关于「小红书:搜索意图、收藏与长尾检索,以及标题、关键词与封面的算术」,《五大平台算法与一鱼五吃》给出的定量结论是什么?

结论:第二章说明对准搜索需求的小红书笔记半年累计浏览是爆款笔记的 2.26 倍,长尾词组合的点击是大词的 2.80 倍,信息型封面的收藏是抢眼封面的 1.50 倍。

数据口径:书中把每条平台说法标为三类:带日期的官方披露、同行评审研究、本书的示意模型;书中的权重、概率与漏斗参数均为示意,不是任何平台的真实设定,全部计算由书中程序生成。 本书是一般性经营参考,不构成法律、税务或投资建议;不涉及假账号、机器人、互赞群、买互动、抓取个人资料或群发私讯;合作内容一律标明;达人酬劳只有单层直付,不涉及任何多层分润。

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: 2.26

cumulative views and the share that came from search
burst note day7: 5,713 views, search share 0.00
burst note day30: 6,291 views, search share 0.04
burst note day90: 6,891 views, search share 0.12
burst note day 1807,791 views, search share 0.22
guide note day7: 2,046 views, search share 0.07
guide note day30: 4,142 views, search share 0.49
guide note day90: 9,542 views, search share 0.78
guide note day 18017,642 views, search share 0.88
guide note overtakes burst note on day 57
day-180 ratio guide / burst2.26
查看完整验算程序
# Chapter 2. p21: feed burst against search long tail for two RedNote-style notes.
# Daily views = feed part F0 * exp(-t / tau) + search part S * min(1, t / ramp).
# The burst note gets a big feed start and little search demand; the guide note gets
# a small feed start but matches steady queries. Every parameter is illustrative;
# RedNote has not published decay constants or search shares for single notes.
import math

NOTES = {"burst note": {"F0": 2000.0, "tau": 2.5, "S": 10.0, "ramp": 14.0},
         "guide note": {"F0": 600.0, "tau": 3.0, "S": 90.0, "ramp": 14.0}}


def daily(n, t):
    feed = n["F0"] * math.exp(-t / n["tau"])
    search = n["S"] * min(1.0, t / n["ramp"])
    return feed, search


print("cumulative views and the share that came from search")
cum = {}
for name, n in NOTES.items():
    feed, search = [], []
    for t in range(180):
        f, s = daily(n, t)
        feed.append(f)
        search.append(s)
        cum.setdefault(name, []).append(math.fsum(feed) + math.fsum(search))
    for day in (7, 30, 90, 180):
        fs, ss = math.fsum(feed[:day]), math.fsum(search[:day])
        print("  %-10s day %3d: %8s views, search share %.2f" % (
            name, day, "{:,.0f}".format(fs + ss), ss / (fs + ss)))
cross = next(t for t in range(180) if cum["guide note"][t] > cum["burst note"][t])
print("guide note overtakes burst note on day %d" % (cross + 1))
print("day-180 ratio guide / burst: %.2f" % (cum["guide note"][-1] / cum["burst note"][-1]))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《五大社交图谱算法异质性、跨平台同源异构分发与达人聚拢母典》第 2 章「小红书:搜索意图、收藏与长尾检索,以及标题、关键词与封面的算术」(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-359 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026359,
  author    = {Fan, Yafu},
  title     = {Algorithmic Heterogeneity across RedNote, Threads, X, Instagram and Facebook: Cross-Platform Adaptation and Creator Aggregation},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论