关于「对话归因:末次点击、Shapley 值与被切断的来源」,《点击直达对话的成交与归因》给出的定量结论是什么?
结论:第三、四章说明,末次点击记给最后触点的功劳可达 Shapley 份额的 2.36 倍,来源被切断时广告被多记到 1.20 倍,预填参考代码降到 1.06 倍;冷受众测出一成增量,每组要 822,401 人,对话层面每组也要 8,158 段。
数据口径:书中的转化率、时延、回复率、样本量与金额均为示意参数,全部计算由书中程序生成;本书是非个人化经营建议,不构成法律或营销合规意见,不承诺任何收入;只写有同意记录的触达,语音条只由真人本人录制,收款只用持牌支付机构的公开功能,紧迫感只来自真实库存与真实截止,推荐奖励只给直接推荐人一层。 第一、二章推导摩擦损耗减半的条件:示意参数下直跳后首句发出率须达 0.650;直跳让对话数涨到 2.500 倍,高意向占比从 0.571 跌到 0.327,每段接待成本 4.94 单位时两条路径打平。
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: 1.06, 1.20
| source | true share | last-click log | with prefilled code |
|---|---|---|---|
| ad | 0.455 | 0.547 | 0.475 |
| creator | 0.347 | 0.182 | 0.301 |
| poster | 0.198 | 0.138 | 0.185 |
| unknown | 0.000 | 0.134 | 0.040 |
| ad over-credit: 1.20 times without code, 1.06 with code | |||
| chats simulated: 20,000 |
查看完整验算程序
# Chapter 3. p32: how device switches, forwarding and screenshots cut the
# source of a chat, and how much a prefilled reference code recovers.
# True origins of 20,000 chats: click-to-chat ad 0.45, creator post 0.35,
# printed QR poster 0.20. Chance the source tag is lost: ad 0.03; creator
# 0.15 device switch + 0.10 forwarding; poster 0.20 screenshot + 0.10
# forwarding. 0.30 of creator-origin chats also clicked an ad later, so a
# last-click log credits the ad. A prefilled code survives with 0.70.
# Lost sources fall into "unknown". Fixed seed; all rates illustrative.
import collections
import random
N = 20000
ORIGIN = (("ad", 0.45), ("creator", 0.35), ("poster", 0.20))
LOSS = {"ad": 0.03, "creator": 0.25, "poster": 0.30}
LATER_AD = 0.30
KEEP_CODE = 0.70
def draw(rng):
names = [name for name, _ in ORIGIN]
return rng.choices(names, weights=[p for _, p in ORIGIN])[0]
def simulate(code):
rng = random.Random(20261120)
true, seen = collections.Counter(), collections.Counter()
for _ in range(N):
o = draw(rng)
true[o] += 1
lost = rng.random() < LOSS[o]
later = o == "creator" and rng.random() < LATER_AD
kept = code and rng.random() < KEEP_CODE
if kept:
seen[o] += 1
elif later:
seen["ad"] += 1
elif lost:
seen["unknown"] += 1
else:
seen[o] += 1
return true, seen
t0, s0 = simulate(False)
t1, s1 = simulate(True)
print("source true share last-click log with prefilled code")
for k in ("ad", "creator", "poster", "unknown"):
print("%-9s %10.3f %15.3f %20.3f" % (k, t0[k] / N, s0[k] / N,
s1[k] / N))
print("ad over-credit: %.2f times without code, %.2f with code" % (
s0["ad"] / t0["ad"], s1["ad"] / t1["ad"]))
print("chats simulated: %s" % format(N, ","))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《东南亚即时通讯成交漏斗、点击直达对话归因与语音权威收口母典》第 3 章「对话归因:末次点击、Shapley 值与被切断的来源」(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-366 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026366,
author = {Fan, Yafu},
title = {Click-to-Messenger Conversion Funnels, Chat Attribution and Voice-Note Closing in Southeast Asian Commerce},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}