Assignment 1:
A company is considering whether to do a test on its marketing campaign. Direct marketing costs €0.40 to send.
The company sells a product for €50 and their margin is 50%
Question 1:
What is the breakeven response threshold, i.e., the response probability this company needs for its mailing
to be profitable? Provide your answer with three decimals separated by a dot, not a comma (e.g., 0.123).
Marketing cost = 0.40
Margin 50%, Sells = 50
P = c/ m = 0.40 / (0.50 * 50) = 0.016
Question 2:
The company estimates a response rate of 0.02 using a randomly selected test sample of 5000. If they decide
to roll it out to the population, assuming the sample is representative of the population, what’s the
probability that the true response rate is lower than the breakeven response rate? Round your answer to
two decimals separated by a dot, not a comma (e.g., 0.12)
p̂ = 0.02
#̂(&'#̂
Standard error (p̂ ) = ! (
= *0.02 ∗ (1 − 0.02)/5000 = 0.00198
NORM.DISTR(threshold; p̂ ; s.e; TRUE)
NORM.DISTR(0.016;0.02; 0.00198;TRUE) = 0.02
Question 3:
How many users should the test sample comprise, assuming they want to test whether the response rate of
0.02 is really higher than the break-even? The company can tolerate a 5% chance of error declaring the test
is above the breakeven when it is not, and a 5% error when declaring the test is not above the breakeven
when it is (the company is only concerned with whether the test is above the breakeven).
Use G-power:
• Test family = exact
• Statistical test = ‘proportion: difference from a constant’
• Set power (1- β): 0.95
• α error probability = 0.05
• Constant proportion = 0.016
• Effect size = p – pBE = 0.02 – 0.016 = 0.004
Total sample size = 11918
,Question 4:
Running a test is expensive in terms of time and money. So the company wants to calculate the value of the
test and see whether it exceeds their estimate of the cost. Their customer base totals 100,000. Their sample
is 5000 customers. In the past, campaigns have either been a success, with a response rate of 0.02, or a failure,
with a response rate of 0.001. Historically, 60% of past mailings have been a failure.
Recall that the direct marketing costs €0.40 to send. The company sells a product for €50 and their margin is
50%.
What is the expected profit without a test? Provide your answer without commas or dots (e.g., 5000)
The answer is 0 as the firm is not likely to rollout the marketing, since the probability of failure > success. If the
firm decides not to roll out, the expectation is 0.
Question 5:
What is the expected profit when the company would do a test? Provide your answer without commas or
dots (e.g., 5000)
Because we make a loss if we have a failure, we won’t roll out if failure. For success we make profits, so we will
roll-out to rest of population:
Success
Profit if text says success: N(psuccess*m - c) = 100,000 *(0.02*25 - 0.4) = €10,000
Profit if text says failure: n(pfailure*m-c) = 5000 * (0.001*25 - 0.4) = - €1875
Expected profit = (10,000*0.4) + (-1875*0.6) = €2875
Question 6
What’s the maximum amount you would spend on the test? Provide your answer without commas or dots
(e.g., 5000)
Maximum amount to spend should be equal to the profit. This is also €2875
, Assignment 2:
A company is considering using RFM segments to target its rollout. The mailing under consideration costs €0.60
to send; if customers respond, they spend on average €50, of which €15 is margin. It conducted a test mailing
of approximately 1000 customers.
The results of this test mailing are in the data set test_RFM.sav.
Create RFM segments by creating three groups each of RFM, using the nested (not independent) binning
method (which first creates segments in R, for each R segment 3 segments of F, and for each RF segment, 3
segments of M).
Question 1:
What is the difference between the average response rate in the most recent group (R = 3) and the least
recent group (R = 1)? Provide your answer with two decimals separated by a dot, not a comma (e.g., 0.12)
Cost: 0.50
Average spending: 50 à margin: 15
Test sample: 1000
Step 1: Create RFM Segmentation
1. Direct marketing à choose technique à Help identify my best contacts à customer data
o Customer identifier: account number
o Transaction date: R
o Total number of purchases: T
o Amount: M
2. Binning:
o Independent
o Number of bins: 3
Step 2: Make RR and NR
Segment customers in 3 groups based on recency:
1. Transform à rank cases
2. Variable: recency
3. Rank types:
a. Ntiles: 3
Step 3: Compare means recency
1. Analyze à compare means
2. Means:
a. Dependent list: response to mailing
b. Independent list: NR
