关于「五分钟要几个人:排队、分级与值班人手」,《五分钟线索与防飞单》给出的定量结论是什么?

结论:五分钟承诺是一道排队题。每小时 18 条表单、首触平均 8 分钟时,九成达标需要 5 个人;投放高峰需要 6 个人,九成五要 7 个人。达标的代价是闲置,安静时段的利用率只有 0.2667。分级能用更少的人守住承诺:三个人做分级,价值已超过四个人做先来先服务;先来先服务要六个人,才追得上四个人的分级。

数据口径:「一百倍」出自 2007 年一份未经同行评审的行业研究,比的是几率而不是概率;2011 年《哈佛商业评论》短文报告的倍数较小,而且该刊是编辑审稿的管理杂志,作者之一与研究所用软件公司有利益关联,这是本书的蓝图修正。 书中参数为示意,全部计算由书中程序生成。

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: 0.2667, 18

A. share of leads touched within 5 minutes
lam/h load a c1 c=2 c=3 c=4 c=5 c=6
60.80 0.2940 0.8920 0.9868 0.9987 0.9999 1.0000
182.40 0.0000 0.0000 0.5552 0.8944 0.9776 0.9958
304.00 0.0000 0.0000 0.0000 0.0000 0.7034 0.9184
B. smallest desk for a target share within 5 minutes
target 0.80: lam 62, 18 -> 4, 30 -> 6
target 0.90: lam 63, 18 -> 5, 30 -> 6
target 0.95: lam 63, 18 -> 5, 30 -> 7
C. utilisation at the 0.90 desk (busy share of each person)
lam6: c = 3, utilisation 0.2667
lam 18: c5, utilisation 0.4800
lam 30: c6, utilisation 0.6667
查看完整验算程序
# Chapter 3. p31: how many people does a five-minute promise need?
# Model: M/M/c queue (Erlang C). Leads arrive at rate lam per hour; a
# first touch (read the form, call or message, log the outcome) takes
# an exponential time with mean 8 minutes; c people are on the desk.
# P(wait <= t) = 1 - C(c, a) exp(-(c mu - lam) t), a = lam / mu, where
# C is the Erlang C probability of waiting. Arrival rates of 6, 18 and
# 30 leads an hour stand for a quiet hour, a normal peak and a campaign
# peak; all illustrative, not MPE data.
import math

HANDLE_MIN = 8.0
TARGET_MIN = 5.0


def erlang_b(c, a):
    b = 1.0
    for k in range(1, c + 1):
        b = a * b / (k + a * b)
    return b


def erlang_c(c, a):
    if a >= c:
        return 1.0
    b = erlang_b(c, a)
    return c * b / (c - a * (1.0 - b))


def p_within(c, lam_h, t_min=TARGET_MIN, h_min=HANDLE_MIN):
    mu = 60.0 / h_min                 # touches per hour per person
    a = lam_h / mu
    if a >= c:
        return 0.0
    pw = erlang_c(c, a)
    return 1.0 - pw * math.exp(-(c * mu - lam_h) * t_min / 60.0)


def staff_for(lam_h, goal):
    c = 1
    while p_within(c, lam_h) < goal:
        c += 1
    return c


if __name__ == "__main__":
    print("A. share of leads touched within 5 minutes")
    print("  lam/h  load a   c=1     c=2     c=3     c=4     c=5     c=6")
    for lam in (6.0, 18.0, 30.0):
        a = lam * HANDLE_MIN / 60.0
        cells = ["{:.4f}".format(p_within(c, lam)) for c in range(1, 7)]
        print("  {:>4.0f}   {:>5.2f}  ".format(lam, a) + "  ".join(cells))
    print("B. smallest desk for a target share within 5 minutes")
    for goal in (0.80, 0.90, 0.95):
        row = [staff_for(lam, goal) for lam in (6.0, 18.0, 30.0)]
        print("  target {:.2f}: lam 6 -> {}, 18 -> {}, 30 -> {}".format(
            goal, *row))
    print("C. utilisation at the 0.90 desk (busy share of each person)")
    for lam in (6.0, 18.0, 30.0):
        c = staff_for(lam, 0.90)
        print("  lam {:>2.0f}: c = {}, utilisation {:.4f}".format(
            lam, c, lam * HANDLE_MIN / 60.0 / c))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《高意向线索黄金五分钟转化、线下派单防飞单审计与转介绍佣金锁死母典》第 3 章「五分钟要几个人:排队、分级与值班人手」(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-348 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026348,
  author    = {Fan, Yafu},
  title     = {Five-Minute High-Intent Lead Conversion, Offline Deal-Leakage Auditing and Referral Commission Lock-In},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论