Tools › Finance and Business › Marketing
Marketing
2 tools in the Finance and Business area.
Cohort Retention and CLV
Customer retention by acquisition cohort and the value of a customer, for subscription and repeat-purchase businesses: paste a retention table (cohort, size, active counts by period), transaction rows (customer, date, amount: the tool builds the monthly cohorts itself) or a single retention curve. The page shows the retention matrix in percent with the size-weighted average curve and the period churn rates, the Kaplan–Meier survival of a customer with Greenwood standard errors (cohorts not yet observed at a period are censored), the shifted-beta-geometric model of Fader and Hardie fitted by maximum likelihood (the churn probability varies across customers by a beta distribution) beside the constant-churn geometric fit, projected retention to a chosen horizon, the discounted expected residual lifetime of a customer by tenure and the residual value of the customer base, the CLV multiple and payback against the acquisition cost, and the simple margin formula CLV = m r/(1 + d − r) of Gupta, Lehmann and Stuart. The default case is the five-cohort table of Fader and Hardie's sBG notes (α = 3.80, β = 15.19; DERL 3.31 to 3.84 at a 10 % discount rate) with their Regular-segment retention curve (α = 0.704, β = 1.182) fitted to years 1–7 and projected over years 8–12.
4 worked examples reproduced, listed on the tool page.
Customer Metrics (NPS)
Net Promoter Score, CSAT and Customer Effort Score from pasted survey responses (one respondent per line with optional segment, period and driver ratings, or a tally of counts per score point): the score with its standard error and confidence interval (NPS by the multinomial variance of Rocks 2016 with his recommended Adjusted Wald interval; CSAT top-two-box and CES favourable share by the Wilson score and Agresti–Coull intervals), the promoter / passive / detractor distribution, comparison tests against an industry benchmark and a prior period, between segments (difference with its standard error, X² test of the category distributions) and between periods, the trend over periods with a weighted-least-squares slope and a single change-point test, the sample size for a target interval half-width, and a key-driver analysis of the driver ratings (Pearson and Spearman correlations with the score, importance–performance quadrant chart, multiple regression with standardized coefficients and Pratt importance). No benchmark tables are shipped: benchmarks are entered by the analyst. The default case is an illustrative file of 400 responses in three segments over four quarters with three driver ratings (NPS +26.5, 95 % interval +18.6 to +34.0); the library is checked against Rocks (2016), the seventeen published Adjusted Wald widths of Sauro and Lewis, Newcombe's Wilson intervals, Agresti's X² table and the NIST StRD regression datasets.
1 worked example reproduced, listed on the tool page.
Other Finance and Business sections
Finance · Accounting · Economics and Damages · Personal Finance · Pricing · Insurance · Operations and Capacity