关于「精算工具与执行清单:订阅现金流月度台账、试算器与预注册检验」,《订阅叠砖、扣款挽回与年付拐点》给出的定量结论是什么?
结论:台账每月一行,按订阅层分表,主动与被动流失分开、失败按四类原因登记、年付与递延收入单列。试算器按台账算出稳态收入、各原因挽回率、下月重试日、年付拐点与警示,示范台账的稳态收入是 251,440 个单位,亮了两盏灯。执行清单把动作落到日、周、月、季。预注册检验给出样本量:重试排程每组 356 笔,首次续费每组 691 位,留存上升检验一个队列 680 位。
数据口径:书中的订阅数、客单、流失率、挽回率、折现率与价格均为示意参数,全部计算由书中程序生成,不对应任何真实客户或成交数据;本书是非法律、会计、税务或投资建议,不承诺任何收入水平,「稳定」只指收入可预期并写明前提与失效条件。 第一至三章说明,稳态收入等于每月新增乘以客单再除以月流失率,新增定斜率,流失率定天花板;订阅者异质时总体留存随月龄上升,移位贝塔几何模型用几个月的队列数据就能外推;前两次续费是筛子,单次干预回报是后段的 1.40 倍,三年总增益却在长尾,示意参数下是 18,598 对 4,760;订阅层只在刚需使用次数明显高于值回票价的门槛时成立。
对照数据表
下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 356, 680, 691
| test | p1 | p2 | n per arm |
| H1 retry schedule | 0.600 | 0.700 | 356 |
| H2 first renewal | 0.650 | 0.720 | 691 |
| H3 renewal 6 vs 1 | 0.847 | 0.917 | 333 |
| H3 cohort size so month-5 survivors reach n | 680 | ||
| H1 simulated power at n = 356 per arm | 0.797 |
查看完整验算程序
# Chapter 8. p82: pre-registered tests and their sample sizes.
# Two-sided alpha 0.05, power 0.80, two-proportion normal approximation:
# n per arm = (z_a * sqrt(2 pbar qbar) + z_b * sqrt(p1 q1 + p2 q2))^2
# / (p1 - p2)^2
# H1 retry schedule: recovery of insufficient-funds failures 0.60 -> 0.70.
# H2 carved layer: first-renewal rate 0.65 -> 0.72 after a usage nudge.
# H3 heterogeneity: retention in renewal 6 (0.917) above renewal 1
# (0.847), the sBG values of p21; cohort size needed so that month-5
# survivors (0.490 of the cohort by p21) reach the n of the test.
# The H1 size is checked by a fixed-seed simulation of 2,000 trials.
import math
import random
import statistics
Z = statistics.NormalDist()
ZA, ZB = Z.inv_cdf(0.975), Z.inv_cdf(0.80)
def n_per_arm(p1, p2):
pbar = (p1 + p2) / 2.0
top = ZA * math.sqrt(2 * pbar * (1 - pbar)) + \
ZB * math.sqrt(p1 * (1 - p1) + p2 * (1 - p2))
return math.ceil(top ** 2 / (p1 - p2) ** 2)
rng = random.Random(20261120)
print("test p1 p2 n per arm")
for name, p1, p2 in (("H1 retry schedule", 0.60, 0.70),
("H2 first renewal", 0.65, 0.72),
("H3 renewal 6 vs 1", 0.847, 0.917)):
print("%-26s %6.3f %6.3f %9s" % (name, p1, p2,
"{:,}".format(n_per_arm(p1, p2))))
n3 = n_per_arm(0.847, 0.917)
print("H3 cohort size so month-5 survivors reach n %s" % "{:,}".format(
math.ceil(n3 / 0.490)))
n1 = n_per_arm(0.60, 0.70)
hits = 0
for _ in range(2000):
a = 0
b = 0
for _ in range(n1):
a += rng.random() < 0.60
b += rng.random() < 0.70
pa, pb = a / n1, b / n1
pool = (a + b) / (2.0 * n1)
se = math.sqrt(pool * (1 - pool) * 2.0 / n1)
if abs(pb - pa) / se > ZA:
hits += 1
print("H1 simulated power at n = %d per arm %.3f" % (n1, hits / 2000))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《月度经常性收入叠砖、扣款失败自动挽回与年付现金流锁定母典》第 8 章「精算工具与执行清单:订阅现金流月度台账、试算器与预注册检验」(版本 1.0.0). 5D Library. DOI 待发布
DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-367 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026367,
author = {Fan, Yafu},
title = {Stacking Monthly Recurring Revenue, Involuntary-Churn Dunning Recovery and Upfront Annual Cash Lock-In},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}