关于「消灭率的上界:类别份额乘自助完成率」,《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:消灭率等于各可自助类别份额乘完成率之和,减去自助流程新引出的咨询。本例中可自助份额是 0.869,这是天花板;把完成率与新咨询的不确定性都抽样进去,九成区间是 0.468 到 0.601。九成之所以不可达,是因为退款、付款与杂项必须留给人,而这一块占了 0.131。

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: 0.131, 0.869

tickets 366 self-serviceable share 0.869
human-only share (refund, payment, other) 0.131
g low mid high perfect
0.000.441 0.581 0.722 0.869
0.050.397 0.538 0.679 0.825
0.100.354 0.494 0.635 0.782
0.90 needs self-serviceable share >0.90 even at completion 1;
here the perfect-completion ceiling is 0.869
human tickets left per 100 orders (12.20 before):
low E0.397 -> 7.36
high E0.679 -> 3.92
查看完整验算程序
# Chapter 2. p21: the ceiling on eliminable tickets
# E = sum over self-serviceable classes of share x completion rate,
# minus g x (eligible share) for new contacts that self-service flows
# create themselves. Shares are the p11 month; completion ranges and g
# are illustrative.
import math

# class, tickets in the p11 month, low, high completion (0 = human)
ROWS = [("not_received", 98, 0.60, 0.90),
        ("wrong_input", 55, 0.55, 0.85),
        ("progress", 73, 0.40, 0.80),
        ("repeat_chase", 38, 0.50, 0.85),
        ("link_expired", 31, 0.70, 0.95),
        ("content_q", 23, 0.10, 0.40),
        ("refund", 18, 0.0, 0.0),
        ("payment", 12, 0.0, 0.0),
        ("other", 18, 0.0, 0.0)]
N = int(math.fsum(r[1] for r in ROWS))


def elim(pick, g=0.0):
    terms = []
    elig = []
    for name, k, lo, hi in ROWS:
        s = k / N
        c = pick(lo, hi)
        terms.append(s * c)
        if hi > 0:
            elig.append(s)
    return math.fsum(terms) - g * math.fsum(elig)


elig_share = math.fsum(k / N for _, k, lo, hi in ROWS if hi > 0)
print("tickets {}  self-serviceable share {:.3f}".format(N, elig_share))
print("human-only share (refund, payment, other) {:.3f}".format(
    1 - elig_share))
print("g     low    mid    high   perfect")
for g in (0.0, 0.05, 0.10):
    lo = elim(lambda a, b: a, g)
    mid = elim(lambda a, b: (a + b) / 2, g)
    hi = elim(lambda a, b: b, g)
    top = elim(lambda a, b: 1.0 if b > 0 else 0.0, g)
    print("{:.2f}  {:.3f}  {:.3f}  {:.3f}  {:.3f}".format(
        g, lo, mid, hi, top))
print("0.90 needs self-serviceable share >= 0.90 even at completion 1;")
print("here the perfect-completion ceiling is {:.3f}".format(
    elim(lambda a, b: 1.0 if b > 0 else 0.0)))
print("human tickets left per 100 orders (12.20 before):")
for label, f in (("low", lambda a, b: a),
                 ("high", lambda a, b: b)):
    e = elim(f, 0.05)
    print("  {:<4} E={:.3f} -> {:.2f}".format(label, e, 12.20 * (1 - e)))

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 2 章「消灭率的上界:类别份额乘自助完成率」(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