cheaper media.
Similarly, some mailing lists will produce better quality customers than others.
Experienced direct marketers have noted with interest that customers who
make their first purchase as a result of a direct mailing tend to demonstrate
higher loyalty and thus greater lifetime value. We do not know for certain why
this is so but in the opinion of the author it is because a mailing tends to contain
3.5 – 10
 Chapter 3.5 : Calculating the lifetime value of a customer

much more information than an email or a TV commercial and thus tends to

generate a considered rather than an impulse purchase.

Among the best new customers will always be those who were introduced to us by
a friend – (through promotions we call MGM (Member-get-Member), Friend-get-a-
Friend, or referral schemes).

As we can see from the following chart (based on real-life case studies), this is the
only acquisition source likely to produce better quality customers than direct
mail.

Figure 3.5.4 Lifetime value by source of customer acquisitions











… 
'LUHFWPDLO $GYHUWLVLQJ 3DFNDJH )ULHQGJHWD
LQVHUW IULHQG

Geodemographic factors
For example, although as first-time buyers they may seem the same, owners of
large gardens are likely to buy garden items in greater quantity or more frequently
than those with smaller properties.

How we use LTV analysis

Planning customer acquisition investment strategies
As we have said, one of the key uses of LTV analysis is to decide how much you
can afford to invest in recruiting new customers. Many practitioners are prepared
to invest 30 per cent or even more of the lifetime value they predict for certain
groups of customers.

Experience will help you decide what is appropriate for your market, but
remember that if you have been prudent in setting the discount level and you have
cut off your predictions when there are still a healthy number of active customers
in the cell, you can invest with more confidence.

To elaborate on that, let us look in table 3.5.3 at a typical cell of 1,000 new
customers projected over five years and making allowance for some attrition
(customers ceasing to buy from us) each year. This is shown as Retention Rate,
3.5 – 11
 Chapter 3.5 : Calculating the lifetime value of a customer

i.e. with a retention rate of 60 per cent we have 40 per cent attrition. Note that
retention rate tends to increase the longer customers have been with us. Annual
sales per customer have been increased by around 3 per cent each year to allow
for inflation.

Table 3.5.3 Five-year projection

Year 1 Year 2 Year 3 Year 4 Year 5

A Customers 1000 600 390 273 204.8

B Retention 60% 65% 70% 75% 80%

C Sales p.a. £600 £620 £640 £660 £680

D Total sales £600,000 £372,000 £249,600 £180,180 £139,230

E Net profit £120,000 £74,400 £49,920 £36,036 £27,846

20%

F Discount rate 1 0.9 0.81 0.729 0.656

G NPV £120,000 £66,960 £40,435 £26,270 £18,270

contribution

H Cum NPV £120,000 £186,960 £227,395 £253,665 £271,935

contribution

I Lifetime value £120.00 £186.96 £227.40 £253.67 £271.94

at net present
value

Our projected five-year lifetime value for each of these 1,000 new customers is
£271.94 at net present value. If we are prepared to invest 30 per cent of this in
recruitment we can afford to spend £81.58 to recruit a single customer. This may
sound like a lot of money but:

G The lifetime value will continue to increase because we still have 205
customers buying from us and retention rate is by now a healthy 80 per
cent, so this cell will deliver good profits for several more years

G We have used a conservative discount rate of 10 per cent (to allow for the
risk factors we mentioned earlier) so the real net present value we realise
may be somewhat higher than predicted.

Allocating acquisition funds using lifetime value

Here we would build several cells, each based on a single medium – these could
be down to the level of individual publications or mailing lists, or if no major
variations occur within a media group, we may simply compare customers
recruited from newspapers with those responding to magazine advertisements,
direct mailings, television, online advertising and so on.

Another aspect we may wish to examine is how lifetime value may vary according
to the type of offer we made to persuade the recruit to place their first order. As
mentioned earlier a discount offer may attract more recruits but their long-term
value may turn out to be lower than a smaller group who bought at the full price.
We can compare as many variables as we wish, providing only that there are
enough customers in any single category to give a significant reading.

3.5 – 12
 Chapter 3.5 : Calculating the lifetime value of a customer

As with all testing, if we want to evaluate the precise effect of varying any single
factor, we must ensure that this change is isolated into a single cell. (For a full
discussion of testing, see chapter 8.2.)

When doing this sort of comparison it can be helpful to allocate some sort of
simple ratio to each result – this makes it much easier to compare multiple cells
and prioritise them on the basis of return on investment. For example, a magazine
publisher analysed the lifetime value and the acquisition cost per customer by
source, and expressed these as ratios of NPV profit to cost, then compared these
ratios to help develop customer acquisition strategy. The data is set out in the
following table 3.5.4 – which source do you think is most valuable?

Table 3.5.4 Lifetime value and acquisition cost by source

Source Lifetime value Cost per customer Ratio

A £12.59 £4.32 2.91

B £3.87 £1.68 2.30

C £20.41 £5.67 3.60

D £7.37 £2.27 3.25

This table shows the danger of relying purely on cost per new customer when
planning acquisition strategies. Although source B generates customers at the
lowest cost, the most productive source of profit is source C – where the cost per
customer is highest of all, but so is the profit per customer (lifetime value minus
cost) and, more particularly, so is the return on investment, indicated in the ratio
column above. The ratio for customers from source C indicates that we generate
£3.60 NPV profit for each £1 we spend on the acquisition campaign.

As we can see the ratio gives us a much clearer picture of the true value of each
new recruit. So, assuming there is a large enough pool of prospects in each
source to enable us to hit our overall sales targets, we would first of all
concentrate on source C and then on source D.

Choosing products for customer acquisition offers

As we said earlier, LTV can depend on the type and value of a customer’s initial
purchase. Once we have identified these causes and effects we can make more
informed decisions about what to feature in our acquisition promotions. Clearly
we would want to feature those products and offers that have been shown to
produce customers with the highest lifetime values.

In the following example, figure 3.5.5, an association plotted the values of its
customers according to which of its services they bought. It found that there were
some very different values in each line of business.

3.5 – 13
 Chapter 3.5 : Calculating the lifetime value of a customer

Figure 3.5.5 Which activity produces most long-term income?

The association markets memberships, subscriptions, products and training

seminars. It has to make important decisions about how to allocate its
marketing funds among these various areas. To make those decisions
correctly, it needed to see how much its members/customers were worth to it
over several years. It found that there were some very different values in each
line of business as illustrated below:










… 
0HPEHUVKLS 6XEVFULSWLRQVWR 3URGXFWVDOHV 7UDLQLQJ
IHHV SXEOLFDWLRQV VHPLQDUV

These results show that training seminar delegates were by far the most valuable,
while product buyers were least valuable. Until the association did this analysis it
had concentrated on acquiring product buyers, as these were the least expensive
to recruit. Clearly by recruiting more seminar delegates it would increase the long-
term value of its customer database significantly.

Two ways to calculate customer lifetime value

All types and sizes of businesses should start to estimate their customers’ lifetime
values. It is easy to build LTV tables using Excel, Lotus or any other good
spreadsheet program.

There are two basic perspectives on lifetime values – historical and projective.

G Historical – here we would take a group of customers acquired in the past

and track all revenues and costs associated with them. The main purpose of
this is to gauge the effectiveness and ROI of an investment made several
years ago. This is the easier of the two options because it simply uses
historical facts rather than building projections based on estimates.

G Projective – in this approach we use previously observed trends and ratios

and project them into the future. This is used when building a case for new
investment, assessing the value of a customer file in calculating the worth of
a company, and so on.

3.5 – 14
 Chapter 3.5 : Calculating the lifetime value of a customer

Table 3.5.5 Advantages of historical and projective lifetime

values
Historical Projective

 Relatively simple to calculate and implement  Reflects most recent

business conditions

 Result is factual, not speculative  Can be done without

many years of historical data

 Easier to tie to financial statements  Easier to do what if

analyses

Calculating historical lifetime value

There are four basic steps in this process:

1. Identify a group of customers, all of whom started as customers at the same

time – perhaps five years ago although this period is variable according to
the type of business you are in. As we explained earlier if your customers
only buy every two or three years you would be more likely to choose a
period between seven and 10 years.

2. Record the sales revenue for this group each year (or whatever period you
decide to use).

3. Calculate costs and thus profits for each period.

4. Discount each profit figure to arrive at the net present value for each period.

Table 3.5.6 shows a worked example of this process:

Year 1 Year 2 Year 3 Year 4 Year 5

A Customers 1000 660 456 338 261.0

B Retention rate 66% 69% 74% 77% 83%

C Sales p.a. £615 £673 £721 £812 £872

D Total sales £615,000 £444,180 £328,776 £274,456 £227,592

E Net profit £123,000 £88,836 £65,755 £54,891 £45,518

20%

F Discount rate 1 0.94 0.89 0.840 0.790

G NPV £123,000 £83,506 £58,522 £46,109 £35,941

contribution

H Cum NPV £123,000 £206,506 £265,028 £311,137 £347,078

contribution

I Lifetime value £123.00 £206.51 £265.03 £311.14 £347.08

at net present
value

3.5 – 15
 Chapter 3.5 : Calculating the lifetime value of a customer

The key points to note are:

Row A – customers – this table plots the sales and profits achieved from a cell of
1,000 customers who started five years ago. The active customer total declines as
attrition takes its toll.

Row B – retention rate – the percentage of customers who continued to buy each
succeeding year.

Note – because this method uses actual data the figures in Rows B and C are
precise rather than rounded estimates.

Row C – sales per annum – this shows the average annual sales per customer of
this group; this multiplied by Row A gives each year’s total sales.

Row E – net profit – in this example we have assumed that the company
maintained its targeted net profit margin of 20 per cent each year; this could be
variable, of course.

Row F – discount rate – note how in the historic model we do not make
additional allowances for possible problems and we simply use a discount factor
that reflects what the true cost of our business finance has been – in this example
approximately 6 per cent per annum.

Rows G and H show the NPV figures, year by year and cumulative.

Row I – we divide the cumulative NPV figure by the starting number of customers
in the cell to show how LTV per customer increases each year.

Many people make the mistake of dividing the final cumulative NPV profit
figure by the remaining number of customers in the cell. This gives a very
highly inflated figure and could lead to a major error in strategic planning.

The 1,000 customers who started five years ago (for example in 2001) have shown
a cumulative £347.08 LTV at net present value in 2001 (i.e. related back to the
time of acquisition five years ago).

We can thus see whether the investment we made in acquiring these customers
has been returned with a profit over the period.

Calculating projective lifetime value

In this model we are predicting retention rates, annual sales and profits, and
interest rates, so we need to be rather more cautious – remembering the risks we
discussed earlier.

Apart from these considerations the basic steps are very similar and the table
layout remains the same. The process is as follows:

1. Segment new customers into cells – these may be based on media, offer,
type of first purchase, first-order value, in fact any significant variable of the
acquisition process. Alternatively, we may set up matched cells and use the
LTV analysis to measure the effect of differential ongoing marketing
programmes. We will demonstrate this process later.

3.5 – 16
 Chapter 3.5 : Calculating the lifetime value of a customer

2. Choose a time period that is appropriate to your business – this could be a

quarter, half year, year or even longer. As mentioned earlier, if doing such
analyses for companies selling high-value durables with a repurchase period
averaging three years, we may use that as the period.

3. Estimate sales and profits for the customer group in each cell – this would
be based on results of similar customers in recent years.

4. Plot your predictions over the period of your analysis.

5. Apply the discount factor for each year and you can then predict the lifetime
value at net present value for this cell.

The layout of the table would be the same as that above except we would be
dealing with predicted rather than historic data.

Our table would look like this:

Table 3.5.7
Year 1 Year 2 Year 3 Year 4 Year 5

A Customers 1000 600 390 273 205.0

B Retention rate 60% 65% 70% 75% 80%

C Sales p.a. £600 £620 £640 £660 £685

D Total sales £600,000 £372,000 £249,600 £180,180 £140,425

E Net profit £120,000 £74,400 £49,920 £36,036 £28,085

20%

F Discount rate 1 0.9 0.81 0.729 0.656

G NPV £120,000 £66,960 £40,435 £26,270 £18,427

contribution

H Cum NPV £120,000 £186,960 £227,395 £253,665 £272,092

contribution

I Lifetime value £120.00 £186.96 £227.40 £253.67 £272.09

at net present
value

Note – if you get your calculator out and check the figures in this table you may
find slight variances – this is simply because Excel has been set to show to three
decimal places. For example in the year five column the discount rate is shown as
0.656 whereas it is really 0.6561. Thus the resulting NPV figure of £18,427 seems
to be incorrect. If you divide the net profit figure by 0.6561 you will see that the
NPV figure is correct.

The key differences between this and the historical version are:

G Retention percentages and sales figures are round numbers as they are
estimates. Discount rate is 10 per cent as discussed earlier.

G Because of these factors the projected LTV figure is less than we saw in the
historic model. This is no bad thing, as it will at least ensure that we do not
overinvest in acquisition.

3.5 – 17
 Chapter 3.5 : Calculating the lifetime value of a customer

To summarise what we have covered so far:

We have seen what lifetime value analysis is and how it works to:

G Help us understand the true value of a customer over time

G Help us measure the true ROI of a previous acquisition programme

G Help us decide on an appropriate level of investment in future new

customer acquisition

We have explained the need to discount future earnings to allow for the cost of
money and demonstrated the basic steps in the process.

Before we move on to showing how LTV analysis can be used to measure the
effects of retention marketing, here is a final example of how we might use the
process to compare the effectiveness of two alternative recruitment sources.

Let us imagine we are comparing two cells of new customers who started with us
around the same time (for example, 3 years ago). The customers in table 3.5.8 all
started as a result of a direct mail campaign.

Table 3.5.8
Year 1 Year 2 Year 3

A Customers 1000 650 455

B Retention rate 65% 70% 75%

C Sales p.a. £600 £620 £640

D Total sales £600,000 £403,000 £291,200

E Net profit £120,000 £80,600 £58,240

20%

F Discount rate 1 0.9 0.81

G NPV £120,000 £72,540 £47,174

contribution

H Cum NPV £120,000 £192,540 £239,714

contribution

I Lifetime value £120.00 £192.54 £239.71

at net present
value

As we can see each of these customers has produced a cumulative net present
value profit of £239.71 over the three years.

Now let’s compare this data with a cell of new customers who started at the same
time as a result of an online advertising campaign.

What differences might we see?

G Perhaps the online recruits will be younger, more impulsive and have more
disposable income?

G So, in year one we may well see an increase in sales value, BUT

3.5 – 18
 Chapter 3.5 : Calculating the lifetime value of a customer

G This may be offset by a reduction in retention as they tend to be more fickle

and may be quicker to seek new suppliers?

Table 3.5.9
Year 1 Year 2 Year 3

A Customers 1000 400 180

B Retention rate 40% 45% 50%

C Sales p.a. £700 £730 £760

D Total sales £700,000 £292,000 £136,800

E Net profit £140,000 £58,400 £27,360

20%

F Discount rate 1 0.9 0.81

G NPV £140,000 £52,560 £22,162

contribution

H Cum NPV £140,000 £192,560 £214,722

contribution

I Lifetime value £140.00 £192.56 £214.72

at net present
value

Comparing these two tables we see:

G Although in the second table sales per customer are higher throughout the
three years:

G The lower retention rate more than compensates for this, and

G After the three years there are only 180 customers left compared to 455
from the direct mail cell. Thus the longer this evaluation is run the more it
will favour the direct mail recruits.

The difference in NPV profit per customer may seem quite small at £24.99 but
overall it represents almost £25,000 per 1,000 customers recruited from one
source rather than the other.

Note – it is important to point out that this is a hypothetical example and is

simply based on an assumption of the differences between the two types of
prospect. It is not to suggest that online recruitment is not a valuable source of
business but simply to demonstrate the process of comparison.

3.5 – 19
 Chapter 3.5 : Calculating the lifetime value of a customer

Using LTV analysis to compare the effectiveness of various

customer development strategies
We are now going to look at another useful application of LTV analysis. In this
case we want to measure the effectiveness of a new retention strategy.

Let us assume we are a large retailer and our customers receive regular
advertising messages through our normal marketing programme. This includes
TV, radio and press advertising, local poster campaigns and so on. Prior to this
campaign we had not generally sent mailings to customers, but the new
programme was set up to see if direct communications would increase our share
of customer spending (i.e. encourage them to visit our store more often rather
than shopping with our competitors).

Three years ago, after launching our customer reward card we started to test a
new strategy. This required us to gather comprehensive data on the shopping
preferences and behaviour of a group of card holders and then use that data to
mail highly relevant offers to these people. We used our store card as the tool for
gathering this information.

We obtained basic demographic and lifestyle information from the customer’s

store card application form. Our rewards programme ensured that customers
used their card every time they shopped with us and this enabled us to gather
additional data such as:

G Frequency of visiting one of our stores

G Amount spent

G Products purchased

This behavioural data, added to the application form information, gave us a very
rich source of data, enabling very precise targeting of offers.

Another separate but matched group of customers was used as a control – they
did not receive the additional communications. They did of course receive
statements of their accumulated points balance but did not receive the highly
targeted and relevant offers made to the test group.

Let us look at the control group first – table 3.5.10 plots their business with us
over the past three years; note we are using the historical method here as we
want to be sure that the money we are spending on the additional
communications is paying off in terms of additional profit.

3.5 – 20
 Chapter 3.5 : Calculating the lifetime value of a customer

Table 3.5.10 Control group with no additional communications

Year 1 Year 2 Year 3

A Customers 1000 680 482.8

B Retention rate 68% 71% 75%

C Sales p.a. £1,615 £1,673 £1,721

D Total sales £1,615,000 £1,137,640 £830,899

E Net profit £80,750 £56,882 £41,545

5%

F Discount rate 1 0.95 0.9

G NPV £80,750 £54,038 £37,390

contribution

H Cum NPV £80,750 £134,788 £172,178

contribution

I Lifetime value £80.75 £134.79 £172.18

at net present
value

The rows are the same as those used previously; as this is an historical
comparison we are again using actual retention percentages, sales figures and
interest rates – in this case five per cent. All normal advertising and marketing
costs are included, i.e. the net profit figure of five per cent is after all normal
marketing costs are paid for.

Let us now look at the test cell in table 3.5.11 – that group of customers with
whom we have developed relationships through highly targeted and relevant
communications. You will notice some additional rows in this second table and
these will be explained in a moment:

3.5 – 21
 Chapter 3.5 : Calculating the lifetime value of a customer

Table 3.5.11 Test cell – with additional customer relationship

programme
Year 1 Year 2 Year 3

A Recommends rate 7% 6%

B Customers gained 70 49

C Customers 1000 820 705.2

D Retention 75% 80% 85%

E Sales p.a. £2,553 £2,628 £2,722

F Total sales £2,553,000 £2,154,960 £1,919,554

G Net profit £127,650 £107,748 £95,978

5%

H Retention activities £15,000 £12,300 £10,578

£15

I Net contribution £112,650 £95,448 £85,400

J Discount rate 1 0.95 0.9025

K NPV £112,650 £90,676 £77,073

contribution

L Cum NPV £112,650 £203,326 £280,399

contribution

M Lifetime value £112.65 £203.33 £280.40

at net present
value

Now, what is different about this table?

First of all we have a new row at the top – the recommends rate.

Row A – recommends rate – when we write to our customers with interesting

and relevant offers we also have an opportunity to ask them to recommend their
friends to join our store card scheme. We offered them a small incentive; for
example, 250 extra reward points for each of their friends who joined the scheme.
In the first year seven per cent of customers recommended a friend who joined
and in the second year six per cent did the same. This gave us an additional 119
customers to add into the cell at Row B.

LTV cells record the actual or projected sales of the customers in the cell, so it is
not permitted to add customers during the period – except in this one
circumstance – the 119 recommended customers are only there because they were
introduced to us by the customers already there; so in this case we can
legitimately add them to the model.

Row D – retention rate – as we are communicating regularly with these

customers it is not surprising to see that the number who continued to buy from
us is higher than in the control group.

Row E – sales per annum – similarly, the mailings and promotional offers we
send to this group will increase the frequency of their visits and the amount they
spend in the year.

3.5 – 22
 Chapter 3.5 : Calculating the lifetime value of a customer

Row G – net profit – this remained at five per cent as the additional marketing
costs for this group are allowed for in Row H.

Row H – retention activities – these costs averaged out at £15 per customer per
year; this was sufficient to cover the necessary data collection and analysis, and
production and despatch of four mailings each year. This amount is deducted
from the net profit figure before discount is applied.

Rows J to M – these are the equivalent of Rows F to I in the control group table.

So let us summarise what these historical LTV analyses show us in table 1.5.12:

Table 3.5.12 Summary of analyses

Year 1 Year 2 Year 3

Table 1 £80.75 £134.79 £172.18

Table 2 £112.65 £203.33 £280.40

Increase £31.90 £68.54 £108.22

1000 £31,900 £68,538 £108,221

customers

5000 £159,500 £342,689 £541,103

customers

Looking at the control group table we see that the three-year lifetime value
amounted to £172.18 per customer, discounted back to the starting point three
years ago.

The group having the relationship management programme on the other hand
showed a three-year LTV of £280.40, an increase of £108.22 per customer.

This figure disregards the setting-up costs of the programme of course, but this is
reasonable, as this cost would be amortised over a much larger customer group
and a much longer period.

The above summary table shows us that, even with only 5,000 customers in the
new customer development programme, it would deliver more than £540,000
additional net profit over three years.
We can thus see that this programme is successful and well worth continuing.

What if the results are not so good?

Well, as we can see, the gains were obvious even at the end of year one, so if the
numbers were not so good, we would also see this early and be able to modify or
even abandon the programme quite quickly.

The need for calculations, not guesswork

In today’s highly competitive environment a well-developed customer database is

a major asset for any company. Competition generates attrition (or customer
churn as it is sometimes called) and even the best managed companies need to
invest in acquiring new customers and reactivating those who have lapsed.

3.5 – 23
 Chapter 3.5 : Calculating the lifetime value of a customer

This process is financially quite sensitive and it is vital to be able to measure the
cost-effectiveness
of our efforts.

Customer lifetime value analysis is the process that brings together the marketer’s
need to develop the customer database and the financial manager’s need to
optimise company investments.

It can be used to:

G Answer the question: “How much can we afford to spend to acquire a new
customer?”

G Justify long-term investments in developing customer relationships and

measure comparative value of different retention (customer relationship)
programmes

G Estimate true return on investment after a programme has been

implemented

G Measure the comparative effectiveness of different recruitment media,

offers, timings, products and so on

G Place an asset value on a company’s customer database

To summarise
Customer lifetime value is the measure of a customer’s worth to the company over
the entire time that customer continues to buy. It can vary according to the
methods or media through which customers are acquired. It is used to inform
investment decisions in many key marketing areas.

3.5 – 24