关于「进度透明与剩余人工值守:催单模型与 Erlang C」,《零客服界面与自助修正状态机》给出的定量结论是什么?

结论:第六章显示可见进度节点从零到两个,每百单催问从 41.0 次降到 16.5 次,并用 Erlang C 说明小店省下的是忙碌时间而非值守人数。

数据口径:在合成的一个月里,可自助类别合计只占 0.869,退款、付款与杂项必须留给人;把完成率与新咨询的不确定性抽样进去,消灭率九成区间为 0.468 到 0.601,要达到 0.9 需要统一完成率 1.086,算术上做不到。 书中参数为示意,全部计算由书中程序生成;本书是非个人化经营建议,不构成法律意见。

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: 16.5, 41.0

nodeswidth_dchases/100 orderslate share
0241.00.350
2216.50.350
4210.30.350
627.40.350
0636.90.142
2614.90.142
469.10.142
666.70.142
查看完整验算程序
# Chapter 6. p61: chase messages versus visible progress nodes
# 2,000 orders per cell, lead time lognormal (median 10 days, log-sd
# 0.25). k visible progress nodes are spread evenly over the lead time;
# the promised window is centred on 10 days with width w days. Hourly
# chase hazard = 0.0003 x (hours since last news / 24), tripled once
# the promised window has closed (an unexplained late wait). A reply to
# a chase counts as news. Shape follows Maister 1985 and the boundary
# found by Buell and Norton 2011 (their study: 62 and 63 percent chose
# the transparent wait at 30 and 60 seconds, 42 and 23 percent the
# opaque one; Buell, Kim and Tsay 2017 report quality up 22.2 percent
# and throughput time down 19.2 percent with two-way transparency);
# all magnitudes here are illustrative.
import math
import random


def cell(k, w):
    rng = random.Random(40161)
    chases = 0
    late = 0
    end = (10 + w / 2) * 24
    for _ in range(2000):
        lead = 24 * 10 * math.exp(rng.gauss(0, 0.25))
        news = [lead * (j + 1) / (k + 1) for j in range(k)]
        if lead > end:
            late += 1
        last = 0
        for t in range(1, int(lead) + 1):
            if news and t >= news[0]:
                last = news.pop(0)
            h = 0.0003 * (t - last) / 24
            if t > end:
                h *= 3
            if rng.random() < h:
                chases += 1
                last = t
    return chases / 20, late / 2000


print("nodes  width_d  chases/100 orders  late share")
for w in (2, 6):
    for k in (0, 2, 4, 6):
        c, l = cell(k, w)
        print("{:>5}  {:>7}  {:>17.1f}  {:>10.3f}".format(k, w, c, l))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《日入百单零客服自愈界面、错填自助修正状态机与售后工单源头消灭母典》第 6 章「进度透明与剩余人工值守:催单模型与 Erlang C」(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-401 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026401,
  author    = {Fan, Yafu},
  title     = {Zero-Ticket Self-Repairing Order Interfaces, Self-Service Correction State Machines and Pre-Emptive Support Elimination},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论