5D Library › Persuasive Copywriting Algebra, Loss-Framing Asymmetry and Checkout Micro-Architecture · 中文
Chapter 6. The Ceiling of the Funnel and the Best Price
A funnel is a product
From “saw the free report” to “paid”, conversion is the product of the pass rates of each stage. The program takes three illustrative stages: offer read 0.40, checkout opened 0.15, payment completed 0.55.
Calculation
# Chapter 6. p11: Where conversion can and cannot be gained.
# A funnel: report viewed -> offer read -> checkout opened -> paid.
# Overall conversion is the product of stage rates. A change that lifts
# one stage by a relative x lifts the total by exactly x; the ceiling for
# a stage is reaching 1. We list the maximum total lift available from
# each stage and the lift from realistic 10% stage improvements.
STAGES = (("offer read", 0.40), ("checkout opened", 0.15),
("paid", 0.55))
def total(rates):
out = 1.0
for r in rates:
out *= r
return out
if __name__ == "__main__":
rates = [r for _, r in STAGES]
base = total(rates)
print(" baseline conversion per report viewer: {:.4f}".format(base))
print(" stage rate ceiling lift +10% relative")
for i, (name, r) in enumerate(STAGES):
up = rates[:]
up[i] = 1.0
ten = rates[:]
ten[i] = min(1.0, r * 1.10)
print(" {:16s} {:.2f} {:10.2f} x {:12.4f}".format(
name, r, total(up) / base, total(ten)))
best = rates[:]
best = [min(1.0, r * 1.10) for r in best]
print(" all three +10%: {:.4f} ({:+.1%})".format(
total(best), total(best) / base - 1))
# baseline conversion per report viewer: 0.0330
# stage rate ceiling lift +10% relative
# offer read 0.40 2.50 x 0.0363
# checkout opened 0.15 6.67 x 0.0363
# paid 0.55 1.82 x 0.0363
# all three +10%: 0.0439 (+33.1%)Overall conversion is 0.0330. Any relative lift in one stage passes unchanged to the total: raising any one stage by a tenth gives 0.0363 overall. The ceilings differ: “offer read” can at most multiply the total by 2.50, “checkout opened” by 6.67 and “payment completed” by 1.82. The largest room is in “checkout opened” (the reader’s interest in the offer), not in the checkout interface of Chapter 3; the latter has room too, but a lower ceiling. Raising all three stages by a tenth lifts the total by 33.1%.
The best price and the sawtooth
Calculation
# Chapter 6. p12: The revenue-maximising price, with and without the
# left-digit effect. WTP lognormal (median RM 30, log-s.d. 1.0, as in
# p02); perceived price as in p03 with theta 0.3. Revenue per visitor =
# price x P(WTP > perceived price). Grid of whole-ringgit prices.
from math import log, erf, sqrt
def lead(p):
s = str(int(p))
return int(s[0]) * 10 ** (len(s) - 1)
def perceived(p, theta):
L = lead(p)
return L + (1 - theta) * (p - L)
def share(p, med=30.0, sd=1.0):
return 0.5 * (1 - erf((log(p) - log(med)) / sd / sqrt(2)))
if __name__ == "__main__":
for theta in (0.0, 0.3):
best = max(range(5, 301),
key=lambda p: p * share(perceived(p, theta)))
r = best * share(perceived(best, theta))
print(" theta {:.1f}: best price RM {} -> RM {:.2f} per visitor"
.format(theta, best, r))
print(" theta 0.3, revenue per visitor near the optimum:")
for p in (38, 39, 40, 41, 48, 49, 50):
print(" RM {:3d}: {:.3f}".format(p, p * share(perceived(p, 0.3))))
# theta 0.0: best price RM 41 -> RM 15.47 per visitor
# theta 0.3: best price RM 39 -> RM 16.55 per visitor
# theta 0.3, revenue per visitor near the optimum:
# RM 38: 16.418
# RM 39: 16.552
# RM 40: 15.472
# RM 41: 15.587
# RM 48: 16.210
# RM 49: 16.276
# RM 50: 15.237In the valuation model of Chapter 1, without a left-digit effect the price maximising revenue per visitor is RM 41 (RM 15.47). With a left-digit effect of the best price moves to RM 39 (RM 16.55), while RM 40 earns only RM 15.472: crossing a multiple of ten, revenue falls off a “cliff”. RM 49 at 16.276 is likewise clearly above RM 50 at 15.237. Revenue against price is therefore a sawtooth, with a local peak just below each multiple of ten.
Hypothesis · not established
The author’s hypothesis (not yet tested): for the author’s reading reports, the revenue difference between the two just-below prices RM 39 and RM 49 is smaller than that between RM 40 and RM 39. What would show it wrong: if, in a three-arm randomised test run as in Chapter 8, RM 40 and RM 39 do not differ significantly in revenue while RM 49 and RM 39 do, the hypothesis fails.
Summary
Overall conversion is a product of stages; lifting any stage by a tenth lifts the total by a tenth; the stage with the highest ceiling is “checkout opened”. The left-digit effect puts the best price just below a multiple of ten and turns the revenue curve into a sawtooth.
Fan, Y. (2026). Persuasive Copywriting Algebra, Loss-Framing Asymmetry and Checkout Micro-Architecture: Prospect-Theory Valuation Ratios, Numerical Specificity Credibility and Cognitive-Fluency Conversion Bounds [说服文案代数、损失框架不对称性与结账微界面工程] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.24090205 · Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).