thankQ Help
The most common scoring method is to sort customers in descending order (best to worst). Customers are then broken into five equal groups or quintiles.
The best receive a score of 5, the worst a 1.
Recency |
customers are sorted by days since last purchase, the lower the number of days, the better the score |
Frequency |
customers are sorted by number of purchases, the higher the number of purchases, the better the score |
Monetary |
customers are sorted by the amount spent. The higher the amount, the higher the score |
Each time customers are scored, a new relative segmentation scheme is created. This has the advantage of quantifying customer behaviour which can be projected into the future. The relatively best customers would always fall into the 5,5,5 category.
It is necessary to identify where the cut off points fall, since they automatically change with each customer scoring.
The customer quintile method has the advantage of yielding equal numbers of customers in each segment. There are five equal groups for RFM, generating 125 equal size segments overall.
Initial analysis would be to contact all customers, look at the performance of each individual cell (cells would have definitions like: 4,3,5 or 2,3,3) and understand how different segments of the customers perform.
With 600,000 customers there would be 4,800 in each cell. A response rate of 2% would yield 96 orders giving you an acceptable sample for analysis. With less than 600,000 customers, it would be highly questionable to evaluate each cell independently.
Instead, the RFM would be evaluated by looking at the relative performance between the R scores, the F scores and the M scores. This may not be as satisfying, but it would provide statistically significant results.
Thus, a 100,000 customer mailing would have 20,000 in each grouping (looking only at one dimension at a time). This method extends the usefulness of RFM down to the neighbourhood of 10,000-25,000.
Recent customers are considered viable for a certain length of time. They are often mailed heavily in the first 12 months and increasingly less often until, say, 36-48 months.
Unlike Frequency and Monetary, customers reset themselves. A three-year-old reordering customer who has purchased an average amount only once moves up in an orderly manner from a 1 in F&M to a 2 in F&M. But in Recency, he jumps from a 1 to a 5! Customers who order often may never have anything other than a 5 score. At the core of Recency is the fact that most of the customers fall into two groups: hot and dead.
