关于「搜索需求的幂律:排名、销量与长尾份额」,《蓝海词与主题合集》给出的定量结论是什么?

结论:排名与销量之间是幂律,排名每一步都很贵;长尾份额对指数极其敏感,指数从 0.871 升到 1.2,十万名以后的份额就从 0.411 降到 0.046。搜索词需求也是幂律,前二十个短语拿走一半以上查询,估计指数时要用最大似然,不要用对数坐标上的直线。两条幂律相乘,结论是:小作者不该追大词,应该追「有人搜、但前排不拥挤」的词。

数据口径:对小作者,关键词带来的期望销量约等于搜索需求、进入前几位的概率、点击率与转化率之积;大词的进前位概率近乎为零,示意数据里周搜索两万次的大词只带来约 0.0254 本周销量,冷僻短语却有 3.1445 本。 核对原文后,本书做了蓝图修正:排名换销量的指数是示意参数而非定值;KDP 关键词规则以官方帮助页为准,每本书至多七个关键词,不得用与本书无关的作者名或自己不拥有的品牌;AI 生成内容须如实披露;本书不引用艾瑞里的任何研究。

English version

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: 0.046, 0.411, 0.871

A. weekly sales by rank, ln Q = 10.526 - 0.871 ln r
rankweekly salessales per day
137,272.15,324.58
105,016.3716.61
100675.196.45
1,00090.912.98
10,00012.21.75
100,0001.60.24
1,000,0000.20.03
B. share of all sales from titles ranked below the cut
exponent bbeyond 1,000beyond 100,000
0.8710.7450.411
1.0000.5170.220
1.2000.1880.046
C. rank 1 versus rank 2: ratio 1.829
rank 5,000 to rank 500 needs 7.43 times the sales
查看完整验算程序
# Chapter 1. p11: rank-to-sales power law and the share of the long tail.
# Model: weekly sales Q(r) = exp(a) * r^(-b) for a title at sales rank r.
# a = 10.526 and b = 0.871 are the illustrative log-linear parameters of
# the kind reported for Amazon book ranks; b is then swept to show how
# strongly the tail share depends on it. Catalogue of N titles; the sum
# over ranks uses exact terms up to R0 and the Euler-Maclaurin integral
# beyond. All figures illustrative, not any real title.
import math

A, B = 10.526, 0.871
N = 3_000_000
R0 = 20_000


def weekly(rank, a=A, b=B):
    return math.exp(a - b * math.log(rank))


def partial(lo, hi, b):
    # sum of r^-b for r = lo..hi (integers), exact part plus integral
    if hi <= R0:
        return math.fsum(r ** -b for r in range(lo, hi + 1))
    head = math.fsum(r ** -b for r in range(lo, R0 + 1)) if lo <= R0 else 0
    start = max(lo, R0 + 1)
    if abs(b - 1.0) < 1e-12:
        tail = math.log(hi) - math.log(start)
    else:
        tail = (hi ** (1 - b) - start ** (1 - b)) / (1 - b)
    tail = math.fsum([tail, 0.5 * (start ** -b + hi ** -b)])
    return head + tail


def tail_share(cut, b):
    total = partial(1, N, b)
    return partial(cut + 1, N, b) / total


if __name__ == "__main__":
    print("A. weekly sales by rank, ln Q = 10.526 - 0.871 ln r")
    print("        rank   weekly sales   sales per day")
    for r in (1, 10, 100, 1_000, 10_000, 100_000, 1_000_000):
        q = weekly(r)
        print("  {:>10,}  {:>13,.1f}  {:>14,.2f}".format(r, q, q / 7))
    print("B. share of all sales from titles ranked below the cut")
    print("  exponent b   beyond 1,000   beyond 100,000")
    for b in (0.871, 1.0, 1.2):
        print("  {:>10.3f}   {:>12.3f}   {:>14.3f}".format(
            b, tail_share(1_000, b), tail_share(100_000, b)))
    q1 = weekly(1)
    q2 = weekly(2)
    print("C. rank 1 versus rank 2: ratio {:.3f}".format(q1 / q2))
    print("   rank 5,000 to rank 500 needs {:.2f} times the sales".format(
        weekly(500) / weekly(5_000)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《百卷数字专著蓝海词卡位、高客单主题合集直售与全自动知识收银母典》第 1 章「搜索需求的幂律:排名、销量与长尾份额」(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-356 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026356,
  author    = {Fan, Yafu},
  title     = {Niche Keyword Positioning for Digital Monograph Catalogues, High-Ticket Thematic Bundle Direct Sales and Automated Knowledge Cashflow},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论