MARK3054 Lecture Notes - Lecture 4: Cash Flow, Sensitivity Analysis
Topic 4: Customer Value Assessment
Issues with preference estimation:
• Many data is not observable for real behaviours (we can know which products a customer
buys but don’t know their potential options)
• Data from research studies (e.g. Conjoint) are detailed, but intentional (suffers hypothetical
bias). Also, difficult to fully mimic real purchase situations in research studies
Methods to use Purchase Records
RFM Model: who is more likely to buy in the next period?
Through a Total RFM Score where:
R (Recency): time/purchase occasions since last purchase
F (Frequency): number of purchase occasions since first purchase
M (Monteary value): amount spent since first purchase
1. Arbitrarily assign an R, F, and M score different recencies, frequencies, and amounts spent
2. Add scores for Total RFM Score
IMPORTANT: When coding scores, highest score must be assigned to the most FAVOURABLE. This is
debatable however, as the highest score can be attributed to either VERY recent or NOT VERY recent
depending on product type (i.e. a customer who recently purchased a car would not purchase again
for some time, while a customer who has not purchased for a while may be considering now)
Issues: arbitrary (low predictive power), only about next period
Customer Lifetime Value (CLV)
Transaction View: Acquisition Cost, base profit (of selling one product)
Long-Term Relationship View: Acquisition Cost, base profit, demand increase, price premium, cost
savings, relationship value (if you have customers similar to the ones you want to acquire, you can
also learn the best promotional/acquisition strategies which will work i.e. reference, referral,
learning, innovation etc.)
Economic CLV= (expected) revenue cash flow – (expected) cost to serve = expected profit cash flow
(adjust for risk) = discounted cash flow
NB: loyalty will lower both expected cost to serve and risk adjustment
CLV= (1+r)/(1-p+r) * (R-C) – A
Excel Model Advantages:
• See what you get
• Simulation (smart worksheet)
• Sensitivity analysis
• Adjustment in the procedure (e.g. growing margins, different discount rate, different
retention rate)
• Visualisation
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Also, difficult to fully mimic real purchase situations in research studies. F (frequency): number of purchase occasions since first purchase. M (monteary value): amount spent since first purchase: arbitrarily assign an r, f, and m score different recencies, frequencies, and amounts spent, add scores for total rfm score. Important: when coding scores, highest score must be assigned to the most favourable. Issues: arbitrary (low predictive power), only about next period. Transaction view: acquisition cost, base profit (of selling one product) Economic clv= (expected) revenue cash flow (expected) cost to serve = expected profit cash flow (adjust for risk) = discounted cash flow. Nb: loyalty will lower both expected cost to serve and risk adjustment.
