05 Class Notes Delivered During Class Five

Class 05 (and 04)
Marketing Analytics
Dhahran Roads
(the importance of time value of money)
Project Profits of 22 SR
have NPV of 7.17.
Project IRR is 41%

(found using =IRR or
goal seek).
Next step: Sensitivity

Analysis. What if costs
over-run? What if
payments are
delayed?
Next step: Use CB to

see how risky NPV is.
TIAVI
Applying time value of money to Customer Relationships
1,000
customers each
have CLV of
$155.77
500 converted
customers have
PV of $128,445
at t=1.
CLV if renew at
t=2 is $256.51
CLV if
converted at
t=1 is $256.89
The Crutchfield Corporation
• Catalogers often segment their customers
based upon RFM
• R is recency
– How many catalogs since last bought.
• F is frequency
– How many times ever bought.
• M is monetary value
– Either average or total dollars spent
1. How much did it cost us to acquire customers using the
Car list?
2. Build an excel model to track the future recencies of 1,000

new customers for the purposes of estimating CLV. (See
class three notes above for more detail.) Should we
prospect the remainder of the Car list?
3. (Optional) Currently we quit mailing catalogs to customers

after recency 24. What do the economics say about this
cut off? Does a quicker cut off make economic sense?
1. How much did it cost us to acquire customers using the
Car list?
2. Build an excel model to track the future recencies of 1,000

new customers for the purposes of estimating CLV. (See
class three notes above for more detail.) Should we
prospect the remainder of the Car list?
3. (Optional) Currently we quit mailing catalogs to customers

after recency 24. What do the economics say about this
cut off? Does a quicker cut off make economic sense?
Using Both
R and F,
there
would be
24x5 = 120
customer
buckets
(states)
Using R
only, there
would be
24 + 1 = 25
customer
buckets
(states)
To the spreadsheet…
Break!
Customer Lifetime Value $
• What it’s good for?

– If we know CLV, we know which lists to roll out
and which to ignore
• Break even response rate = c/CLV = $1/$25 = 0.04
• Same is true for TIAVA. Don’t roll out JJJ.
• Same is true for Crutchfield.
– In deciding how to treat existing customers, maximizing
CLV is the goal.
• Crutchfield cut off should be based on CLV, for example.
• Every Customer Relationship Management (CRM) Decision should
be based on CLV.
– E.g., Cap One experiments to decide how to respond to customer
requests.
– Next class, we will explore using CLV to value companies.
Customer Lifetime Value $
• How to Estimate CLV?

– Cohort and Incubate
• Set aside a bunch of new customers. Keep track of
their subsequent Cash Flows. PV the cash flows for the
cohort back to acquisition date; divide by the number
of customers.
– Just Spreadsheet it.
• TIAVI (.5 conversion followed by .75 renewal rates)
• Crutchfield (repurchase rates depend on recency)
– A SIMPLE FORMULA
The Simplest CLV Model
• The firm receives $M in net cash flow each period
the customer is retained.
– $M is revenue minus costs minus any retention
marketing spending.
• The customer is retained with probability (or
rate) r each period.
– The customer churns with probability 1-r.
• The per-period discount rate is d.
• The customer is “lost for good” the first time she
is not retained.
CLV of an existing customer who was
just retained.
r  $M r  $M r  $M 2 3
CLV     ...
1 d (1  d ) 2
(1  d ) 3
$M  r
CLV 
1 d  r
See “Customer Lifetime Value” note in notebook for more details.
just retained.
r  $M r  $M r  $M
2 3
CLV     ...
1 d (1  d ) 2
(1  d ) 3
$M  r If r=0,
CLV  CLV = $0.
1 d  r
just retained.
r  $M r  $M r  $M
2 3
CLV     ...
1 d (1  d ) 2
(1  d ) 3
$M  r If r=1,
CLV  CLV = $M/d.
1 d  r
CLV of a new customer (if and when
we acquire her).
r  $M r  $M r  $M 2 3
CLV  $M     ...
1 d (1  d ) 2
(1  d ) 3
$M  (1  d )
CLV 
1 d  r
CLV of a new customer (if and when
we acquire her).
r  $M r  $M r  $M 2 3
CLV  $M     ...
1 d (1  d ) 2
(1  d ) 3
$M  (1  d ) If r=0,
CLV  CLV = $M.
1 d  r
An ISP charges $19.95 per month. Variable costs are
$1.50 per account per month. With marketing
spending of $6 per year, attrition is only .5% per month.
At a monthly discount rate of 1%, what is CLV?
• $M = $19.95 - $1.5 - $6/12 = $17.95

• r = 0.995
• d = 0.01
• CLV of existing customer = $M x r/(1+d-r) =

$1,191
• CLV of a new customer = $M(1+d)/(1+d-r) =
$1,209
Assignment for class 5
• Please complete the CLV Exercises (linked from
assignment page). Each team email me their answers
(in one file) before start of class next Thursday. This
counts as a team exercise.
• Be familiar with “Valuation of Netflix” (in notebooks).
• Be familiar with “Netflix Case Study: David becomes
Goliath” (link to this is on the course assignment page)
– What were the keys to Netflix success?
– Which of the three valuation methods is best?
– Use the simple CLV formula (as best you can) to estimate
the CLV of an average Netflix customer at the end of 2008.
• Each team please prepare one slide with their CLV calculation.
To the Cup…..
You get to choose which of the five

exercises you want to present…

05 Class Notes Delivered During Class Five

Uploaded by

Copyright:

Available Formats

05 Class Notes Delivered During Class Five

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

05 Class Notes Delivered During Class Five

Uploaded by

Copyright:

Available Formats

Class 05 (and 04)

Project IRR is 41%

Next step: Sensitivity

Next step: Use CB to

2. Build an excel model to track the future recencies of 1,000

3. (Optional) Currently we quit mailing catalogs to customers

2. Build an excel model to track the future recencies of 1,000

3. (Optional) Currently we quit mailing catalogs to customers

• What it’s good for?

• How to Estimate CLV?

• $M = $19.95 - $1.5 - $6/12 = $17.95

• CLV of existing customer = $M x r/(1+d-r) =

You get to choose which of the five

You might also like