5D Library › Subscription Unit Economics, Net Revenue Retention and AI Inference Scaling Laws · 中文
Chapter 8. A Four-Step Check and a Pre-Registered Price Test
Four questions before trusting a subscription model
This chapter turns the book into four questions to ask of any subscription plan or pitch, including one’s own. They are not investment advice.
1. Which churn, from which model? A lifetime value computed from one churn rate is wrong in a direction that depends on the age of the base it came from (Chapter 1). Ask for the cohort survival curves, the discount rate, and the range of values across at least two models of heterogeneity. If the lifetime value is undiscounted, ask what it is with a discount.
2. Where is NRR in its life? An NRR above 100 % turns the installed base into a second engine (Chapter 2), but expansion is bounded and blended NRR falls as the base ages. Ask for NRR by cohort age, not only blended, and do not extrapolate a young company’s NRR.
3. What is the trough under this growth plan? LTV / CAC says nothing about timing (Chapter 3). Ask for the CAC payback period, the discounted ratio, and the cash trough and its month under the plan’s own growth path. If the plan’s funding is sized to a slower path than its targets, the plan is a plan to run out of money on success.
4. For an AI product: what does the mean user cost? The median user does not set the margin (Chapter 5). Ask for the distribution of usage, the cost per million tokens at the batch sizes and context lengths actually served (Chapter 4), the share of cost from the top 1 % of users, and how the usage of the base has changed with its age.
A pre-registered price test
Chapter 7 ended with a conflict that only an experiment can settle: a higher price raises first-year revenue but, if churn is price-sensitive, may lower value per sign-up. The test below is written before any data exist.
Hypothesis · not established
The study. New sign-ups are randomised at sign-up to a price of 20 or 25 a month; nothing else differs. Primary outcome: gross profit per sign-up over the first 12 months — revenue minus serving cost, discounted 1 % a month — so that conversion, churn and usage cost all count. Primary estimate : mean of the higher-price arm minus mean of the lower; positive favours the higher price. Smallest effect of interest: 5 per sign-up. Existing customers are not affected. Prices are disclosed before payment in both arms. The test is run and analysed by people independent of the pricing decision, and it is not a recommendation of any price.
Calculation
# Chapter 8. A pre-registered price test. New sign-ups are randomised to price
# A = 20 or price B = 25 a month. Primary outcome: gross profit per SIGN-UP
# over the first 12 months (revenue minus serving cost, discounted 1 % a
# month), so conversion, churn and usage cost all count. D = mean(B) -
# mean(A); POSITIVE D favours the higher price. Smallest effect of interest 5
# per sign-up. The standard deviation of the outcome is taken from a
# simulation of the book's own model: trial-to-paid conversion 0.30 at 20 and
# 0.27 at 25, churn c(p) from p15, serving cost per month lognormal with
# median 1.20 and log-sd 1.5 (p11). Two-sided alpha 0.05, power 0.80.
import math
import random
def signup(p, conv, rng):
if rng.random() >= conv:
return 0.0
c = 0.02 * math.exp((p - 20) / 15)
cost = 1.20 * math.exp(1.5 * rng.gauss(0, 1))
gp = 0.0
for t in range(12):
gp += (p - cost) / 1.01 ** t
if rng.random() < c:
break
return gp
SESOI = 5.0
def verdict(d, lo, hi, n_per_arm, planned, srm_p=0.5, price_shown_ok=True):
if not price_shown_ok:
return (
"D = {:+.2f}: NOT INTERPRETABLE - some sign-ups saw the other"
" arm's price".format(d)
)
if srm_p < 0.001:
return ("D = {:+.2f}: NOT INTERPRETABLE - arm sizes differ more than"
"\n randomisation allows".format(d))
if n_per_arm < 0.9 * planned:
return (
"D = {:+.2f}: NOT INTERPRETABLE - only {:,} of {:,} planned"
" sign-ups per arm".format(d, n_per_arm, planned)
)
if lo >= SESOI:
v = (
"the higher price earns more per sign-up,\n by at least the"
" SESOI"
)
elif lo > 0:
v = "the higher price earns more,\n not shown to reach the SESOI"
elif hi < 0:
v = "the higher price earns LESS per sign-up"
elif lo >= -SESOI and hi <= SESOI:
v = "no difference worth acting on"
else:
v = "unclear: the interval is too wide to decide"
return "D = {:+.2f} ({:+.2f} to {:+.2f}): {}".format(d, lo, hi, v)
if __name__ == "__main__":
rng = random.Random(58016)
a = [signup(20, 0.30, rng) for _ in range(100000)]
b = [signup(25, 0.27, rng) for _ in range(100000)]
ma, mb = sum(a) / len(a), sum(b) / len(b)
sd = math.sqrt(
(sum((x - ma) ** 2 for x in a) + sum((x - mb) ** 2 for x in b))
/ (len(a) + len(b) - 2)
)
print(
"Model means per sign-up: A {:.2f}, B {:.2f} (true D = {:+.2f});"
" pooled sd {:.1f}".format(ma, mb, mb - ma, sd)
)
n = math.ceil(2 * (1.959964 + 0.841621) ** 2 * sd ** 2 / SESOI ** 2)
print("Sign-ups needed per arm: {:,}".format(n))
print("\nDecision rule, fixed before the data are examined:")
for c in (
(9.1, 6.2, 12.0, n, n),
(3.0, 0.4, 5.6, n, n),
(0.8, -2.0, 3.6, n, n),
(-6.5, -9.3, -3.7, n, n),
(2.0, -6.0, 10.0, n, n),
(9.1, 6.2, 12.0, n, n, 0.0001),
(9.1, 6.2, 12.0, int(0.6 * n), n),
(9.1, 6.2, 12.0, n, n, 0.5, False),
):
print(verdict(*c))
print("\nThe example intervals above are invented, to"
" show what each branch prints.")
print("Secondary outcomes (conversion, 12-month"
" churn, serving cost per paying")
print("user) are reported with intervals and never change the verdict.")
# Model means per sign-up: A 49.85, B 56.39 (true D = +6.54); pooled sd 112.2
# Sign-ups needed per arm: 7,908
#
# Decision rule, fixed before the data are examined:
# D = +9.10 (+6.20 to +12.00): the higher price earns more per sign-up,
# by at least the SESOI
# D = +3.00 (+0.40 to +5.60): the higher price earns more,
# not shown to reach the SESOI
# D = +0.80 (-2.00 to +3.60): no difference worth acting on
# D = -6.50 (-9.30 to -3.70): the higher price earns LESS per sign-up
# D = +2.00 (-6.00 to +10.00): unclear: the interval is too wide to decide
# D = +9.10: NOT INTERPRETABLE - arm sizes differ more than
# randomisation allows
# D = +9.10: NOT INTERPRETABLE - only 4,744 of 7,908 planned sign-ups per arm
# D = +9.10: NOT INTERPRETABLE - some sign-ups saw the other arm's price
#
# The example intervals above are invented, to show what each branch prints.
# Secondary outcomes (conversion, 12-month churn, serving cost per paying
# user) are reported with intervals and never change the verdict.The book’s own model predicts , just above the smallest effect of interest, with a pooled standard deviation of 112.2 per sign-up — most sign-ups never convert and a few heavy users cost more than they pay, so the outcome is extremely dispersed. That dispersion is why the test needs 7,908 sign-ups in each arm.
The decision rule prints one line per branch. Three gates make the result uninterpretable rather than negative: a sample-ratio mismatch (arm sizes differing more than randomisation allows, the most common sign of a broken experiment (Kohavi et al. 2020)), some sign-ups seeing the other arm’s price, and fewer than 90 % of the planned sign-ups. A result in which the higher price earns less per sign-up is a possible and informative outcome, and the rule names it.
Note
What a result would and would not mean. It would describe sign-ups during the test period at these two prices. It would not say what existing customers would do under a price rise, nor what happens at prices not tested, nor whether the effect lasts beyond twelve months, where Chapter 1’s projection problem returns.
Summary
Four questions summarise the book: which churn and which model a lifetime value rests on, where NRR is in the life of the base, what cash trough the growth plan implies, and — for AI products — what the mean user costs at the batch and context actually served. One conflict the book identifies, between first-year revenue and value per sign-up, can be settled by a randomised price test whose primary outcome counts conversion, churn and serving cost together; it needs 7,908 sign-ups per arm because that outcome is so dispersed.
References
Fan, Y. (2026). Subscription Unit Economics, Net Revenue Retention and AI Inference Scaling Laws: Cohort Survival Mixtures, Negative-Churn Differential Equations, Token Compute Margins and Metcalfe Congestion Limits [订阅制单位经济学、净收入留存与人工智能推理标度律] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.24090164 · Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).