初创公司尽调
尽调报告 Infrastructure / Product Analytics Series E 2026-05-24

PostHog

开发者平台很强,但公开分母仍不足以支撑最新私募价格

PostHog 看起来已是实打实的多产品开发者平台,但仅凭公开证据,仍难支撑以有纪律的方式支付据称 $1.4B 的价格。

封面要素

估值 01
1400 USD M [CV001]
收入下限 02
>$50M [CV004]
2026 年收入目标 03
100 USD M [CV005]
阶段 04
Series E [CV001]
成立时间 05
2020-01-23 [CO003]
付费产品 06
10+ [CO007]
免费公司占比 07
90%+ [CV007]

公司概况

PostHog 是一家私营、远程优先的软件公司,由 James Hawkins 和 Tim Glaser 于 2020 年创立。公司从开源产品分析工具起步,已经扩展成更广的开发者优先平台,覆盖分析、回放、功能开关、实验、类数据仓库工具、CDP 式路由和 AI 辅助工作流。公开证据支持其广泛的自助式、工程驱动分发动作,但还不足以给出承销最新私募估值所需的完整财务分母。

官网
posthog.com
成立时间
2020-01-23
创始人
James Hawkins, Tim Glaser
创立地点
San Francisco, CA
总部
San Francisco, CA (remote-first)
产品
PostHog 面向产品工程师销售一个按用量计费、对开源友好的平台,把产品分析、网站分析、会话回放、功能开关、实验、问卷、数据仓库能力、CDP 式路由和 AI 辅助开发者工作流合在一起。
客户
工程驱动型创业公司、开发者工具公司、产品团队,以及希望采用一体化产品智能栈的相邻增长 / 数据职能。
商业模式
开源核心 / 用量计费 SaaS 订阅
阶段
Series E
融资情况
官方披露 2025 年完成 $70M Series D,随后独立报道称公司在 2025 年晚些时候以 $1.4B 估值完成 $75M Series E。
[CO003, CO004, CO023, CE001, CI001, CV001]

执行摘要

主要优势

  • 开发者优先的一体化平台,覆盖分析、回放、功能开关、实验和数据工具。
  • 透明的按用量计费与开源分发,支撑较强的 PLG 触达。
  • 具名客户案例展示了真实 ROI,也证明其能嵌入工程驱动型产品的日常运营。

主要风险

  • 公开分母质量太弱,无法有把握地承销最新私有估值。
  • 2025-2026 年事故记录抬高了信任风险和企业采用风险。
  • 免费用户基数很大,转化与扩张质量比客户 logo 数量更关键。

未决问题

  • 当前 ARR 或确认收入、毛利率,以及云端与自托管收入结构。
  • NRR、logo 留存、免费转付费转化率和产品附加率。
  • 股权结构表、清算优先权,以及 2025 年各轮融资中新股与老股的拆分。

目录

Chapter 01

01公司概况

1.1 身份、平台边界与公开定位

PostHog 自己的公开页面已经呈现出一个远比 2020 年起步时的开源分析工具更宽的业务。关于页、产品页、定价页和战略页都把公司框定为开发者优先平台:产品分析、会话回放、功能开关、实验、问卷、类数据仓库基础设施、AI 可观测性和工作流自动化被放在同一个地方。这个统一定位很关键,因为它解释了公司的分发模型和资本故事:PostHog 想先用慷慨的免费层切入工程师,再把原本分散在多个点状工具上的预算整合过来。公司也把透明度当作产品和 GTM 资产来经营,明确称代码库开源、手册公开、定价完全发布。后续章节可据此建立基线:PostHog 不是小众分析厂商,而是一个更宽的开发者基础设施平台,目标是成为产品和客户上下文的事实记录系统。[CO001, CO003, CO005, CO006, CO007, CO008]

FO002: 公司画像逻辑

PostHog 把开源透明度、产品广度、定价和开发者优先分发连成一套公司逻辑。

[CO005, CO006, CO008, CO009, CO036, CO040]

1.2 创始人、运营模式与治理可见度

公开创始人叙事异常清晰,尽管更完整的治理图景并不透明。PostHog 的手册、Y Combinator 档案和第三方研究都指向同一个事实:James Hawkins 和 Tim Glaser 于 2020 年 1 月创办公司,Hawkins 为 CEO,Glaser 为 CTO。Contrary 的备忘录补充了有用背景:两人都曾在 Arachnys 工作,商业与技术经验互补,解释了为什么 PostHog 既会讲强产品故事,又保持很深的技术取向。运营模式同样明确。招聘页和外部远程工作资料显示,公司围绕小型自治团队、异步优先习惯和全球分布式招聘搭建,而不是围绕办公室管理。与此同时,公开治理披露比创始人披露薄得多。除创始人和运营副总裁 Charles Cook 等少数签字人外,审阅材料中没有同样可见的公开董事会页面或完整高管名单。尽调上,这意味着关键人依赖是真实风险,正式治理需要直接核验,不能只从品牌强度推断。[CO002, CO003, CO004, CO010, CO011, CO012]

领导层和创始人表
人员公开角色背景 / 证据覆盖强度关键人依赖
James Hawkins联合创始人,CEOHandbook、YC 和第三方研究都把 Hawkins 与创立、战略、融资和公开叙事联系起来。Contrary 称他此前曾升任 Arachnys 销售副总裁。
Tim Glaser联合创始人,CTOHandbook、YC 和第三方研究将 Glaser 标识为技术联合创始人和架构负责人;Contrary 将他与 Arachnys 的产品和 R&D 工作联系起来。
Charles Cook运营副总裁公开 DPA 预览将 Charles Cook 列为运营副总裁和签署高管,是所审材料中少数清晰的非创始人领导层引用之一。有限

公开领导层披露并不完整:创始人非常可见,但董事会构成和完整高管名单没有同等公开。

[CO003, CO004, CO010, CO011, CO012, CO016]

1.3 融资历史、投资人图谱与资本策略

PostHog 的公开资本历史显示,公司从种子轮到独角兽估值爬升得异常快,同时把主要融资故事绑定在产品宽度和创始人控制上。手册时间线记录了 2020 年 4 月 $3.025M 种子轮、2020 年 12 月 GV 领投的 $9M Series A,以及 2021 年 6 月 Y Combinator 领投的 $15M Series B。更近的资本形成则由官方 Series D 文章和 2025 年新闻报道支撑。2025 年 6 月,PostHog 称其完成 Stripe 领投的 $70M 融资,估值 $920M,同时提到较小的 Series C 式新股部分和扩大后的员工流动性安排。到 2025 年 9 月,独立报道称 Peak XV 领投 $75M Series E,估值 $1.4B,把 PostHog 推入独角兽行列。反复出现的投资人名字很重要:YC、GV、Stripe、Peak XV、1984 Ventures 和 Formus Capital 都是可信度锚点。公司的招聘页也强调老股交易和要约回购,说明管理层不只用融资轮支持扩张,也在降低员工对近期退出的压力。[CO018, CO019, CO020, CO021, CO022, CO023]

利益相关方或投资者地图
利益相关方在股权结构或分发中的角色重要性公开证据尽调问题
Y Combinator加速器、早期投资者、后续 Series B 领投方提供进入创业公司生态的早期分发,并在多个时间点提供验证。手册记录种子支持;公司资料和 Series B 时间线公开可见。核实现有持股,以及 YC 是否仍有董事会席位或观察员权利。
GVSeries A 领投方和持续投资者体现其早期对开发者基础设施 / 分析投资逻辑的强信心。手册记录 GV 领投 Series A,并有后续参与引用。确认 2025 年两轮融资后的当前持股和治理角色。
1984 Ventures早期种子投资方公司最早绑定的外部基金之一。手册种子轮时间线和 1984 投资组合页面。澄清该基金是否仍持有有意义股份。
Stripe2025 年 Series D 领投方锚定 $920M 估值跃升,并支持员工流动性。官方 Series D 文章加后续新闻报道。了解 Stripe 关系是纯财务,还是也有战略 / 商业含义。
Peak XV2025 年 Series E 领投方帮助 PostHog 推至 $1.4B 独角兽估值,并释放成长期投资需求信号。Peak XV 投资组合页面加独立融资报道。核实董事会权利、清算优先权和任何地域扩张预期。
Formus Capital2025 年融资参与者中的现有投资者显示后续融资中支持的连续性。官方 Series D 公告。确认持股规模,以及支持是新股、老股还是两者都有。

这是一张基于公开可见性的利益相关方地图,不是完整股权结构表。它突出官方和新闻来源中反复点名的基金和相关方。

[CO018, CO019, CO020, CO021, CO022, CO023]
里程碑表
日期事件类型金额 / 状态参与方含义
2020-01-23PostHog 创立创立公司成立James Hawkins;Tim Glaser几次上线前转向之后,当前公司起步。
2020-02MVP 在四周编码后发布到 Hacker News产品几天内 300 次部署创始人;早期开源社区验证快速发货和早期开发者拉力。
2020-04种子轮融资完成融资$3.025M 种子轮YC Continuity;1984 Ventures公司在 YC 之后获得早期资本。
2020-12Series A 公布融资$9M Series AGV 领投标志早期机构化扩张。
2021-06Series B 公布融资$15M Series BY Combinator 领投确认重复投资者支持。
2022-12管理层报告收入增长 6x,并以 $10M ARR 为目标、毛利率目标 70%规模战略里程碑公司管理层显示公司从产品市场契合转向财务效率。
2024首次员工老股交易完成治理流动性事件员工;管理层表明公司愿意将融资用于员工流动性,而不仅是新股资本。
2025-06Series D 加小规模 Series C 式新股部分融资$70M,估值 $920M;约 $10M 新股资本投资方:Stripe;YC;GV;Formus Capital估值跃升的轮次,同时扩大员工流动性和创始人控制权。
2025-08-15披露安全公告 PSA-2025-00001反向中等严重度;已解决PostHog 安全团队公开提醒:透明度也包括承认授权漏洞。
2025-09-30Series E / 独角兽公告融资$75M,估值 $1.4BPeak XV 和现有投资者设定进入 2026 年时的当前公开估值锚。
2026-02-20发布日志数据丢失公开事故复盘反向影响客户的事故PostHog 工程团队为后续章节提供具体运营风险标记。
2026-04-27发布工作流事故公开事故复盘反向影响客户的事故PostHog 工程团队显示透明度延续到 2026 年,但也说明可靠性问题仍然存在。

这是基础里程碑的公开记录时间线。它有意混合融资、规模、治理和反向事件,因为后续章节需要一条可复用时间线,而不是重复的年表。

[CO017, CO018, CO019, CO020, CO021, CO023]
FO001: 公司里程碑时间线

公开里程碑显示,公司从 YC 阶段发布到独角兽融资推进很快,同时仍保持可见的运营事故披露。

[CO017, CO018, CO019, CO020, CO023, CO037]

1.4 规模信号、披露缺口与后续需要带入的问题

公开规模信号方向上很强,但仍不够一致,后续章节应把多个指标当作区间,而不是定论。积极面是,PostHog 关于页称平台被 190,254+ 个团队使用,定价页称超过 90% 的公司免费使用,未来页称管理层希望到 2026 年底实现 $100M 年收入。手册也展示了历史招聘和收入扩张,包括 2022 年收入增长 6x,员工数从 2021 年中 25 人增长到 2022 年底 38 人。但当前规模图景没有被干净交叉验证。招聘页和人员页暗示公司有 200+ 人,而外部远程工作资料称团队更接近 110 人。ARR 同样混杂:Sacra 发布了 2026 年估计,但审阅过的官方页面没有披露权威当前 ARR。最后,PostHog 的透明度有利也有弊:公司公开列出事后复盘和安全公告,文化上是正面信号,但也留下具体风险标记,包括一个已解决的 2025 年查询可见性安全公告,以及多个 2025-2026 年面向客户的事故,值得在风险章节单独处理。[CO002, CO007, CO024, CO025, CO026, CO027]

快照 KPI 表
指标公开信号来源时间窗可信度限制
创立日期2020-01-232020-2026 年公开来源手册与 YC 来源口径稳定。
总部 / 法律主体地址注册地址:2261 Market Street #4008, San Francisco, CA当前法律页面这是法律 / 控制者地址;运营模式是远程优先,而不是以办公室为中心。
创始人James Hawkins(CEO)与 Tim Glaser(CTO)2020-2026 年公开来源更广泛管理层披露弱于创始人披露。
产品宽度10+ 个付费产品,覆盖分析、回放、功能开关、实验、数据仓库、AI 和工作流2026 年网站状态产品数量为公司口径,可能很快变化。
最新披露估值$1.4B2025 年新闻报道由 2025 年独立报道支持,不是公开股权结构表文件。
上一轮披露估值$920M2025 年官方融资文章官方 Series D 文章把它描述为接近独角兽,而不是完整公开股权结构细节。
定价模式用量计费,免费层很大;据称 90%+ 公司免费2026 年定价页面公司披露的采用组合。
客户 / 使用信号190,254+ 个团队;每届 YC 中有 65%2026 年关于页面公司口径,且时间敏感。
当前员工数信号官方页面暗示 200+ 人;外部远程工作资料称约 110 人2026 年混合来源公开来源分歧明显,因此应使用区间而非点估计。
ARR 披露2026 年目标为 $100M 年收入;当前 ARR 仍未官方披露2026 年手册和第三方估计存在第三方 ARR 估计,但不是公司权威披露。

汇总官方、合作伙伴和独立来源。公司未发布权威 KPI 的地方,本表保留区间或空值式限制说明,而不是强行填一个数字。

[CO002, CO003, CO004, CO007, CO023, CO024]
FO003: 披露与经营信号快照

公开快照混合增长、流动性和事故透明度信号,而不是复印 KPI 表。

[CO026, CO027, CO028, CO031, CO037, CO038]

1.5 展示要点

Chapter 02

02市场分析

2.1 市场边界、纳入支出与替代方案

定义 PostHog 市场的最干净方式,是把它看作开发者优先的产品智能栈,而不只是单一分析品类。PostHog 自己的产品、Product OS、CDP 和竞品比较页面显示,公司想捕获过去分散在网站分析、产品分析、会话回放、功能管理、实验、类数据仓库集成和行为激活工具上的预算。因此,最相关的市场边界应包括用于埋点用户行为、分析行为、在产品内采取动作,并把数据路由到运营系统的软件。广义商业智能和通用营销云不应纳入,除非这些系统被拿来绕路承担产品埋点或客户旅程分析。最常见的现状替代仍是一套拼接栈:GA4 负责流量,Amplitude 或 Mixpanel 负责产品分析,FullStory 负责回放,LaunchDarkly 或 Statsig 负责功能开关,再加上独立的数据或激活管道。PostHog 的机会来自把这些品类压缩成一个工程友好的控制平面。[CM001, CM002, CM003, CM004, CM005, CM006]

市场定义表
品类 / 支出桶是否纳入 PostHog 市场边界?典型买方重要性现状替代方案
产品分析是 - 核心产品、工程、数据跟踪激活、参与度、留存、漏斗和生命周期健康度Mixpanel、Amplitude、Heap
Web 分析 / 流量分析是 - 相邻核心增长、营销、产品通常是第一层行为数据,也是预算的常见切入点GA4 及类似纯 Web 工具栈
实验和功能管理是 - 核心相邻工程、产品、增长把发布控制与 KPI 验证、上线安全连接起来LaunchDarkly、Optimizely、Statsig、VWO
会话回放 / 定性 UX是 - 核心相邻UX、支持、工程、产品解释用户为何流失,并减少调试时间FullStory 和回放优先工具
CDP / 行为激活 / 数据仓库路由是 - 相邻扩张营销运营、销售运营、数据、工程把产品事件转为下游动作,并减少集成蔓延独立 CDP 和 ETL / 反向 ETL 工具
通用 BI 和广义营销云部分纳入 / 大多排除财务、RevOps、分析负责人只有当买方把它们当作产品埋点或分群替代品时才相关用作权宜方案的数据仓库仪表盘或营销套件

定义 PostHog 作为产品智能栈的边界。它有意纳入相邻品类,因为买方经常把多种工具拼在一起完成同一项工作。

[CM001, CM002, CM003, CM004, CM005, CM006]
FM004: 市场边界图

PostHog 的竞争发生在几个历史上分离的工作流品类交汇处。

[CM001, CM004, CM005, CM026, CM027, CM037]

2.2 规模测算视角:先看产品分析,再看更广的数据工具

公开市场规模来源没有收敛到一个精确数字,但方向一致:产品分析已经是有意义的软件品类,仍在以健康的双位数速度增长。Grand View、Expert Market Research 和 Mordor 都把当前市场描述为低双位数到高双位数十亿美元级别,预测 CAGR 从大约中双位数到约 20% 不等。这些差异更多来自方法和年份,而不是对方向的分歧。Grand View 采用更宽的 2024 年基数,给出最高当前值;Expert 使用 2025 年数值并拉长预测周期;Mordor 则锚定 2026 年市场规模估计,并按企业规模和部署方式做额外拆分。StartUs 提供了有用的外沿边界:更宽的高级分析相邻市场远大于产品分析本身,这很重要,因为 PostHog 正在向数据仓库、激活和工作流用例扩张。另一条 SAM 视角来自开发者人口数据:SlashData 估计全球有 48.4 million 名开发者,提醒我们开发者中心型分析和实验的买方基础远大于传统只面向产品经理的工具。[CM010, CM011, CM012, CM013, CM014, CM015]

TAM / SAM / SOM 或规模测算视角表
发布方 / 视角年份地理范围数值增长 / 份额方法说明可信度局限
Grand View Research 产品分析市场2024 年基准 / 2030 年预测全球当前 USD 19.92B;预测 USD 58.78B19.8% CAGR,2024-2030带细分详情的广义第三方产品分析市场规模测算本组中最高的当前估计;不是 PostHog 专属切片
Expert Market Research 产品分析市场2025 年基准 / 2035 年预测全球当前 USD 12.03B;预测 USD 49.09B15.1% CAGR,2026-2035长周期第三方市场预测预测期更长,相比 2030 和 2031 年估计的可比性较弱
Mordor Intelligence 产品分析市场2026 年估计 / 2031 年预测全球2026 年 USD 13.04B;2031 年 USD 25.73B14.55% CAGR,2026-2031包含部署方式和企业规模拆分当前数值窄于 Grand View,基准年也不能直接比较
StartUs 高级分析相邻市场2025 / 2029 年预测全球当前 USD 57.01B;预测 USD 139.92B25.2% CAGR,2025-2029更宽的高级分析相邻市场,用作外层市场边界远宽于 PostHog 可直接触达的品类
SlashData 开发者人口Q3 2025全球48.4M 开发者n/a以开发者人口衡量技术可触达基数人口不是软件支出数字,若无定价假设,不能转换为收入
Mordor 企业 / 部署拆分2025-2026全球大型企业份额 60.18%;云份额 87.6%中小企业 CAGR 19.7%更适合判断细分方向,而不是总体 TAM份额数据说明谁在购买,但不是 PostHog 可直接套用的 SOM

这张表保留不同的市场规模口径,不强行拼出一个合成 TAM。方法论列刻意把收入估算和人群估算拆开。

[CM010, CM011, CM012, CM013, CM014, CM015]
FM001: 市场估算区间

当前公开产品分析市场估算因方法不同而分散,但都指向一个已达 100 亿至接近 200 亿美元量级的品类。

这些数值来自不同基准年份和方法的点估算,这里作为可比公开区间展示,而不是数学上调和后的中点。

[CM010, CM011, CM012, CM013, CM036]

2.3 买方、用户、付费方与采用路径

买方证据指向多线程采购动作,而不是单一画像。PostHog 的客户故事显示,工程和产品团队经常主导采用,因为他们负责埋点、漏斗、实验和发布质量。但同一批案例也显示,平台嵌入后,领导层、UX、营销和数据角色都会成为相邻用户。Y Combinator 在 Startup School 和 Co-Founder Matching 中使用 PostHog,领导层、产品和工程都参与其中。Hasura 从工程主导的新用户引导分析起步,随后扩展到 UX 和营销。竞品和相邻平台页面进一步强化了这个模式:VWO 明确面向产品经理、工程师、增长营销和 UX/分析团队;Statsig 的客户引语来自数据工程、CTO、PM 和工程角色;Forrester 认为功能管理仍由开发者主导,而实验越来越服务产品和体验设计人群。实际含义是,初始用户常常是技术角色,但随着更多团队消费数据或使用控制能力,预算和续约逻辑会变成交叉职能问题。[CM007, CM008, CM009, CM021, CM022, CM026]

细分市场 / 买方地图
细分市场主要买方主要用户可能付款方 / 预算负责人工作流 / 触发点与 PostHog 的关系
早期 B2B SaaS / PLG 初创公司创始人、工程师、产品工程师和 PM创始人 / CTO / 产品预算需要快速埋点和低摩擦实验贴合 PostHog 自助式、透明定价的切入打法
成长期软件团队产品负责人和工程经理产品、工程、数据产品 VP / 工程 / 分析预算需要看清生命周期、激活、留存和发布信心支持从分析先落地,再扩到回放、功能开关和实验
开发者工具和基础设施厂商工程和开发者体验负责人工程师、PM、数据工程 / 平台预算需要技术化埋点、数据仓库兼容性和开发者级控制与开源和 SQL 优先定位高度契合
跨职能优化团队增长、营销、UX、数据增长、UX、生命周期团队增长 / 数字化 / RevOps 预算需要实验、旅程分析和个性化激活解释了它与 VWO、Optimizely 和实验套件的重叠
受监管或隐私敏感团队工程、数据治理、安全工程和分析用户IT / 数据 / 平台预算需要掌控数据驻留、云区域或自托管逻辑当隐私和架构约束排除广告技术式分析时,PostHog 受益
延伸分析工程的企业数据团队分析工程、数据平台分析师、数据工程师、业务团队数据平台预算需要受治理的事件数据、质量控制和可供 AI 使用的上下文支撑 Product OS / CDP / 数据仓库扩张逻辑

概括客户案例、基准页面和相邻平台定位中最常见的买方—用户—付款方模式。

[CM007, CM008, CM009, CM017, CM021, CM022]
FM002: 买方 / 细分群体图

采用通常从技术埋点开始,随后扩展到产品、增长和高管利益相关方。

[CM003, CM008, CM009, CM021, CM022, CM028]
FM003: 采用漏斗 / 价值链图

典型路径从埋点到洞察,再到实验,最后进入跨职能激活。

[CM003, CM024, CM030, CM034, CM035, CM037]

2.4 增长驱动、采用约束与仍需证明的事项

几乎所有来源都指向同一个主增长驱动:软件增长越来越取决于理解产品内行为,而不是继续购买更多漏斗顶端流量。Mixpanel 的 2026 年基准文章认为,增长已经移入产品内部,激活和留存比原始获客更重要。Grand View 和 Mordor 补充了结构性原因:云原生交付降低部署摩擦,AI 和自助分析提高了丰富行为数据的价值,监管收紧后,隐私安全的增强能力和区域数据托管选项更重要。PostHog 自己的 GA4 和客户页面展示了这个论点的运营侧:工程师想要更少的割裂工具、更少手工埋点,以及更好地处理广告拦截器、cookies 和下游激活。约束同样可见。Mixpanel 提到,实验需要足够流量和清晰假设;dbt 强调数据质量差是最常见挑战;Forrester 称市场本身正在在开发者主导的功能管理和产品主导的实验之间分裂。合起来看,采用应会继续增长,但买方教育、数据治理和市场定义模糊会继续影响品类扩张速度。[CM017, CM018, CM019, CM020, CM021, CM022]

增长驱动与约束表
驱动因素 / 约束方向时点对 PostHog 的影响尽调问题
增长转向产品内部正向当前利好一体化产品行为平台,而不是只看渠道的分析工具确认 PostHog 能否把更多免费使用转成由留存驱动的扩张
激活和留存取代粗放获客,成为关键杠杆正向当前利好把分析、实验和发布控制合在一起的厂商要求客户证据,证明激活工作能带来付费扩张
云原生成本和部署优势正向当前支撑初创公司和中小企业自助采用、更快落地量化规模化后的云端毛利率影响和支持负担
AI 辅助分析和实验正向近期把品类价值从仪表盘扩到决策引导和自动化把持久工作流价值和短期 AI 功能热度拆开看
数据质量和信任缺口负向当前糟糕埋点会挡住采用,或压低实际 ROI检查 PostHog 如何处理埋点治理、schema 控制和脏事件清理
功能管理与实验的用户画像分裂负向当前可能拆散预算线,并让品类叙事更复杂测试 PostHog 能否靠一个经济买方拿单,还是需要多线程销售 / 采用
隐私、驻留和合规需求混合当前在受监管 / 自托管场景利好 PostHog,但也抬高落地复杂度核实隐私需求多常带来胜单,还是把买方推向内部自建
有效实验所需流量门槛负向持续对低流量产品或小团队限制价值,除非打包工具仍能证明支出合理厘清客户采用分析以外产品所需的最低规模

把市场顺风和实际卡点放在一起,后文才能区分品类增长和公司自身执行风险。

[CM017, CM018, CM019, CM021, CM022, CM023]

2.5 展示要点

Chapter 03

03竞争格局

3.1 直接同业、相邻控制平面与现状替代

PostHog 面对的不是一个镜像式竞争对手。分析侧最接近的直接同业是 Mixpanel、Amplitude、Heap 和 Fullstory,它们都已从经典事件图表向会话回放、实验或 AI 辅助分析的某些组合延伸。第二组竞争的是发布控制和实验预算:LaunchDarkly、Statsig、GrowthBook,以及现在承载在 Harness FME 内的 Split 能力。第三组并非直接同业,却仍然赢得同一个买方任务:Google Analytics 仍是免费、默认的营销和网站分析选项,VWO 销售优化和实验但没有变成完整开发者数据栈,内部自建或仓库原生方案对想控制存储、部署或成本的技术团队仍有可信度。按类别拆分很重要,因为当买方想要一个工程主导平台时,PostHog 能赢;但买方只要满足于专业层或现状栈,PostHog 就会输。[CP005, CP007, CP011, CP013, CP015, CP020]

竞品画像表
竞品 / 替代方案类别公开规模 / 融资信号目标团队或买方产品范围定价 / 打包信号战略方向 / 局限
PostHog一体化 Product OS 锚点定价页披露 60,000+ 客户;开源并提供自托管选项工程驱动的初创公司、产品团队、重视数据的构建者分析、回放、功能开关、实验、问卷、数据仓库、CDP、工作流按用量计费,各产品提供慷慨免费额度本组里最宽的一体化栈,但对非技术 PM 来说 UI 不是最容易上手
Mixpanel直接分析竞品公开产品和定价页更强调成熟数字分析平台,而不是当前融资披露需要分析优先工作流的产品驱动型 SaaS 团队分析、网站分析、回放和热图、实验和功能开关、KPI 树1M 事件免费;Growth 档按用量计费;Enterprise 定制成熟且覆盖面广,但部署控制以 SaaS 为先,打包宽度仍窄于 PostHog
Amplitude直接分析竞品官方定价展示套件深度;独立评测强调企业适配,而不是公开当前融资数据大型组织的产品、增长和高管团队分析、问卷、功能实验、网页实验、激活、回放、AI 工具免费档,Plus 每月 $49;更高档位定制PM 和治理定位强,但企业经济性更难公开对比
Heap直接分析 / 回放竞品10,000+ 家公司使用,现为 Contentsquare 旗下想要自动捕获和更易上手设置的团队带自动捕获的分析、AI 助手、回放附加项、更高套餐的数据仓库集成免费入门至每月 10k 会话;更高套餐和回放更偏定制自动捕获和 DX 网络背书有帮助,但模块宽度薄于 PostHog
Fullstory回放优先的相邻竞品定位为行为数据平台;审阅材料未给出当前公开融资细节需要定性旅程洞察的产品、UX、支持和 CX 团队自动捕获、行为分析、AI 洞察、指南和问卷、激活采用申请演示式打包,而非透明自助标价回放深度和行为细节仍强,但更宽的产品开发套件较窄
LaunchDarkly功能管理老牌厂商强调企业工程规模;审阅材料页面未显示当前客户数或融资优先考虑运行时控制、审批和发布治理的工程组织功能开关、渐进式交付、实验、可观测性、智能体控制Developer 套餐免费起步;企业档定制治理驱动的老牌厂商,工作流可信度强,但不是完整分析 OS
Statsig实验和功能管理套件客户证据显示竞争胜单;定价页未披露私募融资大规模跑实验的数据驱动型产品和工程团队分析、实验、功能管理、回放、网站分析、配置Developer 档包含 2M 计量事件免费;更高档按用量扩展快速增长的一体化控制平面;部署可移植性弱于开源对手
GrowthBook开源 / 数据仓库原生相邻竞品主页披露 3,000+ 家公司;审阅材料未披露私募融资需要数据仓库原生实验的技术型产品和数据团队实验、功能开关、产品分析、数据仓库原生部署、自托管Starter 免费,Pro 每席 $40,Enterprise 定制开源和成本叙事强,但产品宽度和分发仍小于 PostHog 或 Amplitude
Google Analytics 4现状型老牌方案更像 Google 生态默认项,而不是独立初创公司的规模故事营销、网页和归因团队绑定 Google 广告栈的客户旅程和 ROI 分析免费营销分析里摩擦最低的默认项,但卖点不是统一产品开发控制平面
内部自建 / 数据仓库原生栈替代方案没有共享规模信号;经济性取决于内部团队产能和数据基础设施优先考虑控制权和可组合性的技术团队围绕既有数据仓库和开源层自建分析、功能开关、回放或实验资本开支是工程时间,而非标价控制权和可移植性强,但价值兑现更慢,运营负担更重

各行综合保留的官方产品和定价页面,以及少量独立团队适配评论;只有审阅公开材料明确披露时,才列出融资和客户规模。

[CP004, CP005, CP006, CP007, CP008, CP009]
FP001: 竞争定位图

0 到 10 的序数分比较 x 轴部署或数据控制,y 轴捆绑产品广度。分数是有证据支撑的综合判断,并非厂商上报指标。

x 轴分数反映自托管、开源姿态和对数据平面部署的控制;y 轴分数反映各厂商在保留页面上公开捆绑了多少实质上不同的任务。

[CP001, CP005, CP007, CP011, CP013, CP015]

3.2 竞品画像、能力宽度与打包信号

从官方产品和定价页看,PostHog 仍以套件宽度突出:产品分析、网站分析、会话回放、功能开关、A/B 测试、问卷、数据仓库层、CDP 式路由和工作流工具,都被包装成一个 Product OS。但这种宽度优势在方向上已不独特,只是在程度上更强。Mixpanel 现在把回放、实验、功能开关、KPI 映射和数据仓库连接器与分析一起销售。Amplitude 展示分析、实验、激活、问卷、回放和多项 AI 功能。Statsig 在同一平台营销分析、实验、功能管理、会话回放和开发者配置,GrowthBook 则销售仓库原生实验、功能开关和产品分析,并支持自托管或云部署。定价也强化了同一模式。PostHog、Mixpanel、Statsig 和 GrowthBook 都发布透明的免费入口或自助定价,而 Fullstory、LaunchDarkly 企业计划和企业版 Amplitude 更不透明。市场看起来不像一个主导型既有厂商,而更像一组正在向同一份买方期望清单收敛的套件菜单。[CP001, CP002, CP003, CP005, CP006, CP007]

功能 / 能力矩阵
厂商核心分析回放 / 定性洞察实验 / 功能开关数据或控制平面层部署 / 信任说明
PostHog通过数据仓库和 CDP 形成强能力开源并提供自托管选项
Mixpanel通过回放和热图提供通过实验和功能开关提供通过数据仓库连接器提供托管 SaaS;成熟分析工作流
Amplitude通过会话回放提供通过功能和网页实验提供通过激活和 AI 分析层提供托管 SaaS;对治理友好的企业工作流
Heap部分通过回放附加项提供审阅材料未强调更高档部分通过数据仓库集成提供Contentsquare 平台内的托管 SaaS
Fullstory部分分析能力,加上强行为数据审阅材料未强调部分激活工作流托管行为数据平台
LaunchDarkly非核心非核心强运行时控制和审批聚焦治理的托管 SaaS
Statsig配置和指标引擎强带一体化实验引擎的托管 SaaS
GrowthBook正在成形的产品分析审阅材料中非核心强数据仓库原生部署云端或自托管部署
GA4营销和网站分析强审阅材料未见原生回放审阅材料未见一体化功能开关与 Google 广告栈强绑定Google 生态托管默认项

单元格只使用公开证据简写:强 = 保留页面明确主推;有 = 有营销但不是品类定义能力;部分 = 相邻能力或附加项;非核心 = 在审阅来源中不是核心。

[CP001, CP003, CP005, CP007, CP010, CP011]
定价 / 打包对比
厂商公开入门价 / 免费档计量单位或打包模式入门档覆盖宽度公开不透明度影响
PostHog免费;各产品免费额度之后按用量计费事件、录制、功能开关请求、行数和其他产品单位超额计费前已包含多个核心产品最适合想先广泛试用、再付费的团队
Mixpanel每月 1M 事件以内永久免费分析事件,Growth 按用量计费付费扩容前包含核心分析和有限回放自助部分不透明度低;Enterprise 定制分析经济性直观,但额外规模仍转向用量驱动
Amplitude免费档,Plus 每月 $49按席位和产品套件打包,更高档位定制分析优先套件,实验、回放和 AI 功能可见中等不透明度,因为更高档位定制入门门槛可达,但企业 TCO 需要实时报价验证
Heap每月 10k 会话以内免费入门入门按会话计费,更高档位定制打包分析优先;回放附加项和数据仓库功能上移到高端自动捕获取向买方的好起点,但可比套件经济性没那么透明
Fullstory申请报价 / 演示按套餐类别打包,没有公开自助美元标价按角色划分的行为数据和分析套餐不走销售流程,更难与用量型套件对标
LaunchDarklyDeveloper 免费起步按套餐计价,并提供定制企业价格以运行时控制和实验切入,不含完整分析 OS入门以上中高不透明度对治理优先买方有吸引力,但价格无法直接与分析套件对比
StatsigDeveloper 档含 2M 计量事件免费按用量计量事件和曝光入门即包含功能开关、配置、实验和分析入门档不透明度低对实验密集团队,自助经济性进攻性强
GrowthBookStarter 免费;Pro 每席每月 $40云端按席位计价,并提供自托管选项Starter 含无限实验和功能开关;分析 beta 和高级统计在更高档公开档位不透明度低对封闭式功能管理老牌厂商形成强价格压力
GA4免费使用成本大多藏在 Google 生态经济性里仅限营销和客户旅程分析入门档不透明度低便宜默认项,可能推迟切换,直到产品团队超出营销分析范围

这张表只比较保留页面中的公开入门信号和披露的打包逻辑;企业折扣、最低承诺额和谈判条款仍是未解决证据缺口。

[CP002, CP006, CP008, CP010, CP012, CP014]
FP002: 功能广度 / 能力图

按类别展示能力地图,显示公开证据下各竞争类别的强项。色调只代表相对判断。

该图有意按战略类别归组厂商,避免重复详细厂商矩阵;色调总结保留证据,而不是厂商上报分数。

[CP023, CP024, CP031, CP032, CP034, CP038]

3.3 切换成本、多栖使用、分发力与信任姿态

PostHog 最强的切换成本论点不是硬锁定,而是运营便利。如果一个团队能在一个登录入口、一份合同里买到分析、回放、功能开关、实验、问卷和数据管道,采购和实施摩擦都会下降。Cotera 的比较明确描述了在同一个工程驱动栈里替换 LaunchDarkly 和 Fullstory。但同一证据也显示了护城河上限。PostHog 开源且可自托管,GrowthBook 也是如此,因此部署控制帮助技术买方,也降低了离开厂商云的痛感。多栖使用仍然现实:团队仍可把 GA4 与回放搭配,或把分析与独立功能开关厂商搭配,或从 LaunchDarkly、Statsig、VWO 购买专门的实验层。分发力也并不均衡。GA4 受益于 Google 的默认营销覆盖,LaunchDarkly 受益于企业治理和审批,Amplitude 受益于对 PM 友好的工作流和高管报告。信任也由这条分裂塑造。PostHog 和 GrowthBook 借代码访问与部署选择建立信任,而既有 SaaS 对手借托管治理、审批和合规姿态建立信任。[CP017, CP021, CP025, CP027, CP028, CP032]

3.4 护城河耐久性、商品化风险与 PostHog 最暴露的位置

反向解读很直接:分析、实验和功能管理正在收敛为共同能力层,任何单一模块长期保持专有的概率下降。Forrester 已经把功能管理加实验框定为一个合并市场,厂商页面也显示这种收敛已经落地。GrowthBook 正在推进开源和低成本实验。Statsig 把分析与发布控制打包。Mixpanel 和 Amplitude 都在向经典分析之外扩展。因此,PostHog 的护城河不再落在某个单一功能上,而落在特定组合上:一体化开发者工作流、透明用量定价和开放部署控制。这些是真优势,尤其适合工程驱动型创业公司,但并非所有买方都优先看重。独立评论也保留了两个重要反向点:PostHog 对非技术 PM 不是最容易上手的选择,而最佳单项专家仍然占住会话回放深度、企业治理或生态默认等高端用例。因此,耐久性问题取决于细分市场。买方想要一个技术平台时,PostHog 最强;买方已经信任既有工作流,或只需要一个专业层时,PostHog 最弱。[CP023, CP024, CP026, CP027, CP028, CP034]

护城河耐久性 / 竞争风险登记表
护城河主张或压力点证据威胁向量严重程度当前缓释或抵消因素尽调要求
产品整合广度PostHog 打包的模块比多数直接对手更多品类趋同意味着对手也在补相邻模块一体化套件仍能帮技术团队减少工具蔓延按队列索取附加率和多产品留存
开源 / 自托管控制PostHog 和 GrowthBook 让买方可选择部署方式部署选择降低供应商锁定,退出时痛感也可能更小隐私和基础设施控制仍能形成差异验证自托管转向付费云增购与流失各占多少
透明定价PostHog 公布慷慨免费档和按量费率Statsig、GrowthBook、Mixpanel 和 GA4 也给出有吸引力的免费入门条件透明度仍能推动自助采用并建立信任收集真实企业报价,检验规模化后定价是否仍透明
开发者工作流强度独立评测认为 PostHog 最适合开发者非技术 PM 可能更偏好 Amplitude 或 Mixpanel目标客群仍是工程主导团队用同时包含 PM 和工程师的参考客户测试真实工作流
回放深度对比专业厂商Fullstory 仍以回放为核心,行为数据更丰富部分团队可能发现,最佳单点回放工具仍跑赢套件便利性PostHog 把回放留在同一技术栈内,借此缩小差距在真实部署中比较回放分析深度和上线速度
企业治理与审批LaunchDarkly 和 Amplitude 在大型组织中仍有治理可信度即便功能重叠增加,既有工作流仍会阻碍替换当治理要求较轻或更多由开发者掌控时,PostHog 有机会胜出要求安排企业参考客户访谈,核验审批、可审计性和 RBAC 适配度
现状方案与专业工具多栖GA4 加单点工具仍可行,且通常便宜买方可以混搭不同层级,而不是完全标准化到一个套件一体化合同仍能节省协调成本衡量套件整合带来的实际采购和管理节省
功能趋同与商品化Forrester 和厂商页面显示,实验加功能开关正在变常见单一模块可能不再稀缺到足以支撑溢价护城河从原始功能转向工作流、数据模型和分发按竞争对手类型跟踪路线图速度和赢单 / 输单原因

严重程度衡量其对 2026 年采购流程的竞争影响,不是量化财务损失模型。

[CP023, CP024, CP025, CP026, CP032, CP034]
FP003: 护城河 / 就绪度 KPI

这些紧凑的公开指标勾勒 PostHog 身边入门级和规模级竞争压力已经有多强。

每个 KPI 都直接来自保留页面,但该图混合不同单位,只为概括竞争压力;它不是归一化评分模型。

[CP004, CP006, CP009, CP016, CP018, CP041]

3.5 展示要点

Chapter 04

04财务情况

4.1 收入模型、公开标价与价格卡没有说明的部分

对一家私营软件公司而言,PostHog 的公开定价表面异常透明。公司按产品发布免费额度和超量计费器,覆盖分析、回放、功能开关、问卷、数据仓库、管道、AI 可观测性、AI、工作流和日志;配套文档还解释,用量而非席位是核心变现原语。这是模型标价侧的强证据。但它不足以把用量机制转成实际收入。同一套公开文档也称,客户可以获得创业公司赠额、非营利折扣、自定义优惠券、定制价格层、首层固定收费结构,以及预付固定、不按表计费的方案。换句话说,PostHog 对标价透明,但对净价不透明。正确解读是,自助价格卡真实且有战略价值,但企业和促销结构仍会改变各客户队列的实际 ARPA、递延收入节奏和毛利率。因此,投资人应把公开价格视为入口经济性,把实际经济性视为尽调请求,而不是已经披露的事实。[CI001, CI002, CI003, CI004, CI005, CI006]

收入来源表
收入来源变现机制计费单位当前公开状态证据质量尽调要求
产品分析按量自助或企业合同事件公开标价和免费档已披露按队列索取云端与自托管的付费量和合同占比。
会话回放按量自助或企业合同录制数公开标价已披露;实际附加率和留存未公开索取每条录制的存储成本和付费附加率。
功能开关 / 实验按量自助或企业合同请求数公开标价已披露,本地评估计费机制也有文档索取前端与后端评估占比,以及付费功能开关采用率。
调研问卷按量自助或企业合同回答数公开标价已披露;实际变现结构未公开索取付费调研采用率和回答量分布。
数据仓库按量自助或企业合同行数公开标价和免费历史同步已披露按托管仓库工作负载和同步类型索取毛利率。
数据管道按量自助或企业合同事件 + 行数公开标价已披露;导出结构未公开索取批量导出与实时目的地的结构,以及相关 COGS。
AI 可观测性按量自助或企业合同事件免费档已披露;付费采用率和模型成本未公开索取付费 AI 可观测性客户数、使用结构和模型支出。
PostHog AI按量自助或企业合同积分免费积分已披露;付费转化未公开索取付费转化率、平均积分消耗和毛利率。
工作流按量自助或企业合同各渠道消息数免费档已披露;已变现量未公开索取工作流附加率、消息结构和超额用量兑现情况。
企业积分 / 定制方案谈判确定的预付或混合合同积分、自定义档位或固定费用手册支持,但未公开标价审阅订单、折扣梯度和收入确认政策。

各行区分公开标价和仍未公开的商业路径。「当前公开状态」反映官方页面可见内容,不代表实际净价或经审计的收入结构。

[CI003, CI008, CI009, CI012, CI013, CI014]
定价 / 变现表
产品或商业动作公开标价 / 免费额度对公开标价的解读实际定价可能变化的因素来源
产品分析前 1M 个事件免费;之后每个事件 $0.0000500 至 $0.0000090透明按量标价企业档位、优惠券或创业公司积分会改变净价定价页 + 分析定价文档
会话回放前 5k 条录制免费;之后每条录制 $0.0050 至 $0.0015透明按量标价存储时长、移动端占比和企业折扣未公开定价页
功能开关前 1M 次请求免费;之后每次请求 $0.000100 至 $0.000010透明按量标价本地评估可能放大等价请求量;套件经济性未公开定价页 + 成本估算文档
调研问卷前 1.5k 份回答免费;之后每份回答 $0.100 至 $0.010透明按量标价回答量和付费采用情况未公开定价页
托管仓库前 1M 行免费;之后每行 $0.000015 至 $0.000001透明按量标价历史同步免费,但按工作负载拆分的基础设施毛利未公开定价页
创业公司计划最高 $50k 积分促销计划,不是长期标价资格条件、领取率和付费转化未公开计费 FAQ
非营利 / 特殊情形折扣未披露公开价格梯度由销售介入的折扣路径实际折扣梯度和审批规则未公开计费 FAQ
定制企业合同支持固定首档、预付积分、优惠券或不计量方案非标商业路径订单、最低承诺和递延收入处理仍未公开计费手册

本表将公开标价与实际定价拆开。凡提到积分、折扣或定制方案的行,都说明公开单价不等于实际净收入。

[CI004, CI005, CI006, CI007, CI008, CI009]
FI001: 收入模式桥接图

公开证据支持基于用量的自助式入门模式,并存在企业例外,但企业侧定价并不完全透明。

[CI001, CI003, CI008, CI014, CI024, CI044]

4.2 GTM 动作与销售效率代理证据

PostHog 不披露 CAC、回本周期或 NRR,因此公开材料中读销售效率,最好看动作和客户行为,而不是财务比率。这个动作很强地呈现产品驱动、工程优先。官方页面强调透明定价和无需销售介入的接入,计费 FAQ 称,潜在客户通常先免费启动并观察用量,反而能得到更好的支出估计。客户证明方向上也一致。Y Combinator 称 PostHog 比 Google Analytics 捕获多 30% 的数据,看重与工程师直接 Slack 沟通,并用实验带来 40% 更多消息和 35% 更多接受请求。Hasura 称 PostHog 驱动的新用户引导调整让转化率提升 10-20%,使用也从工程扩展到 UX 和营销。这些是可信的落地扩张信号,但仍是代理。它们没有告诉我们混合付费转化、自助到销售辅助的转化率,或增长中有多少来自单账户用量提升、多少来自更多付费账户。公开证据支持低摩擦采用和清晰客户价值,但还不能构成完整的单位经济模型。[CI018, CI019, CI020, CI021, CI022, CI028]

FI002: 单位经济性桥接图

公开资料包能看出使用量如何变成账单,但折扣、流失和支持成本之后的队列经济性没有披露。

发票之后的节点仍只能定性描述,因为公开来源没有披露实际 ARPA、NRR、CAC 回本周期或毛利率。

[CI009, CI010, CI011, CI035, CI036, CI067]

4.3 成本结构、公开牵引力下限与上市公司参照

尽管缺少审计视角,公开牵引力下限仍比许多私营同业更强。招聘页称收入已经超过每年 $50 million,手册未来页目标到 2026 年底达到 $100 million,Y Combinator 档案称收入月增约 10%。官方表面还显示一系列规模信号:60,000+ 客户、190,254+ 团队、176k+ 历史注册、200+ 人团队和 25 个计划招聘名额。这些数据点确认了商业动能,但没有证明实际 ARR 质量。成本侧,产品和计费文档暗示的是经典云软件成本基座:计算、存储、事件处理、回放留存、数据仓库行、AI 工作负载,以及庞大的 R&D/支持表面。因此,合适的公开基准不是硬件或金融科技,而是高毛利开发者软件。Atlassian FY2025 报告显示毛利率 83%,R&D 占收入 51%,销售和营销占 22%。Datadog FY2025 10-K 暗示毛利率约 80%,GTM 强度更高,约 28%,R&D 约 45%。这些数字有助于框定规模化软件平台可能长什么样,但它们是基准,不是 PostHog 事实。[CI018, CI019, CI020, CI021, CI022, CI023]

单位经济性表
指标公开数值 / 状态置信度重要性尽调要求
收入下限招聘页披露超过 $50M/year给出最低规模下限,但不能把该数字等同于 ARR 或经审计收入。索取月度 ARR 与确认收入的桥接表。
2026 年收入目标年收入目标 $100M,需要约 7% 月增长揭示管理层增长门槛和隐含经营计划。索取 2026 年至今月度实际收入与计划对比。
当前增长节奏YC 资料页显示月收入增长约 10%指向强劲扩张,但未经审计,也未与签约收入和确认收入对账。索取董事会指标定义和队列桥接表。
客户规模代理60,000+ 客户、2025 年 176k 次注册、190,254+ 个团队显示采用面广,但看不出付费客户密度或集中度。索取付费账号数、免费转付费转化率和前 20 大客户集中度。
实际净价 / ARPA公开标价看不出折扣、创业公司积分或企业最低承诺。按客户队列、产品和折扣代码导出账单。
毛利率毛利决定用量增长是放大现金生成,还是推高烧钱速度。按产品和托管模式披露云端毛利率。
NRR / 扩张队列能留住并加深消费时,落地扩张叙事才成立。按客群提供总金额留存和净金额留存。
CAC 回本周期 / 销售营销效率客户案例不能替代可量化的获客效率。提供回本周期、混合 CAC 和销售辅助漏斗转化率。
现金余额 / 现金跑道默认存活状态看不出可用流动性或融资风险。提供现金、债务、烧钱速度和现金跑道模型。
员工数 / 运营费用代理200+ 名员工,计划招聘 25 人即便薪酬未披露,仍能看出薪酬支出增长压力的方向。按职能提供全口径薪酬、托管和承包商支出。

空值表示该指标为私有,或无法从已审阅的公开材料支撑。Figure FI003 中的上市公司类比仅作为基准,不代表公司事实。

[CI018, CI019, CI020, CI021, CI022, CI023]
FI003: 财务估算区间

公开软件可比公司提示了成规模开发者平台可能落入的毛利率和运营费用区间,但这些只是基准,不是 PostHog 披露值。

这些区间来自 Atlassian 和 Datadog 的公开文件,只作为投资测算类比纳入。它们不是关于 PostHog 自身毛利率、运营费用结构或烧钱速度的主张。

[CI052, CI053, CI054, CI057, CI059, CI060]

4.4 资本充足性、回款纪律与仍未解决的阻碍

2025 年融资历史让 PostHog 看起来暂时资金充足,但只是暂时。官方口径是,公司 2025 年 6 月以 $920 million 估值融资 $70 million 新股资本;第三方报道后来称 Peak XV 领投一轮 $75 million 融资,估值 $1.4 billion。与此同时,官方 Series D 文章和招聘页都明确员工流动性很重要,这意味着投资人不能假设每一美元披露融资都像纯新股一样延长现金跑道。运营计费文档提供了另一个有用但需要谨慎的视角。它们显示了正式回款控制:信用方案采用 Net 30 开票,预付合同用银行转账结算,随用随付客户可重试四次,逾期付款者可能被暂停服务、退回免费层或被要求预付款。这对纪律是正面信号,但也提醒我们,收入运营可能带来流失或数据丢失风险。核心阻碍很直接:公开材料仍未披露现金、烧钱速度、现金跑道、毛利率、NRR 或云与自托管组合。因此,财务结论是,模型看起来像软件公司,也可能具备资本效率,但当前资本充足性仍无法只靠公开证据完全承销。[CI037, CI038, CI039, CI040, CI041, CI042]

资本充足性表
资本项目公开数值 / 状态证据质量重要性尽调要求
最新披露轮次2025 年 9 月以 $1.4B 估值融资 $75M给出最新公开估值锚点,但不披露资产负债表上的当前剩余现金。核验交割日、新股与老股拆分,以及当前剩余现金。
此前新股融资2025 年 6 月以 $920M 估值完成 $70M 新股融资确认近期发生过一笔与现金跑道有关的大额融资。对账到账现金、费用,以及任何 side letter 或流动性分配。
资金用途官方 Series D 文章称,资金将支持更多产品,以及更多支持、销售和营销用例表明 2025 年资本计划用于扩展产品广度和 go-to-market。按职能和产品线索取 2026 年预算。
员工流动性2024 年老股交易、2025 年 tender,以及 2025 年轮次层面的流动性均已公开披露流动性目标可能削弱新增资本对现金跑道的实际延长。索取募资收益在新股、老股和费用之间的分配。
当前在手现金没有公开资产负债表,无法干净计算现金跑道。提供现金、短期投资、债务和最低现金政策。
月度烧钱 / 自由现金流PostHog 没有公开烧钱倍数或 FCF 趋势。提供过去 12 个月烧钱额和 2026 年预测。
回款纪律Net 30 预付账单、用量账号四次重试,并升级到暂停服务或回退免费档显示收入运营纪律,但也带来一定非自愿流失风险。索取 DSO、坏账、退款和非自愿流失指标。
债务 / 项目融资义务已审阅材料中未发现公开风险债务或项目融资义务隐性杠杆可能大幅改变稀释和现金跑道测算。提供债务明细、契约条款、租赁承诺和任何表外融资。

公司概况章节覆盖完整融资时间线。本表只重述判断未来资本充足性所需的融资事实;凡现金、烧钱和现金跑道属于私有信息,均明确保留为空。

[CI037, CI038, CI039, CI040, CI041, CI042]
公开财务缺口表
缺失的私有指标对投资判断的影响公开代理指标或基准具体尽调路径
实际 ARPA / 混合净价没有该指标,公开标价无法转化为实际收入质量或回本判断。定价文档显示公开档位;计费文档显示积分、折扣和定制方案。按产品、客户、折扣和合同类型导出 12 个月账单收入。
按产品拆分的 ARR / 确认收入无法评估收入结构集中度和增长持久性。招聘页称 >$50M/year,future 页面称目标 $100M,Sacra 发布了非官方估计。提供 2025 年 1 月以来的月度确认收入、ARR 和产品结构。
按产品和托管模式拆分的毛利率阻碍判断用量增长能否盈利性扩张。Atlassian 和 Datadog 文件暗示软件毛利率约 80-83%,但 PostHog 可能存在实质差异。披露云端 COGS、自托管经济性、存储 / 计算成本和毛利率瀑布图。
NRR / logo 留存 / 流失没有队列留存,就无法检验落地扩张叙事。客户故事显示扩张,但没有公开队列统计。按客群提供总金额留存、净金额留存、logo 流失和扩张。
CAC 回本周期和销售辅助转化自助服务案例看不出获客效率。没有公开 CAC 或回本周期;只能看到支持和采用代理指标。提供漏斗转化、按渠道拆分的支出和按队列拆分的回本周期。
现金 / 烧钱 / 现金跑道资本充足性仍不可知,也无法判断下一轮融资时点。公开信息只有融资事件和 default-alive 表述。提供当前现金、月度烧钱、现金跑道和下行情景计划。
云端与自托管收入结构托管结构会改变毛利、支持负担和收入确认。开源和自托管定位已公开,但收入结构未公开。提供付费云端 ARR、自托管支持收入和迁移趋势。
企业合同负债 / 递延收入混合合同可能让现金回款节奏与收入确认错位。计费手册确认存在预付和额度结构;但没有公开披露递延收入。提供合同原型、递延收入余额,以及按产品拆分的收入确认政策。

每一行都是尽调阻断项,而不是隐藏结论。提到上市公司基准的地方,只作为框架类比,不能替代公司数据。

[CI014, CI015, CI042, CI066, CI067, CI068]
FI004: 资本强度 / 现金流地图

公开证据显示公司近期完成融资且计费控制较严,但回款之后,真实流动性桥接仍消失在私有数据里。

由于 PostHog 不公布现金、烧钱速度或现金跑道,该图为定性图。它标出公开证据在哪里结束、尽调必须从哪里开始。

[CI037, CI038, CI042, CI046, CI048, CI049]

4.5 展示要点

Chapter 05

05产品与技术

5.1 平台定义与 PostHog 正在销售的工程工作流

PostHog 的产品定义比「产品分析」更宽。官方 Product OS 文档把分析、网站分析、回放、功能开关、实验、问卷和数据仓库打包成一个系统;代码仓库和 YC 档案又把范围扩展到错误跟踪、CDP/数据管道、LLM 可观测性和 AI 产品助手。按工作流看,公司想拿下一个工程师主导的闭环:用 SDK 和自动捕获埋点行为,用分析 / 回放 / SQL 检查行为,在功能开关和实验背后发布变更,再把由此产生的数据激活或导出到栈内其他系统。 这个打包工作流是真正差异点,因为它减少了早期和中期工程团队的工具拼接。公开资料包也让商业形态清晰:定价按产品表面用量计费,免费层覆盖几乎所有主要模块。取舍在于,模块宽度本身已经成为尽调问题。核心分析、回放和功能开关在公开文档中看起来成熟;较新的 AI、工作流和更广的激活表面可见且已变现,但它们各自 GA 还是预览的确切组合,在留存资料包中并不同样明确。[CE001, CE002, CE003, CE004, CE005, CE006]

产品模块 / SKU 矩阵
模块 / 产品面主要用户任务公开状态 / 成熟度公开计量口径或部署单位主要尽调缺口
产品分析 + Web 分析量化用户行为、漏斗、队列和增长核心 / 成熟事件和 Web 分析额度企业版功能分层的细节,不如核心用量计量口径清晰。
会话回放诊断 UX 摩擦、支持问题和性能核心 / 成熟录屏留存资料只部分交代了高级回放信号覆盖和存储经济性。
功能开关 + 实验渐进发布、远程配置和 KPI 验证核心 / 成熟功能开关请求;实验随功能开关计费公开资料把机制讲得清楚,但大型企业治理细节偏弱。
调研问卷在产品流程内收集定性反馈相邻模块 / 已成型回复资料展示了打包方式,但按客群拆分的采用深度公开证据不多。
数据仓库 + SQL将外部数据与产品数据打通并直接分析战略模块 / 成熟到可售行数和查询用量详细来源 / 导入覆盖不如顶层定位充分。
CDP + 数据管道 + 工作流转换、路由并运营化产品数据扩张模块事件、行数、消息和目的地活动新激活模块在 GA 与预览之间的成熟度,公开披露仍不均衡。
AI 产品面:AI 可观测性、PostHog AI、MCP将产品数据带入 AI 调试和智能体工作流已有明确上线信号的扩张模块事件、额度和免费 MCP 访问公开资料展示了访问方式和打包方式,但没有完整公开路线图或附加率数据。

表格把长期成熟的行为类产品,与较新的激活和 AI 产品面分开。模块地图清楚呈现打包宽度,但每个新模块的公开成熟度细节并不一样深。

[CE001, CE002, CE008, CE010, CE011, CE031]
工作流 / 用例表
用户任务当前工作流需求PostHog 路径公开证据主要限制
快速埋点产品行为工程师希望不用手工埋每个细碎交互,也能拿到可用分析SDK、posthog-js 自动捕获和产品分析Product OS、代码仓库 README 和 JS 文档自动捕获减少设置工作,但不能替代周密的事件设计。
排查转化或支持问题团队既需要行为上下文,也需要技术细节分析、回放、控制台 / 网络视图和 SQL回放文档、FullStory 对比和代码仓库 README资料能证明工作流跑得通,但不能证明每个模块都比同类最佳工具更强。
在生产环境安全发布渐进发布功能并衡量影响功能开关、实验、本地评估和发布阶段功能开关文档、本地评估文档和发布手册治理、审批和大型企业运营细节,不如核心机制公开。
运营化产品数据将清洗后的产品数据送入下游工具CDP 转换、目的地和工作流CDP 页面和代码仓库 README目的地广度清楚,但目的地级别的质量和 SLA 细节偏薄。
在 AI 原生工具里使用产品数据开发者希望从编辑器或智能体中拿到上下文并触发动作MCP 服务器加向导式安装,接入编程客户端MCP 文档和 GA4 对比安全模型有文档,但实际采用和 ROI 数据未公开。

工作流表刻意以工程师为中心,因为留存公开资料反复把 PostHog 描述成开发者平台,而不是营销人员优先的分析工具。

[CE003, CE005, CE007, CE008, CE010, CE023]
FE001: 产品系统流

PostHog 的产品循环从采集开始,进入分析、受控发布、下游激活,再从开发者工具里做 AI 辅助查询。

该图综合已记录的工作流界面,而非复刻内部系统图。它只反映留存公共资料包中明确可见的产品流。

[CE003, CE007, CE008, CE010, CE031, CE032]

5.2 模块地图、架构与共享数据平面

PostHog 披露的运营栈多于许多私营软件公司。架构文档点名 Django 用于 Web 应用和 API,Rust 用于高吞吐捕获 / 功能开关 / 回放服务,Kafka 作为传输层,Celery 加 Temporal 加 Dagster 处理工作进程,一个 Node.js CDP 工作进程,以及 ClickHouse/PostgreSQL/Redis/blob 存储作为核心数据平面。ClickHouse 文档又把摄取模式具体化:Kafka-engine 表、物化视图、分布式表和分片位于事件捕获与查询负载之间。 SDK 表面同样明确。浏览器侧文档展示了代码片段和包管理器安装路径、框架专用集成、可选扩展包、跨域追踪、回放触发器和命名实例;服务端功能开关文档则暴露多语言本地评估。结果是,PostHog 的架构更像一个可观测数据与控制平面,而不是单一仪表盘产品:一个捕获层把分析、回放、功能开关、实验、数据仓库查询和下游 CDP 动作都喂给同一个共享底座。[CE003, CE012, CE013, CE014, CE015, CE016]

技术 / 运营架构表
层 / 组件作用公开证据依赖薄弱时的风险
浏览器 + 服务端 SDK从应用端捕获事件、识别用户并评估功能开关文档和包注册表证据强客户端埋点质量、项目配置和 SDK 更新状态捕获质量差,会同时污染分析、回放和实验。
Django Web 应用 / API面向用户的控制平面、API 和编排入口官方架构文档PostgreSQL、Redis、工作器服务和认证UI 和 API 不稳定,会削弱整个操作面。
Rust 捕获 / 回放 / 功能开关服务高吞吐摄取和快速评估路径官方架构文档Kafka、blob 存储和上游 SDK 行为这些服务一旦滞后或故障,核心产品闭环会快速劣化。
Kafka 传输层连接摄取、存储和 CDP 处理的中央总线官方架构 + ClickHouse 文档生产者可靠性、topic schema 和消费者健康度队列积压或 schema 错误会波及分析和激活。
ClickHouse 分析后端主分析存储和查询引擎官方 ClickHouse 文档Kafka 消费者、物化视图、分片和查询设计查询隔离或摄取设计薄弱,会带来准确性和安全问题。
PostgreSQL / Redis / blob 存储运营元数据、缓存,以及录屏 / 对象数据官方架构文档应用服务和工作器这些存储承载状态和回放资产,是面向用户产品的依赖。
Celery / Temporal / Dagster 工作器短任务、可靠工作流和定时数据管道官方架构文档队列健康度、权限和管道定义工作流或管道故障会直接变成用户可见事故。
CDP 工作器和目的地层转换入站事件并发送到下游工具官方架构和 CDP 页面Kafka、映射、目的地 API 和质量规则转换错误或导出中断会破坏激活和下游信任。

架构表只反映留存公开资料中明确可见的层。它不纳入隐藏内部服务,也不推断微服务边界。

[CE012, CE013, CE014, CE015, CE016, CE023]
FE002: 模块 / 能力地图

产品栈把采集接口和共享数据基础设施垫在底层,上面承载分析、控制、激活和 AI 工作流。

层级按功能角色分组,不按内部归属或计费 SKU 边界分组。

[CE001, CE002, CE010, CE012, CE014, CE031]

5.3 部署、集成与产品变更如何发布

PostHog 给客户两种非常不同的部署姿态。云端是推荐路径,也是明显为规模优化的路径。自托管仍可通过 MIT 许可的 Docker Compose 爱好者部署获得,文档强调产品表面相同,但基础设施和支持模型不同。运营方要自己承担扩容、URL、升级和故障风险;公开 README 还暗示,开源部署在大约每月 100k 个事件以内才实际可行,之后迁移到云端会成为推荐动作。任何买方或投资人如果想把「开源」等同于完整企业私有云支持策略,都需要重视这个限定条件。 发布机制也异常透明。手册称,新产品先放在功能开关后面,再经过 alpha、beta 和 GA,并用用户级预览和即将上线界面发现早期需求。这与技术产品形态一致:浏览器 SDK 模块可以选择性打包,服务端功能开关可以本地评估,MCP 向导可以把 PostHog 的 AI 访问层直接安装到开发者工具里。较弱的是路线图具体性。公开路线图 URL 存在,但留存的审计抓取只看到加载壳,因此近期详细时间表仍更多依赖发布流程披露,而不是可读的公开路线图导出。[CE017, CE018, CE022, CE023, CE024, CE025]

集成 / SDK 覆盖表
产品面公开支持关键细节部署注意证据
JavaScript Web SDK代码片段或包管理器安装框架专项文档、命名实例、跨域 cookie、回放触发器和退出捕获扩展加载和 CSP / Electron 约束,可能要求选择 slim 或 no-external bundle 包JS 文档
服务端功能开关评估Node、Ruby、Go、Python、C#/.NET、PHP、Java 和 Rust本地评估在后台拉取功能开关定义,并减少请求数调用方必须提供所有相关属性;无状态运行时需要共享缓存或远程评估功能开关本地评估文档 + PyPI
分析 / SQL / 数据仓库产品面可视化分析、SQL 访问和外部数据导入SQL 定位为基于共享产品 / 客户数据的无限制自定义分析详细导入来源覆盖不如顶层定位充分Product OS + 代码仓库 + YC
CDP / 管道 / 目的地实时转换、webhook 和下游同步支持数据增强、映射、校验、PII 清理和运营触发器各目的地的实现质量公开描述不深CDP 页面 + 代码仓库 README
MCP / AI 客户端PostHog Code、Cursor、Claude Code/Desktop、Codex、VS Code 和 Zed免费 MCP 服务器,支持 OAuth 或项目级 API 密钥,并按区域路由LLM 工作流带来提示注入审查要求MCP 文档
自托管部署Docker Compose 爱好者部署与云端相同的产品面,并可选 TUI 试用无带标签的版本发布、无保证,也不再新增付费开源 Kubernetes 部署自托管文档 + README

表格聚焦公开实现细节最清楚的集成产品面。它不试图枚举每一种语言 SDK 或每一个下游目的地。

[CE017, CE018, CE019, CE020, CE022, CE023]
路线图 / 发布 / 开发阶段表
阶段 / 机制公开页面说法当前含义证据剩余缺口
准备中初始规划和 alpha 开发放在功能开关后新产品面可以在客户广泛可见之前存在发布手册 + 功能开关文档没有公开材料把每个当前模块绑定到这套阶段分类。
Alpha已沟通过的客户会被逐步加入功能开关手工早期发布是标准运营模式的一部分发布手册没有披露公开客户名单或 alpha 遥测。
Beta向所有想选择加入的用户开放选择加入式 beta 是广泛发布前的常规关口发布手册留存资料没有把每个当前产品映射到 beta 或 GA 状态。
GA完整发布包括定价和营销GA 时商业和产品页面应当对齐发布手册 + 定价具体发布日期因产品而异,且常未披露。
功能预览 / 即将推出用户可以在用户级别切换预览或登记兴趣PostHog 在更广泛推出前,会在产品内发现需求发布手册留存资料没有完整抓到预览清单本身。
JS SDK 节奏文档称团队「快得离谱」地发版,npm 显示运行日前两天有一次发布浏览器产品面看起来维护活跃JS 文档 + npm速度本身不能证明稳定性。
Python SDK 节奏PyPI 显示 2026-05-21 上传,且带可信发布来源非 JS SDK 产品面同样显得活跃且新PyPI注册表更新不证明各语言功能对齐。
公开路线图页面路线图 URL 在线,但留存抓取只返回加载壳经审计公开资料无法读到近期路线图细节路线图抓取具体即将推出事项、负责人和日期仍需内部尽调。

发布流程证据强于路线图可读性。公开流程存在,但留存资料里的公开单项时间表仍有一部分不透明。

[CE011, CE029, CE030, CE035, CE036, CE051]
FE003: 部署 / 集成流

PostHog Cloud 是首选运行路径;自托管仍是同一产品的一条分支,但由客户自己运营,支持和规模假设也不同。

这张依赖地图强调部署选择和集成端点,不强调法律实体结构,也不穷尽每条内部服务边。

[CE025, CE026, CE027, CE029, CE031, CE032]

5.4 差异化:一个数据 / 控制平面,再叠加 AI 辅助工作流

PostHog 最强的产品差异化不是某个孤立模块,而是各模块共享同一个数据平面的方式。Product OS、公开代码仓库和 GA4 比较都指向同一条判断:分析、回放、功能开关、问卷、SQL、数据仓库查询和数据激活应该放在一起,因为工程师想在一个地方发布和调试,而不是对齐多个厂商。MCP 把这个判断延伸到更新的工作流。PostHog 不把分析只当作仪表盘目的地,而是暴露一个 MCP 服务器,让 Cursor、Claude Code、Codex、VS Code 和 Zed 等 AI 原生开发者工具能访问项目数据和产品动作。 不过,这个套件没有消除专业化风险。Forrester 认为功能管理和实验正在按人群分化,Mixpanel 认为 AI 正把门槛从「报告」抬到「决策支持」。因此,正确解读需要分层:PostHog 通过共享数据、SQL、捕获、控制和 AI 访问,具备可信的平台级差异化;但部分买方仍会按工作流评估最佳单项深度,而不是天然奖励套件宽度。开发者信号在这里仍是正面:公开代码仓库、npm 包和 Python 包都显示当前仍在积极发布的表面,而不是一个封在闭源商业核心外面的冻结开源壳。[CE010, CE031, CE032, CE033, CE034, CE035]

5.5 信任、隐私与质量控制——保留限定条件

信任姿态是 PostHog 记录得更好的领域之一。法律页面覆盖托管和自管理安装、在应用内执行 DPA、EU 与 US 托管,以及用户级捕获控制。官方比较页还提出 EU 托管、SOC 2、GDPR 就绪、HIPAA 就绪,以及相关套餐 BAA 等主张。产品文档进一步说明,捕获和回放可以通过触发器或退出选项收窄。对于一个处理敏感事件数据的开发者优先产品,这些都是有意义的控制,不只是营销口号。 但公开记录也把下行保留在明面上,本章就应该这样处理。安全公告页称今天没有活跃公告,却记录了一个已解决的 2025 年 SQL 编辑器问题:该问题暴露了跨团队查询文本,并用具体授权术语解释了修复。公开事后复盘索引还单独列出最近发生在工作流、日志、回放和功能开关上的可靠性故障。自托管进一步放大了取舍:PostHog 透明地说明客户可以自托管,但同样明确自托管实例没有保证,可能意味着数据丢失风险。因此,信任结论是透明度和控制为正面,但可靠性确定性只能算中等——尤其对自管理或高度监管买方来说,他们需要精确支持边界和经审计的范围文件,而不是姿态表述。[CE020, CE038, CE039, CE040, CE041, CE042]

信任 / 安全 / 合规控制表
控制 / 信号公开状态范围帮助点剩余限制
隐私政策 + 自托管遥测退出公开且在线的法律文本托管服务、网站和自托管安装澄清控制者立场、遥测收集和退出路径政策文字不等于外部审计,也不等于逐产品控制图。
应用内 DPA 生成器在线公开预览加应用内执行路径处理者义务和会签 DPA 工作流为数据处理条款提供法律机制公开预览只供参考;具约束力的执行需要进入应用。
美国 / 欧盟托管姿态公开且在线的法律 + 产品定位美国区域、欧盟托管云使用德国、云提供商披露,以及区域感知 MCP 认证支持数据驻留和区域选择论证资料不包含底层架构证明或客户级部署图。
回放触发器和捕获退出JS 文档有说明会话录制和用户级捕获行为帮助收窄数据收集,减少过度捕获围绕这些控制的精确运营默认值和审计日志未完全公开。
无活跃安全公告 + 安全公告计划公开维护的手册页面安全审计、披露流程和活跃状态页面体现安全透明度和流程成熟度当前页面不能替代安全白皮书或渗透测试摘要。
PSA-2025-00001 修复已解决的中等级别公告SQL 编辑器查询可见性问题展示通过 team_id 和测试变更完成的具体授权修复单个公告不能证明相邻授权缺陷不存在。
公开复盘计划活跃公开事故列表对客户或数据有直接影响的重大事故提升可靠性透明度留存资料只抓到索引,没有抓到每篇完整 RCA 正文和指标。
自托管无保证姿态文档明确说明支持、可靠性和运营者责任让云端与自托管的信任取舍更清楚这削弱了把自托管视为同等企业托管产品的论证。

控制项最强的地方在透明度、法律框架和用户可配置的数据收集行为。准确的审计控制范围和自托管支持边界,仍比公司姿态声明更薄。

[CE020, CE038, CE039, CE040, CE041, CE042]
FE004: 信任 / 可靠性 KPI 快照

公开信任信号在透明度和法律框架上最强;自托管保障和完整事故细节仍较弱。

[CE038, CE039, CE040, CE041, CE042, CE043]
Chapter 06

06客户情况

6.1 客户分层在工程驱动产品里最强,但公开客户数含义模糊

PostHog 可见的客户证明集中在快速构建和发布软件的团队。可访问的具名案例聚集在 Y Combinator 的创始人产品、Hasura 和 Supabase 的开发者基础设施、Phantom 的加密应用、ElevenLabs 和 Lovable 的 AI 产品、Arena 的模型比较平台、Exa 的 AI 搜索 API,以及 ResearchGate 的科研网络。贯穿这些案例,反复出现的买方和日常用户是产品、工程、增长,有时还有营销团队,而不是经典集中采购部门。这个画像与 PostHog 自身作为工程师「Product OS」的定位吻合,也解释了为什么产品常常通过埋点、实验和运营调试被采用,而不是通过自上而下的 CIO 指令进入。 不过,规模披露需要谨慎处理。PostHog 官方页面使用了几种不同单位:关于页称 190254+ 个团队,同一页称超过 190254+ 个客户和近 25 万名工程师,定价页称 60000+ 个客户。定价页还称超过 90% 的公司免费使用产品。这些表述方向上都支持广泛自助式足迹,但不能直接比较,也不能读成干净的付费客户阶梯。正确尽调立场是,PostHog 几乎肯定拥有很宽的漏斗顶端采用,但公开材料没有干净区分公司、团队、用户、注册、免费账户和付费客户。[CU001, CU002, CU003, CU004, CU005, CU007]

客户分层图
客群代表客户可见买方 / 用户 / 付款方主要 PostHog 用例战略价值关键缺口
创业加速器 / 创始人生态Y Combinator买方:产品 / 工程负责人;用户:产品、工程、管理层;付款方:中央平台预算面向创始人产品的产品分析和实验显示出异常强的创业公司分发契合度和 YC 批次渗透没有披露加速器式队列的付费转化或 ACV
开发者基础设施 / 数据平台Hasura、Supabase、Exa买方:工程 / 产品;用户:工程、增长、营销;付款方:产品驱动型开发者工具预算分析集中化、漏斗、回放、SQL、AI 助手、增长归因与偏好自助服务和数据所有权的高技术团队高度契合公开证据对企业席位扩张或续约条款着墨很少
消费级加密 / 金融科技应用Phantom买方:CTO / 联合创始人;用户:工程、产品、管理层;付款方:核心产品基础设施预算可靠性监控、DAU / 量级仪表盘、功能开关控制显示 PostHog 不只适用于 B2B SaaS,也能影响高频消费行为没有公开合同规模或付费计划深度
AI 原生构建者和模型产品ElevenLabs、Lovable、Arena买方:增长 / 产品 / 工程;用户:工程、增长、营销;付款方:产品或增长预算留存跟踪、AI 可观测性、实验、推出控制、调研闭环在快速发版和队列测试重要的场景,公开证据最强客户样本偏向 AI 公司,可能高估更广泛垂直行业多样性
大规模知识平台 / 科学网络ResearchGate买方:产品工程领导层;用户:工程、数据科学、产品;付款方:企业产品平台预算功能开关、实验、漏斗分析,以及超大规模企业支持证明在 25M+ 用户规模和自定义套餐支持下可行头部账户经济性和支持毛利仍未披露

表格按客户类型和可见运营模式归类可获取的具名证据。买方 / 用户 / 付款方来自案例和客户公司定位推断,不来自合同。

[CU008, CU010, CU014, CU017, CU022, CU026]
FU001: 客户采用 / 细分地图

公开证据指向一条开发者主导路径:自助入口容易、团队内扩散快,然后扩展模块,偶尔进入定制企业支持。

这是基于定价、产品定位和具名案例综合出的客户旅程,不是披露了实测转化率的漏斗。

[CU004, CU005, CU007, CU022, CU026, CU033]

6.2 采用轨迹指向广覆盖,但公开证明漏斗很快变窄

最宽的采用解读来自官方使用量主张,而不是披露的付费客户递进。PostHog 在关于页和定价页之间指向 190254+ 个团队、60000+ 个客户,以及超过 90% 公司免费使用的自助动作。除此之外,PostHog 称每一届 Y Combinator 批次中有 65% 使用其产品,这是最重要创业渠道之一里的有意义分发信号。独立评论数量又提供了一个外部标记:G2 档案显示 950 条评论和 4.5/5 评分,这个表面积远大于纯概念客户基础应有的水平。 但这仍不是干净的队列或部署阶梯。当证据从广泛官方计数转向独立可见用户,再转向具名客户故事时,公开证明很快变薄。本次审阅中,九个公开客户故事页面可访问,其中八个包含量化结果或规模标记。这足以说明采用真实存在,但仍是被策展的子集。因此,漏斗支持一个有层次的结论:PostHog 可能拥有很大的漏斗顶端和广泛产品熟悉度,但公开记录更能说明部分客户如何使用产品,而不是每个变现队列里到底有多少付费客户。[CU001, CU003, CU004, CU009, CU041, CU056]

结果 / ROI 表
客户指标 / 结果数值证据时效重要性限制
Y Combinator实验提升消息数增加 40%;匹配数增加 35%近期但未标日期的案例显示 PostHog 能影响核心双边市场行为闭环单次实验,没有完整分组经济性
Hasura入门转化提升 10-20%近期但未标日期的案例显示产品和 UX 团队用 PostHog 改变入门行为未披露基准转化率或收入影响
Supabase增长影响每周新增用户获取提升 10X围绕 2024 年末调整的近期叙事强力佐证归因和增长伙伴发现价值归因链未经独立审计
Phantom可靠性影响失败率降低 90%;持续目标为 1% 或更低历史改善,并给出当前运营目标证明 PostHog 不只适用于传统营销 / 产品分析场景结果由基础设施变化驱动,并非 PostHog 单独造成
Arena规模与参与度月用户 5M+;排行榜页面平均停留 19 分钟;六个月内事件增长 19×当前叙事显示 PostHog 在高容量 AI 产品规模下仍有用页面停留时长不是留存指标
ResearchGate测试规模超过 25M 用户;数亿次会话;全年模型测试当前叙事大规模实验和支持能力的强佐证没有公开 ROI 或合同金额
ElevenLabs发布纪律监控周留存,并将年度定价实验推至 100% 用户当前叙事显示 PostHog 进入以激活 / 留存驱动的发布闭环未发布提升百分比
Lovable供应商响应速度客户要求的 LLM playground 不到一个月上线当前叙事支持将扩张和路线图响应速度视为留存代理功能交付速度不等于续约证据

本表混合了直接量化结果和运营 ROI 代理。时效反映本轮可见的公开叙事; 许多客户故事没有明确发布日期。

[CU012, CU013, CU015, CU021, CU024, CU028]
FU002: 采用 / 证据深度漏斗

当证据从宽泛的漏斗顶部使用量主张推进到具名案例和独立评价时,公开证明会变窄,也更具体。

这是证据深度漏斗,不是字面意义的销售阶段漏斗。前两层故意展示不可直接比较的官方计数单位,以明确这种歧义。

[CU001, CU003, CU041, CU056, CU058]

6.3 具名客户证明扎实,且大多达到生产级,但新鲜度不均

PostHog 的具名客户证明明显强于通用 Logo 墙,因为最佳案例描述了活跃工作流、具名操作者和可衡量结果。Y Combinator 讨论其匹配产品上的实验;Hasura 和 Supabase 描述具体分析、漏斗和增长工作流;Phantom 每天用 PostHog 监控可靠性和 DAU;ElevenLabs 把它接入发布、周留存检查和问卷;Arena 用它在百万级用户规模上做实验护栏;ResearchGate 则依靠它在数亿会话上跑长达一年的算法测试。这些不是「工具不错」的轶事引语,读起来像对客户有实际重要性的运营部署。 新鲜度仍不均衡。最强案例使用现在时,并提到当前产品或近期时间窗口,但许多页面没有暴露清晰发布日期或续约时间戳。因此,可从当前功能引用中方向性推断新鲜度,却无法像有日期的新闻稿或客户文件一样严格审计。生产级与试点的分类,对具名故事本身最强,对更广客户名册最弱。公开证据支持可访问具名引用大多是真实生产账户,但不足以证明更大的不可见基础具备同样深度、年龄或耐久性。[CU010, CU014, CU017, CU022, CU026, CU029]

具名客户证据表
客户客群部署 / 用例生产 / 试点公开结果佐证 / 限制
Y Combinator创业加速器 / 创始人产品在 Startup School、Startup Library 和 Co-Founder Matching 中跑分析与实验生产环境一次实验中消息数增加 40%、匹配数增加 35%PostHog 案例与 YC 官网可佐证;未披露商业条款或续约数据
Hasura开发者基础设施漏斗、入门分析、回放,以及更广的网站 / 产品分析生产环境入门转化率提升 10-20%PostHog 案例与 Hasura 官网可佐证;未披露合同金额或合作期限
Supabase开发者基础设施服务端分析、SQL、AI 辅助分析、归因,以及增长伙伴发现生产环境每周新增用户获取提升 10XPostHog 案例与 Supabase 官网可佐证;采用广度清楚,但支出不透明
Phantom加密 / 消费金融科技日常仪表盘、可靠性指标和功能开关控制生产环境失败率降低 90%,稳态失败目标为 1% 或更低PostHog 案例与 Phantom 官网可佐证;未披露付费计划细节
ElevenLabsAI 语音平台人群画像追踪、周留存分析、回放、调研和发布实验生产环境分组测试后,年度定价实验推至 100% 用户PostHog 案例与 ElevenLabs 官网可佐证;仍无公开续约指标
LovableAI 应用构建器功能开关、实验,以及用于智能体调试的 AI 可观测性生产环境客户要求的 LLM playground 不到一个月上线部署真实,但 Lovable 也公开使用重叠供应商
ArenaAI 模型比较平台受控实验、分组分析、错误追踪和增长落地页生产环境月用户 5M+,六个月内事件量增长 19×结果规模清楚,但案例日期和合同数据不透明
ExaAI 搜索 API集中化分析、回放、功能开关,以及替代分散栈的 PostHog AI生产环境分析收进一个系统,仍有更多模块待采用整合价值的佐证不错,但公开 ROI 量化不足
ResearchGate科研网络高流量下的算法实验、功能开关、企业支持和漏斗分析生产环境超过 25M 用户,全年模型测试覆盖数亿次会话企业级佐证强,但定价和集中度仍不透明

覆盖范围有意保持不完整:这些是本轮在公开名单中能访问到的具名客户故事,并非完整客户清单。 「生产环境」按公开叙事判断,不等同于已签署的实施证明。

[CU010, CU014, CU017, CU022, CU026, CU029]
FU003: 客户证明新鲜度 / 类型矩阵

具名运营方、具体工作流和量化结果同时出现时,证明质量最强;但留存和合同经济性整体仍弱。

由于许多案例页面没有清晰发布日期,新鲜度根据明确时间标记和现在时部署表述推断。

[CU010, CU015, CU021, CU024, CU032, CU038]

6.4 耐久性代理信号有利,但没有正式留存披露

审阅的客户资料包没有公开 NRR、GRR、流失、合同期限或续约率披露,因此无法直接用公开证据承销耐久性。公开记录能提供的是一组有用代理。Y Combinator 在多个面向创始人的产品中使用 PostHog;Phantom 每天在每周全员会中使用,并把功能开关当作运营安全控制;ElevenLabs 在广泛推出功能前监控周留存;Arena 把留存和回访行为称作北极星指标;ResearchGate 基于 PostHog 跑了长达一年的信息流模型测试,并与专家支持互动。这些都是可信信号,说明产品嵌入了重复工作流,而不是偶尔用于一次性分析。 独立满意度证据有支持性,但仍不完整。G2 在 950 条评论上的 4.5/5 评分,对熟悉度和覆盖宽度是正面信号;Sacra 的商业模型框架也符合落地扩张型 PLG 动作。但评论评分不是续约指标,即便最好的公开证明集合,也很少说明客户是否续约、扩张合同价值,或完全替代竞品。正确解读是建设性但谨慎:公开证据支持重复使用和多产品扩张,但在 PostHog 披露队列经济性或给出一致付费客户指标前,客户耐久性仍是管理层会议室问题。[CU023, CU027, CU033, CU036, CU039, CU040]

留存 / 持久性信号表
指标 / 代理数值细分置信度重要性尽调要求
NRR整体审阅材料未公开核心持久性指标按主要产品线和客户规模带提供 NRR
GRR / logo 流失整体没有流失披露,精选案例可能夸大粘性提供年度 logo 流失率和总金额留存率
合同期限 / 续约条款整体续约机制会影响持久性和采购摩擦披露标准期限、年度预付占比和续约结构
免费到付费结构>90% 的公司免费使用 PostHog;付费占比未披露自助漏斗显示漏斗顶部覆盖很广,但已变现客户群可见度弱提供活跃付费客户数,以及云端 / 自托管拆分
创业生态重复使用YC 各届中有 65% 使用 PostHog 产品早期创业公司表明核心分发渠道内有强创始人 / 推荐闭环按创业公司批次展示留存或扩张
嵌入式工作流使用YC、Phantom、ElevenLabs、Arena 和 ResearchGate 都描述了持续的日常、每周或全年使用具名账户重复运营使用是实用的持久性代理提供具名参考客户的续约率和合作年限分布
独立评论信号G2 上 950 条评论,评分 4.5/5评论用户支持广泛用户认知和当前反馈量分享原始 NPS/CSAT 和评论征集政策
公开数量口径含混about 页 190254+ 个团队;pricing 页 60000+ 个客户整体直接把 logo 或注册数映射到付费客户经济性很危险披露一个权威付费客户指标,并保持口径一致

null 表示本次审阅的公开记录未披露该指标。非 null 行只是代理, 不应误认为正式留存或续约披露。

[CU001, CU003, CU004, CU009, CU023, CU027]
FU004: 留存 / 持久性 KPI 快照

紧凑展示最好的公开持久性信号,以及仍然挡住完整留存判断的测量缺口。

该 KPI 图有意把数字代理指标和未披露留存指标的“缺失”标记放在一起,因为公开记录不包含 NRR 或 GRR。

[CU003, CU004, CU009, CU041, CU045, CU056]

6.5 扩张路径可见,但集中度、独占性和证明质量风险仍未关闭

扩张证据是真实的。PostHog 的套件已经明显铺得很宽,多家具名客户称使用场景从核心分析延伸到功能开关、回放、调研、AI、可观测性, 或更深的产品埋点。Exa 把更多分析栈集中到 PostHog;ResearchGate 的使用规模已经需要定制打包和专人支持。这条公开扩张路径最清楚: 先以低摩擦自助切入,再在团队内铺开,随后带动更多模块采用;部分大客户再叠加企业级支持。这个路径也符合 Sacra 的判断: 用量计费先扩大工程团队采用面,再随事件量和产品挂载扩张。 未解决的风险在于,公开证据仍回答不了客户集中度和钱包份额问题。公开样本明显偏向 AI、创业公司和开发者主导客户。 Lovable 明确说自己在 PostHog 之外还同时使用多家可观测性厂商;这能证明 PostHog 被真实使用, 但也提醒我们:生产环境使用不等于供应商独占标准化。G2 也暴露出崩溃和文档上的产品质量担忧。最关键的是, PostHog 没有公开大客户收入占比、付费客户组合或续约经济性。因此,本章对客户喜爱度和扩张潜力给正面结论, 但对集中度,以及公开客户样本是否能代表全基数的货币化留存,仍需谨慎。[CU006, CU007, CU030, CU037, CU039, CU043]

集中度 / 扩张风险表
扩张驱动因素集中度 / 摩擦信号可能影响尽调路径
免费自助切入公开信息显示免费使用很广,但未披露付费客户转化和变现客户群结构logo 或团队数可能夸大耐久收入集中度质量索取按计划划分的活跃付费账户、云端 / 自托管拆分,以及免费到付费转化
多产品套件10+ 个产品和客户故事支持交叉销售,但未披露各批次的模块挂载和降配情况扩张可能真实存在,但各模块不均衡索取主要产品线的挂载、续约和流失数据
开发者和 AI 偏重的佐证集公开具名佐证偏向创业公司、AI 和开发者主导买方公开佐证可能低估企业非技术买方或受监管采购约束索取按垂直行业和客户规模划分的 ARR 与客户数
大规模企业支持ResearchGate 展示了定制套餐打法,但未披露前十大客户暴露少数大客户的重要性可能高于公开 logo 数暗示索取前十大客户的 ARR、续约日期和支持模式
供应商重叠Lovable 公开称在 PostHog 之外还运行多个可观测性供应商产品契合有用,但不一定意味着供应商独占或完整钱包份额按细分市场索取竞争替换和赢回数据
产品质量疑虑一名可见 G2 评论者提到崩溃和文档混乱可靠性或开发者体验问题会拖慢更多团队的深度采用索取总流失、支持响应指标和事件驱动降级数据
精选公开佐证具名故事很强,但仍由公司撰写,且很少披露合同金额或续约历史公开参考客户可能放大平均客户成功索取按 ARR 档、合作年限和部署是否仍在运行划分的活跃参考账户

本表聚焦于扩张证据确实存在、但承销仍依赖管理层披露而不能只靠公开证据的环节。

[CU006, CU007, CU030, CU039, CU043, CU054]
FU005: 扩张路径流

扩张看起来从免费或狭窄功能入口起步,进入更广模块采用;部分客户还会走向定制企业支持,但集中度和质量担忧仍未解决。

该流程是概念图,展示公开扩张证据如何连接到未解决的尽调问题,而不是每个账户披露的固定顺序。

[CU006, CU007, CU030, CU037, CU039, CU057]

6.6 图表

Chapter 07

07风险

7.1 按严重度排序的风险登记表

PostHog 的剩余风险不是一个孤立红旗,而是一组模式。公司公开披露了真实的隐私、安全和可靠性控制,但也公开记录了足够多的反向证据, 说明运营和信任风险并非理论问题。最重要的风险簇是安全和可靠性叠加:2025 年 11 月 npm 供应链遭入侵 5 小时, 2026 年 2 月确认发生日志数据丢失,以及一串公开事故复盘,覆盖工作流、功能开关、回放、调研和数据库迁移问题。 这段历史比任何单个漏洞更重要,因为 PostHog 正从分析扩展为更宽的 Product OS,组件更多、软件包暴露面更广、下游集成也更多。 投资含义是,企业信任才是需要承销的稀缺资产,不只是功能宽度。隐私传输和法律角色分配风险排在其次; 随后是围绕 GitHub 或 npm 的依赖风险,以及在 2026 年内把 90%+ 免费基数转化为 $100M ARR 目标的商业模式风险。[CR012, CR018, CR019, CR020, CR022, CR024]

排序风险登记表
排名风险发生概率严重性缓释成熟度残余暴露投资含义
1多产品面不断扩张,安全与可靠性倒退严重只有企业信任和事件管理尽调足够明确,才应承销。
2PostHog 控制与客户行为在跨境隐私或法律角色上错配中高中高需要围绕传输机制、客户误用边界和事件历史做法律尽调。
3GitHub 或 npm 供应链依赖事件复发严重中高事件若重演,将直接威胁客户信任和开发者驱动的分发引擎。
4日志、工作流、功能开关或回放不稳定,拖慢企业采用中高可能压制扩张,尤其在买方期待统一平台降低而不是放大运营风险时。
5PLG 变现跑输 2026 年 ARR 目标中低一旦转化或扩张失速,$1.4B 估值锚会更脆弱。
6自托管和套件蔓延削弱高端市场适配度中等可能放慢打入受监管或运营保守买方的速度。

严重性和发生概率来自保留来源的定性评估。本登记表按已披露缓释措施后的残余暴露排序, 而不是按控制前的原始风险排序。

[CR018, CR020, CR022, CR024, CR031, CR033]
FR001: 风险严重度矩阵

PostHog 最重的格子是高概率或中概率、且影响为高至关键的风险,尤其集中在信任、安全和货币化交汇处。

可能性和影响是基于截至运行日期留存公共来源的定性评估。

[CR031, CR043, CR051, CR052, CR053, CR054]
FR003: 风险传导图

主要传导路径从安全和可靠性问题进入信任,再进入企业采用、扩张和估值支撑。

[CR043, CR051, CR052, CR053, CR054, CR055]

7.2 监管、法律和隐私传输风险

法律和监管叙事强于一般私有软件供应商,但并非低风险。PostHog 对控制者与处理者角色、跨境传输、受监管使用支持和客户责任讲得异常明确。 这是利好,因为相比模糊的营销页面,客户拿到的是更清楚的工具。与此同时,同一批文件也保留了真实残余敞口。 PostHog 依赖有效的 DPF 认证,并在需要时以 SCC 兜底;这在今天可行,但公司仍暴露于未来传输机制被挑战, 或客户侧误用事件数据的风险。公司还用低责任上限在合同上限制自身财务敞口,并把所收集客户数据的合规责任放在客户身上。 PSA-2025-00001 的公开披露又添了一层信任褶皱:问题虽已解决,公司也称目前没有有效安全公告, 但该公告确认曾存在跨团队查询可见性缺陷,且美国历史窗口的一部分无法完全核验。这正是高度受监管买方会记住的事件。[CR001, CR002, CR004, CR005, CR006, CR007]

监管 / 法律风险登记表
暴露司法辖区 / 规则当前公开状态发生概率严重性缓释措施残余暴露尽调路径
跨境传输机制被挑战EU/UK/Swiss 个人数据传输;DPF 加 SCC 兜底DPF 参与仍有效,并记录了 SCC 兜底DPF 参与、SCC 兜底、EU 托管选项中:传输法律可能比合同变化更快索取当前传输影响评估,以及 DPF 充分性被挑战时的兜底计划。
控制者 / 处理者角色错配或客户误用GDPR、CCPA、HIPAA 及类似法律PostHog 文档说明了角色,并称客户仍要对自己收集的内容负责中高清晰文档、BAA、隐私控制、法律角色划分高:敏感数据采集不当的客户风险更高审查敏感客户的实施模式和默认数据最小化设置。
PSA-2025-00001 查询可见性暴露跨租户访问控制 / 隐私信任已解决,已无仍生效的安全公告中低team_id 修复、类似表审计、计划中的自动化测试中:历史窗口无法完全核验索取内部事件报告、客户通知和修复后测试覆盖。
受影响安装中的 CVE-2025-1520应用安全 / 自托管部署公开 CVE 目录条目仍可见补丁管理和云优先指引中高:补丁卫生薄弱的自托管运营方暴露更大索取修复版本、披露历史和任何官方修复说明。
合同责任上限和客户赔偿结构商业条款现行条款将累计责任限制在 $1,000 或一年费用二者较高者中等标准 SaaS 合同姿态中:若发生大型事件,补救可能偏薄按从供应商处只能有限追偿来建模下行。
受监管用途支持边界HIPAA / 隐私监管工作负载可提供 BAA 支持,但法律合规仍部分落在客户身上中高可提供 BAA、EU 托管、隐私控制中:误配置风险更偏运营,而非纯法律核查 BAA 或企业合同范围内包含哪些产品功能、地区和分包处理方。

本表只覆盖公开可见的法律和隐私风险,不涵盖私下的监管往来或未公开的企业合同条款。本文不构成法律意见。

[CR002, CR004, CR005, CR006, CR007, CR008]

7.3 运营、可靠性和安全风险

读 PostHog 的运营风险,最好从它自己的透明度入手。公开事故复盘列表和独立状态跟踪器显示,事故压力横跨多个组件, 而不是一个孤立子系统。2026 年 2 月的日志事故尤其重要:它既确认了客户数据丢失,又披露这个较新产品的备份深度只有三天, 日志集群也明显不如公司的核心数据栈成熟。2025 年 11 月的 Shai-Hulud 事故从信任角度更严重: PostHog 的软件供应链可能成为生态级 npm 蠕虫的分发向量。公司的响应看起来可信且细致, 但这起事件把供应链安全从假设性的姿态问题,变成了已经验证的失效模式。评论证据还显示,部分用户经历频繁崩溃和文档不清; 这与更大的叙事一致:产品宽度可能跑在可靠性打磨前面。风险不在于 PostHog 隐藏问题;风险在于问题覆盖的表面足够多, 已经会影响企业采用。[CR018, CR019, CR020, CR021, CR022, CR023]

运营 / 可靠性 / 安全风险矩阵
失效模式可能性严重性缓释成熟度剩余风险未解决缺口
npm 或 CI/CD 供应链再次失守极高中-高需要直接证据证明 Shai-Hulud 之后,工作流加固和密钥权限收窄已通过独立验证。
Logs 数据丢失,或新产品备份深度不足公开资料没有披露当前 RPO,也没有说明备份标准是否已与核心集群对齐。
应用、工作流、功能开关、回放和查询链路频繁出现组件事故中-高独立监测数据能看出故障面很宽,但看不到各组件错误预算,也看不到按客户分层的影响。
自托管补丁滞后或 CVE 暴露低-中中-高未留存 CVE-2025-1520 的公开修复说明,补救透明度仍不完整。
类似 PSA-2025-00001 的跨租户授权缺陷低-中历史日志缺口未解决,靠公开证据无法核验此前暴露的准确范围。
可靠性和文档摩擦先消耗信任,销售介入补救来得更晚中等评论证据真实但样本薄,无法公开衡量企业账户中的影响范围。

运营严重性按潜在信任、正常运行时间和客户环境影响判断,不只看内部工程投入。

[CR013, CR014, CR019, CR020, CR021, CR022]
FR002: 事故时间线与已披露暴露标记

公开披露事故集中在 2025 年 8 月至 2026 年 5 月,显示安全和可靠性事件有一定节奏,并非一次性异常。

[CR013, CR019, CR020, CR022, CR031, CR032]

7.4 合作伙伴、依赖、人员和模式风险

PostHog 的依赖和模式风险更偏软件,而不是特定供应商,但后果仍然重要。Shai-Hulud 事故复盘显示, 公司直接依赖 GitHub Actions 和 npm 发布工作流。Product OS 与 CDP 定位又带来另一层敞口:PostHog 想在多个产品面和数百个外部工具之间移动数据, 这增加了集成、合规或权限控制出错的触点数量。自托管进一步增加边界风险,因为 PostHog 仍把开源分发当作增长漏斗, 同时又明确说规模化时云端才是正确选择。人员和模式风险上,公开证据指向一家仍在拉伸到下一阶段的公司。 招聘页面称 200+ 人且仍在招聘,future 页面说 2026 年底 ARR 达到 $100M,Sacra 2026 年 2 月估计 ARR 约 $57.5M, 意味着仍有不小缺口要补。90%+ 的免费用户占比对分发非常好,但估值支撑取决于转化、扩张和可靠性同时站得住。[CR033, CR035, CR036, CR037, CR038, CR039]

合作伙伴 / 依赖风险表
依赖项交易对手 / 层级作用集中度失效情景严重性缓释剩余风险
CI 工作流权限链GitHub Actions 加机器人凭据构建、评审并发布 SDK 包特权工作流或令牌滥用重新打开供应链暴露面极高工作流评审加固和密钥管理调整该失效模式已经发生过一次,剩余风险仍然实质性。
包分发路径npm registry 和开发者包管理器安装把 JavaScript SDK 更新分发到 CI 和开发者环境恶意包发布攻破客户机器或流水线极高迁移到 Trusted Publisher,并采用更安全的包管理器默认设置即便技术控制改善,开发者信任受损的风险仍高。
云托管与传输机制栈AWS 区域,加 DPF 和 SCC 传输机制托管云数据,并决定司法辖区选择中-高区域事故或传输机制中断迫使迁移或合同流失EU 或 US 托管选项,以及已记录的传输机制区域、法律和供应商风险会叠加,因此为中-高。
集成与激活层CDP 目的地和数百个外部工具将数据移动或同步到下游系统权限、模式或下游服务故障把风险扩散到核心分析之外集中式平台和有文档记录的集成面端点越多,故障或误处理数据的路径越多,因此为中-高。
客户自管自托管客户运维团队及其安全姿态在 PostHog Cloud 之外运行 PostHog客户会要求自运维环境达到云端级可靠性中等-高PostHog 明确告知客户,规模化时优先选择云端预期错配仍会伤害品牌信任,因此为中。
供应商管理不透明分包处理方和 AI 功能供应商支撑存储、交付和可选 AI 功能Unknown关键供应商失效,或制造司法辖区 / 集中度问题DPA 覆盖的分包处理方清单和公司声明的极简供应商原则当前名单和集中度未在留存资料中公开,因此为中。

集中度评级是定性判断。留存的公开资料证明了关键平台依赖存在,但不能证明背后的具体商业或技术冗余。

[CR004, CR007, CR024, CR025, CR033, CR035]
人员 / 执行 / 经营模型风险登记表
角色或模型杠杆依赖或缺口可能性严重性缓释尽调路径
2026 ARR 目标相对可获得的外部 ARR 估计,公开目标偏激进品牌强、套件宽、PLG 分发索取到 2026 年底的月度 ARR 桥接表和转化队列。
免费层转化90%+ 公司免费使用 PostHog,变现取决于扩张和付费转化按用量计价,且有多个附加产品面复核免费转付费、付费账户结构和队列留存。
员工规模与招聘负荷200+ 人叠加持续招聘,会增加管理和事件响应复杂度中等-高远程优先的运营模式和透明招聘品牌检查领导梯队厚度、工程经理管理跨度和支持覆盖。
远程异步执行协作全球分布时,事件处理和受监管客户尽调会更难中等少开会文化和自主权能加快小团队交付复核事件战情室流程、全球接力覆盖和升级责任。
品类与产品聚焦PostHog 试图打包更多产品面,但功能管理和实验可能按用户画像分裂中等-高共享数据平面仍能撬动交叉销售按用户画像和产品线测试近期企业交易的赢单 / 输单原因。
拥挤市场里的估值锚$1.4B 估值叠加数千家竞争对手,留给执行失误的空间更小中等-高2025 年融资强劲,透明度差异化对标当前私募市场胃口,以及增长放慢情景下的下行空间。

本登记表把人员和经营模型风险放在一起,因为两者的公开证据都很薄,且高度绑定:执行质量决定模型假设是否可信。

[CR036, CR037, CR039, CR040, CR041, CR042]
FR004: 依赖关系图

PostHog 位于上游软件交付依赖和下游数据目的地之间,旁路还有客户自运营的自托管部署。

[CR024, CR025, CR035, CR038, CR049, CR050]

7.5 缓释、监控和投资逻辑破裂触发点

缓释因素真实存在,所以这不是一概回避的案子。PostHog 显得异常透明,发布事故复盘和安全公告,说明隐私角色, 提供 EU 托管和 BAA 支持,并在 npm 蠕虫之后以具体 SDLC 加固回应。这些都是有意义的利好,因为它们同时提高信任和尽调效率。 但投资上正确的用法是把这些缓释因素转化为更尖锐的监控,而不是据此忽略下行风险。值得盯的公开指标包括:宕机节奏, 是否出现新的事故复盘或安全公告,DPF 参与状态是否保持有效,关于崩溃或文档的评论投诉是否恶化, 以及招聘和产品蔓延是否继续跑在公开 ARR 轨迹前面。如果 PostHog 再次发生重大跨租户暴露、再次供应链入侵, 或在较新的日志表面之外出现核心数据丢失事故,投资逻辑就会破裂;任何一种情况都会削弱这样一个论点: 公司的透明文化正在转化为可持续更安全、更可靠的平台。[CR025, CR026, CR031, CR051, CR057, CR058]

缓释与监测表
风险可监测触发因素阈值 / 事件当前公开基线行动含义
跨租户隐私或授权回退新的安全公告、事后复盘或客户通知任何新确认的跨租户暴露目前没有活跃公告,但历史上有 PSA-2025-00001立即重新审视投资逻辑;要求披露根因、影响半径和控制验证。
供应链攻破复发安全公告、npm 包下架,或 PostHog 发布事故文章任何新的恶意发布或客户环境被攻破Shai-Hulud 已解决,发布工作流已加固新控制通过独立复核前,暂停为开发者信任假设背书。
运营可靠性滑坡Logs、工作流、功能开关、回放或应用再次发生长期影响客户的事故一个季度内公开承认两起或以上事故,或任何核心数据丢失事件2025-2026 年近期事故历史可见重切企业扩张假设,并索取 SLO、备份深度和支持人员配置。
传输机制或隐私治理承压DPF 状态变化、SCC 遭挑战,或重大企业隐私投诉DPF 不再有效,或主要客户因传输担忧暂停DPF 有效,且 SCC 后备机制已有文档升级法律尽调和合同风险建模。
PLG 变现失手ARR 更新、招聘节奏和外部估计进一步偏离 $100M 目标公开轨迹显示,到 2026 年底该目标在结构上难以达到公开目标是 2026 年达到 $100M ARR;外部 2026 年 2 月估计约 $57.5M加码前压力测试估值、烧钱速度和招聘假设。
自托管或企业适配度侵蚀更多云优先警示、评论摩擦,或因可靠性和支持边界导致的输单自托管或受监管买家因 PostHog 支持姿态而拒绝采购,且形成输单模式云优先指引和部分评论摩擦已经公开在支持和 SLA 边界更清楚前,限制上行情景中的高端市场扩张假设。

这些阈值是投资人启发式指标,不是合同约束。它们用于把公开信号转成决策点,而不是预测精确财务结果。

[CR012, CR020, CR022, CR026, CR031, CR035]
FR005: 监控项与打破投资假设的 KPI

公开可观察指标重点看信任、宕机节奏、传输状态连续性,以及增长目标能否跟上已披露的规模信号。

[CR005, CR019, CR031, CR036, CR039, CR041]

7.6 图表

Chapter 08

08估值

8.1 投资逻辑与反向逻辑

乐观逻辑很直接,也值得认真对待。PostHog 搭出了一个可信的开发者优先平台:用量计费透明、产品宽度足够、免费分发漏斗很大, 还有客户证据表明,团队能把实验和用户引导改进转化为可衡量的业务结果。公司所在品类也仍被多家分析机构描述为十几个百分点中段到高段的 CAGR 增长, SlashData 的 48.4 million 名全球开发者基数进一步说明技术买方池子足够大。换句话说,PostHog 有一条可信路径, 可能长成有规模的 PLG 开发者基础设施公司,而不只是小众分析工具。反向逻辑是,当前价格已经要求投资者承销这个结果, 却没有给出公开市场要求的分母质量。最新报道的 $1.4 billion 融资,无法和已披露 ARR 基数、NRR、毛利率或干净股权结构表对照。 只用官方 >$50 million 收入下限,倍数也不低于约 28x;即使用 Sacra 的 2026 ARR 估计,仍约 24x。 这个水平高于 Datadog 的公开市场溢价倍数,也远高于 Atlassian 或 Amplitude。在这个价格上,PostHog 不只是产品质量下注; 也是对转化质量、利润率质量和治理质量的下注。[CV004, CV006, CV007, CV008, CV012, CV014]

投资逻辑 / 反向逻辑表
维度乐观情景反向逻辑什么会改变判断
平台广度Product OS 的广度支持整合多条产品工作流。广度也可能掩盖过多产品面带来的毛利复杂性和可靠性拖累。按产品披露毛利率,并按产品面展示稳定可靠性。
变现透明按量计价支持 PLG 采用和落地后扩张。90%+ 免费基数意味着,logo 数本身不能证明付费经济性。披露免费转付费和产品附加率。
市场背景独立报告仍显示品类双位数增长和庞大开发者 TAM。公开可比公司的纪律重置后,市场增长不保证溢价倍数。展示高质量留存和高效增长。
公开可比公司支撑Datadog 证明高溢价开发者工具倍数仍然存在。Atlassian 和 Amplitude 说明,当规模、成熟度或品类质量不同,倍数压缩会很快。让可比公司集合更接近高溢价开发者工具,而不是分析软件。
资本策略投资人反复支持,且 2025 年估值上台阶,说明融资动能真实。优先股堆叠、新股与老股拆分、真实稀释仍未披露。打开股权结构表和 2025 年条款清单。
退出路径如果 PostHog 达到 $100M+ 收入并披露上市级 KPI,未来 IPO 有可能。信任事件或增长失手会快速推迟或缩小退出选项。维持可靠性并完成收入里程碑。

反向逻辑主要针对价格质量和披露质量,不是否认产品价值。

[CV012, CV015, CV017, CV021, CV024, CV026]
FV001: 从公开证据到建议的情景桥接

公开证据支撑强公司的叙事,但还不足以在最新私募价格上清晰承销。

[CV015, CV020, CV021, CV024, CV025, CV043]

8.2 估值背景、可比公司质量和入场纪律

融资记录显示真实动能,也暴露分母问题。公开证据支持 2025 年 6 月官方 $70 million 融资、$920 million 估值, 之后独立报道又称公司以 $1.4 billion 估值融资 $75 million。这次重估发生时,多位观察者都说后期 SaaS 投资已经比 2021 年高点更有纪律。 公开软件倍数仍高度分散,但市场不会宽容那些拿不出留存质量或利润率质量的公司。最相关的公开可比集不是泛消费软件, 而是优质开发者基础设施和相邻分析品类的组合。Datadog 仍享受溢价倍数,因为它同时有强增长和监管文件级透明度。 Atlassian 规模巨大、毛利率优秀,但收入倍数低得多。Amplitude 是最接近的公开品类类比,交易倍数只有低个位数。 PostHog 也许因为增长和产品宽度值得私人市场溢价,但当前名义价格已经要求投资者像业务已经拥有 Datadog 级市场支撑一样付费, 却没有 Datadog 级披露。因此,入场纪律必须从两个条件开始:价格明显更好,或证据明显更好。[CV001, CV002, CV003, CV013, CV024, CV025]

当前融资与稀释背景
项目公开披露暗示什么稀释 / 悬挂风险解读仍未知
Series D(2025 年 6 月)官方披露由 Stripe 领投,$70M 融资,估值 $920M。成为独角兽前已有真实新股资金需求。根据投前 / 投后口径不同,稀释可能在个位数到低双位数。准确股数、证券类型和清算优先权。
据报道 Series E(2025)独立报道称,Peak XV 领投 $75M,估值 $1.4B。公司估值快速上台阶,进入独角兽区间。如果主要是新股,名义稀释可能看起来在中个位数,但公开证据不足以确认。估值是投前还是投后,以及谁出售了股份。
员工流动性 / 要约收购官方材料强调员工流动性,招聘页面也突出老股交易 / 要约收购。并非每一美元披露融资都应视为纯粹延长现金跑道。老股交易量能缓解创始人 / 员工压力,但不会一比一增加资产负债表现金。2025 年两轮中新股与老股的拆分。
优先股堆叠未找到公开股权结构表或条款包。仅靠公开证据无法干净建模回报测算。潜在优先权悬挂仍未知。清算顺位、参与权和投资人保护条款。
入场纪律含义当前名义价格必须对照公开可比公司质量判断,而不能只看产品质量。这一轮需要溢价级私营指标,或非常强的前瞻路径。作出投资判断前,需要更好价格或更强证据。董事会级 KPI 包和已签署条款清单。

本表保留分母和条款缺口,不强行给出虚假的稀释点估计。

[CV001, CV002, CV003, CV041, CV052]
可比估值表
公司 / 参考状态当前分母或规模信号估值倍数 / 价格背景重要性关键限制
PostHog私营官方 >$50M 收入下限;Sacra 估计 2026 年 2 月 ARR 为 $57.5M基于官方下限为 <$28x;按 Sacra 估计,在 $1.4B 估值下约 24x标的公司;显示当前价格多大程度押注未来执行。没有公开 NRR、毛利率、股权结构表或经审计分母。
Datadog上市FY2025 收入 $3.427B;Q1 2026 收入 $1.006BSaaSValuation.io 显示 18.3x ARR高溢价上市开发者工具参照,公司增长强,披露达到监管文件级透明度。规模大得多,可观测性版图也比 PostHog 更广。
Atlassian上市公司FY2025 收入 $5.215B;毛利率约 83-85%隐含收入倍数为低个位数说明即便是超大 PLG 软件平台,交易倍数也可能远低于高溢价开发者工具区间。协作 / 生产力套件,并非纯分析可比公司。
Amplitude上市公司FY2025 收入 $343M;2026 Q1 收入 $93.5M隐含收入倍数约 2.3x最接近的上市产品分析品类可比公司。平台更窄,增长也慢于 PostHog 的私募叙事。
2026 SaaS 基准区间市场参照上市 SaaS 中位数约 6x-7x;私募 SaaS 为 3x-7x ARR,中位数 4.5x全部 SaaS 平均约 10.4x,但分布很散为单一公司之外的后期估值纪律提供参照。不是直接经营可比组。

可比行故意混合类别接近度和披露质量,因为判断一轮私募估值时,两者都重要。

[CV024, CV025, CV027, CV028, CV030, CV033]
FV003: 可比定位图

按当前可见分母,PostHog 的现有私募定价甚至高于享受溢价的公开可比公司。

增速和倍数混合了公开季度数据与公开基准筛选;PostHog 使用 Sacra 的 ARR 估计和报道的私募轮次。

[CV025, CV026, CV027, CV030, CV032, CV033]

8.3 乐观、基准和悲观情景

由于 PostHog 没有公布权威的当前 ARR 分母,估值更适合用前瞻情景区间框定,而不是给出虚假精确的点估计。 悲观情景假设公司大幅落后于公开目标,2027 年收入只长到 $75 million 至 $90 million, 并在 5x 至 7x 出清,因为后期 SaaS 投资者会惩罚留存或信任证明缺失。基准情景给管理层完成或小幅超过公开 $100 million 目标的信用, 但只给 7x 至 10x,因为公开市场纪律仍紧,溢价区间仍需要比今天更强的披露。乐观情景假设 PostHog 显著跑赢公开计划, 宽套件货币化保持非常强,并拿到 10x 至 14x 区间,接近分层市场里最强的软件公司。按这些假设, 当前 $1.4 billion 名义估值已经落在乐观结果低端附近,而不是基准情景中点。这不能证明该轮融资错了, 但意味着新投资者接受这个价格前,需要异常强的非公开证据。[CV005, CV006, CV038, CV039, CV044, CV045]

情景假设表
情景概率信号2027 收入假设倍数区间隐含价值区间必须成立的前提
悲观真实风险情形75M-90M5x-7x0.4B-0.6B免费用户转付费不及预期、信任成本上升,且公开市场不给予高溢价区间。
基准公开证据下最平衡的情形100M-120M7x-10x0.7B-1.2B管理层达到或小幅超过公开目标,但仍缺少 Datadog 式高溢价指标。
乐观需要溢价证据130M-160M10x-14x1.3B-2.2BPostHog 显著超过公开计划,并证明一流的留存、转化和利润率质量。

情景只是基于公开证据和明确分母假设的估值区间示意;它们不能替代真实数据室后的逐项建模估值。

[CV044, CV045, CV046, CV047, CV056, CV057]
FV002: 按情景划分的估值区间

据报道的最新轮次落在牛市情景区间低端附近,而不是基准情景区间中点。

[CV044, CV045, CV046, CV047]
FV005: 估值对情景中点的敏感性

基准情景中点仍低于当前轮次,牛市情景中点也只是略高于它。

[CV044, CV045, CV046, CV047]

8.4 建议、信心、退出准备度和投资逻辑破裂触发点

建议是继续研究,不是因为 PostHog 缺产品质量,而是因为公开证据还不足以支撑有纪律地支付当前价格。 这正是那类“公司可能好,但估值可能错”的标的:平台范围宽、PLG 分发强、投资人质量也真实。但承销缺口仍太大。 投资者不知道两轮 2025 融资后的实际股权结构堆叠,不知道新股和老股的精确拆分,不知道超过 90% 公司免费这个基数的付费转化质量, 也不知道扩张中产品组合的利润率结构。因此,退出准备度仍在成形,而非成熟。如果公司弥合当前 >$50 million 下限与公开 $100 million 目标之间的差距, 并披露接近上市公司口径的留存、利润率和可靠性指标,就有能力变得具备 IPO 条件。一旦信任事故反复、管理层严重错失 2026 目标, 或后续融资把估值重置到当前名义估值之下,投资逻辑就会破裂。在这些关口变化前,价格纪律比欣赏更重要。[CV040, CV041, CV042, CV048, CV049, CV050]

建议摘要表
维度评估为何落在这里什么会改善它
投资建议继续研究公开证据不足以干净支撑据称 $1.4B 的价格。披露分母、留存、毛利率和融资条款。
信心资料基础在定价、融资轮次、可比公司和市场背景上较强,但私营公司经济性证据偏薄。补充类审计 KPI 披露或数据室确认。
风险评级信任事件、转化质量和股权结构表不透明都会直接影响回报测算。展示 2026 年内稳定的可靠性和留存。
估值立场充分到偏贵当前轮次已经接近公开市场估值区间下乐观情景的低端。给出明显更好的入场价,或证明能拿到异常高的溢价区间。
公开分母质量不完整公开信息只有官方 >$50M 收入下限、$100M 目标和一个外部 ARR 估计。发布当前 ARR,或足够用来三角验证的 KPI。
入场纪律等待证据或价格这是一家高质量公司,但价格敏感度比叙事质量更重要。更低的老股价格 / 平轮,或更完整披露,会改变判断。
退出准备度正在成形,但未就绪规模抱负真实,但上市公司级披露和可靠性信号还不到位。达到上市公司式指标,并守住信任。
主要尽调关口股权结构表 + NRR + 毛利率这三项决定其是否值得溢价倍数。拿到投资意向书、队列指标和毛利率桥接。

本表总结当前价格下的投资判断,不抽象评价公司质量。

[CV048, CV049, CV050, CV052, CV053, CV054]
打破投资逻辑的触发项清单
触发项阈值或事件为什么重要行动含义
信任倒退再次发生重大安全或数据丢失事件信任变弱时,高溢价软件倍数会很快压缩。暂停承销,直到补救措施和客户影响明确。
增长落空管理层明显未达到公开 2026 收入目标当前价格已经依赖强劲的远期增长。把情景重置到悲观侧,并重新审视估值。
留存不及预期私下 NRR 或流失数据低于高溢价区间常模高溢价私募 ARR 倍数需要强扩张质量。未重新评级前,不应支付溢价倍数。
轮次重置后续轮次或老股交易价格显著低于 $1.4B这会直接证伪当前定价逻辑。将当前轮次视为完全减记,并重新评估。
股权结构悬置风险优先股堆叠或老股占比高的结构比预期更差即便经营执行扎实,回报数学也可能失效。重新接触前,按完整条款建模下行情景。

该清单聚焦会推翻当前价格支撑的事件,而非泛化的经营噪音。

[CV041, CV042, CV047, CV051, CV052]
最终尽调索取清单
主题缺失证据为什么重要负责人 / 尽调路径
股权结构表和条款当前股数、证券类型、清算优先权、参与权、按比例认购权这些输入决定回报数学和下行保护。索取最新股权结构表和已签署条款清单。
新股与老股拆分2025 年每轮融资中实际进入资产负债表的金额,以及流向出售股东的金额现金跑道和稀释不能只靠新闻标题推断。审阅交割备忘录和资金流向表。
留存质量按产品和客户细分划分的 NRR、总流失率和队列扩张高溢价倍数需要证明增长质量可持续。审阅董事会 KPI 包和客户队列表。
付费转化免费转付费转化率、附加率和大客户货币化路径90%+ 的免费用户基数会改变 logo 转化为企业价值的方式。检查账单队列和 PLG 漏斗分析。
毛利率云毛利率、自托管 / 云混合,以及按产品线拆分的基础设施成本没有分母和利润率质量,就无法做上市可比比较。索取财务桥表和托管成本分摊。
可靠性 / 信任2025-2026 事故后企业客户留存,以及事后复盘趋势线信任冲击能迅速改变退出时点和折现率。抽样受影响账户,并查看事故跟进报告。

这些索取项是把公开证据视角转成可定价承销案例所需的最低数据集。

[CV052, CV053, CV054]
FV004: 建议 KPI 快照

公司在产品广度和市场相关性上得分较高,但承销卡点在分母质量和条款透明度。

[CV048, CV049, CV050, CV052, CV053, CV054]

免责声明

本报告是基于公开证据的尽调快照,不构成投资建议。重要的财务、法律、技术和合同事实仍未公开;作出任何投资决策前,应直接向管理层核验,并查阅一手文件。

证据索引

结论
编号陈述可信度来源
CO001 PostHog publicly identifies PostHog Inc. (formerly Hiberly Inc.) as the controller for its hosted services, websites, and self-managed installations covered by its privacy policy. SO011
CO002 PostHog's legal materials list 2261 Market Street #4008, San Francisco, CA 94114 as the company address and also reference UK and Germany subsidiaries or affiliates. SO011, SO012, SO013
CO003 PostHog says James Hawkins and Tim Glaser founded the company on January 23, 2020. SO006, SO016
CO004 Public company and partner profiles identify James Hawkins as CEO and Tim Glaser as CTO. SO016, SO018
CO005 PostHog currently markets itself as devtools and product data infrastructure for building successful products rather than as a single analytics tool. SO003, SO008
CO006 PostHog's public docs and repository materials enumerate a suite that includes product analytics, web analytics, session replay, feature flags, experiments, error tracking, surveys, a data warehouse, data pipelines, AI observability, and workflows. SO014, SO015
CO007 The about page says PostHog has grown into 10+ paid products and is used by 190,254+ teams. SO001
CO008 The pricing page describes PostHog as a usage-based product with transparent published rates and generous free tiers across products. SO004
CO009 PostHog says it wants to offer every tool engineers need in one place and price the platform more like a utility than a high-touch enterprise software vendor. SO001, SO008
CO010 Contrary says Hawkins built commercial and data experience at Arachnys while Glaser built product and R&D experience there before founding PostHog. SO018
CO011 Y Combinator and the handbook say the founders pivoted multiple times before launching PostHog during the YC W20 batch. SO006, SO016
CO012 The public DPA preview identifies Charles Cook as VP Operations, making him one of the few non-founder executives explicitly named in reviewed public materials. SO012
CO013 The careers page says PostHog is adding 25 team members and organizes work around small autonomous teams. SO009
CO014 The reviewed role page on PostHog careers is remote and scoped across time zones from GMT +2 to GMT -8. SO009
CO015 A third-party remote-work profile describes PostHog as fully remote, async-first, with no offices and employees distributed across 25+ countries. SO019
CO016 Because reviewed public materials center on the founders and a narrow set of signatories rather than a full board or executive directory, public governance visibility is limited and key-person dependence remains meaningful. SO002, SO012, SO013
CO017 The handbook says PostHog launched on Hacker News in February 2020 after four weeks of coding and reached over 300 deployments within a couple of days. SO006
CO018 PostHog says it raised a $3.025M seed round in April 2020 from YC Continuity and 1984 Ventures. SO006, SO022
CO019 PostHog says it raised a $9M Series A in December 2020 led by GV. SO006
CO020 PostHog says it raised a $15M Series B in June 2021 led by Y Combinator. SO006
CO021 In its Series D announcement, PostHog said it raised $70M at a $920M valuation led by Stripe, with YC, GV, and Formus Capital participating. SO010, SO023
CO022 Peak XV's company page says it partnered with PostHog in 2025 and identifies the company as founded in 2020. SO021
CO023 Economic Times and Entrepreneur India reported that PostHog raised $75M at a $1.4B valuation in September 2025 with Peak XV leading the round. SO020, SO023
CO024 The about page says 65% of every Y Combinator batch uses PostHog's products. SO001
CO025 The pricing page says more than 90% of companies use PostHog for free. SO004
CO026 The Y Combinator profile says PostHog had been averaging roughly 10% monthly revenue growth at the time of the interview. SO016
CO027 The future page says PostHog wants to hit $100M in annual revenue by the end of 2026 and thinks it needs about 7% monthly growth to get there. SO007
CO028 The careers page says PostHog held its first employee secondary in 2024 and executed its first tender offer in 2025. SO009
CO029 The careers and people pages imply that PostHog currently sees itself as a 200+ person company. SO009, SO002
CO030 A third-party remote-work profile says PostHog has approximately 110 employees, materially below the 200+ signal on official pages. SO019
CO031 Public headcount signals therefore conflict and should be treated as a range rather than a canonical single figure. SO009, SO002, SO019
CO032 Sacra estimates that PostHog reached about $57.5M ARR in February 2026 and about $182M of total funding. SO017
CO033 Because official pages reviewed here do not publish a canonical current ARR or cumulative funding total, third-party figures should be treated as estimates rather than settled company facts. SO006, SO010, SO017
CO034 The handbook says PostHog had 25 people in 10 countries by June 2021, 30 people in 12 countries by September 2021, and 38 people by December 2022. SO006
CO035 The handbook says PostHog grew revenue 6x in 2022, set a $10M ARR goal, and targeted 70% gross margin. SO006
CO036 PostHog's public strategy pages say the company wants to help engineers make product decisions and consolidate fragmented tools into one source of truth. SO001, SO008
CO037 PostHog's security advisories page discloses a resolved medium-severity August 2025 incident where an overly permissive SQL table exposed query text across unrelated teams. SO025
CO038 PostHog's public post-mortems page lists customer-affecting 2025-2026 incidents including logs data loss, feature-flag outages, replay SDK issues, and a 2025 supply-chain attack. SO024
CO039 PostHog's legal materials emphasize GDPR-related handling, encryption, SOC 2 and HIPAA claims, and formal legal governance, but they expose compliance mainly through legal artifacts rather than a simple trust-center narrative. SO011, SO012, SO013
CO040 PostHog's GitHub materials say the main repository is open source under MIT expat except enterprise code, and the company also open-sources its handbook. SO014, SO008
CM001 PostHog's products page says the platform combines the tools needed to collect and analyze product usage data and to build and ship new features in one place. SM001, SM007
CM002 PostHog Product OS documentation says the platform ties analytics, replay, feature flags, experiments, surveys, data warehouse, data pipelines, and SQL access into a shared data foundation. SM007, SM001
CM003 PostHog's CDP page says product events can drive marketing activation, sales enablement, data-science workflows, and operations or monitoring tasks. SM008
CM004 PostHog's GA4 comparison page says GA4 is primarily built for web analytics and marketing attribution, while PostHog spans product analytics plus replay, feature flags, A/B testing, error tracking, and surveys. SM006
CM005 The same GA4 comparison page says PostHog lets customers choose EU or US hosting, a meaningful differentiator for privacy-sensitive buyers. SM006
CM006 Contrary describes PostHog as a developer-first platform that tries to consolidate analytics, session replay, feature flags, A/B testing, and adjacent capabilities for engineering-led product teams. SM009
CM007 PostHog's customer pages show adoption across SaaS, education, devtools, AI, crypto, identity, and public-sector-adjacent organizations rather than one single vertical. SM003
CM008 The Y Combinator customer story explicitly lists leadership, engineering, and product as users of PostHog. SM004
CM009 The Hasura customer story explicitly lists engineering, UX, and marketing as users of PostHog. SM005
CM010 Grand View Research says the global product analytics market was worth USD 19.92B in 2024 and could reach USD 58.78B by 2030. SM013
CM011 Expert Market Research says the global product analytics market reached USD 12.03B in 2025 and could reach USD 49.09B by 2035 at a 15.1% CAGR. SM014
CM012 Mordor Intelligence says the product analytics market should grow from USD 11.39B in 2025 to USD 13.04B in 2026 and USD 25.73B by 2031 at a 14.55% CAGR. SM015
CM013 Mordor says large enterprises held 60.18% of 2025 product-analytics revenue while SMEs are projected to grow faster at a 19.7% CAGR. SM015
CM014 Mordor says cloud accounted for 87.6% of the product analytics market in 2025 and highlights cloud-native cost advantages plus privacy-safe enrichment as growth drivers. SM015
CM015 StartUs says the broader advanced analytics market could grow from USD 57.01B in 2025 to USD 139.92B by 2029, showing that PostHog participates in a much larger adjacency than standalone product analytics alone. SM016
CM016 SlashData estimates there were 48.4 million developers globally in Q3 2025, providing a large technical-user population lens for PostHog's addressable base. SM017
CM017 dbt Labs says 45% of surveyed analytics teams planned to increase AI-tooling investment and 38% planned to increase data quality or observability investment over the next 12 months. SM019
CM018 dbt Labs says 30% of respondents reported budget increases and 40% reported headcount increases for data teams. SM019
CM019 dbt Labs says poor data quality was cited by more than 56% of respondents as the most frequent challenge for data teams. SM019
CM020 JetBrains reports that 66% of developers do not believe or are not sure that current metrics reflect their real contribution. SM018
CM021 Mixpanel's 2026 benchmark write-up says growth has moved decisively inside the product rather than being primarily bought through external channels. SM010
CM022 Mixpanel says acquisition without activation no longer counts and that retention is the most dependable growth lever for digital products. SM010
CM023 Mixpanel says analytics is becoming an active system that predicts behavior, identifies friction, personalizes experiences, and aligns teams around outcomes. SM010
CM024 Mixpanel's experimentation guide says product experimentation spans A/B tests, multivariate tests, feature-flag rollouts, and phased releases tied to retention, adoption, and task-success metrics. SM011
CM025 The same guide says experimentation needs enough traffic, clear hypotheses, and can be inappropriate where compliance or ethics prevent differentiated experiences. SM011
CM026 Forrester says feature management is primarily developer-led progressive delivery, while experimentation increasingly serves product, marketing, and experience-design personas. SM020
CM027 Forrester says the two use cases are increasingly being served by separate technology markets rather than one stable combined category. SM020
CM028 VWO markets experimentation to product managers, engineers, growth marketers, and UX or analytics teams. SM021
CM029 Statsig's customer page features quotes from product, engineering, data, and executive roles describing experimentation and feature management as core decision infrastructure. SM022
CM030 Amplitude's product analytics guide says effective product analytics starts with 5-10 critical events, a tracking plan, engineering instrumentation, a North Star metric, and lifecycle dashboards. SM012
CM031 Heap's buyer guide says product analytics helps teams track both known user behavior and behavior they would not otherwise notice, using end-to-end journey insight. SM023
CM032 PostHog's pricing page says more than 90% of companies use the platform for free and only pay as usage expands across products. SM002
CM033 PostHog's Product OS documentation says ClickHouse-backed infrastructure and SQL access reduce the need to engineer data across many separate vendors. SM007
CM034 Y Combinator says it preferred PostHog over GA because Google Analytics lost roughly 30% of data to adblockers or third-party cookies and PostHog autocapture reduced setup work. SM004
CM035 Hasura says PostHog helped it improve onboarding conversion by 10-20% after funnel analysis and replay exposed precise drop-off points. SM005
CM036 Grand View and Mordor both say product analytics demand is driven by customer-behavior tracking, UX optimization, AI-enabled analytics, and self-service or cloud advantages. SM013, SM015
CM037 PostHog's market overlaps with feature management, replay, CDP, and routing categories because buyers increasingly want one system to instrument, decide, release, and activate instead of stitching multiple vendors together. SM001, SM007, SM008
CM038 The effective buyer is often multi-threaded: technical teams instrument and govern the platform, product or growth teams consume insights, and leadership or functional owners justify budget based on activation, retention, and ROI. SM004, SM005, SM020, SM021, SM022, SM012
CM039 Public sources do not isolate a clean PostHog-specific SOM inside overlapping product-analytics, experimentation, feature-management, and CDP markets.
CP001 PostHog publicly presents Product OS as a bundled stack spanning product analytics, web analytics, session replay, feature flags, A/B testing, surveys, and a data warehouse foundation. SP019, SP021
CP002 PostHog pricing discloses a free tier with 1 million analytics events, 5,000 session replay recordings, 1 million feature-flag requests, and 1 million data-warehouse rows before paid overages apply. SP018
CP003 PostHog also positions CDP sources, destinations, realtime transformations, and workflow triggers as part of the same product data stack rather than a separate vendor layer. SP020, SP019
CP004 PostHog says more than 90% of companies use the product for free and that the company serves over 60,000 customers. SP018
CP005 Mixpanel publicly markets analytics, web analytics, session replay and heatmaps, experiments and feature flags, metric trees, and warehouse connectors from the same platform family. SP002, SP001
CP006 Mixpanel pricing publishes a free tier capped at 1 million monthly events and a growth plan charging $0.28 per 1,000 events after the free allowance, with Enterprise handled separately. SP001
CP007 Amplitude pricing presents a multi-product suite including analytics, guides and surveys, feature experiment, web experiment, activation, AI feedback, AI assistant, and session replay. SP003
CP008 Amplitude exposes a self-serve Plus tier starting at $49 per month, while Growth and Enterprise remain custom-priced. SP003
CP009 Heap says it has joined Contentsquare and is used by more than 10,000 companies, making it both a product-analytics tool and part of a broader digital-experience stack. SP006
CP010 Heap pricing centers on analytics first, keeps session replay as an add-on, and moves data-warehouse integration and region-specific storage up to higher plans. SP005
CP011 Fullstory positions itself as a behavioral-data and analytics platform built around automatic interaction capture, AI-powered insight generation, in-product guides and surveys, and activation workflows. SP008, SP009
CP012 Fullstory publishes plan categories but still routes buyers to request pricing and demos instead of listing self-serve dollar rates for comparable enterprise use. SP009
CP013 LaunchDarkly describes its pricing scope around runtime control, feature flags, progressive delivery, experimentation, observability, and agent control. SP010
CP014 LaunchDarkly's pricing page exposes a Developer plan that is free to start and Enterprise or Guardian options that shift to tailored pricing for scaled engineering teams. SP010
CP015 Statsig's homepage markets 5 or more integrated products in one platform, including product analytics, experimentation, feature management, session replay, web analytics, and developer configuration layers. SP013, SP012
CP016 Statsig pricing gives a Developer tier with 2 million metered events each month at no charge and includes gates, configs, experimentation, and analytics in that entry tier. SP012
CP017 Statsig customer proof says at least one buyer evaluated Optimizely, LaunchDarkly, Split, and Eppo before choosing Statsig for end-to-end integration spanning data ingestion, stats engine, and experimentation workflows. SP014
CP018 GrowthBook markets experimentation, feature flags, product analytics, warehouse-native deployment, integrations, and security and compliance from one platform and says it is trusted by more than 3,000 companies. SP016
CP019 GrowthBook pricing offers a free starter with up to 3 users, unlimited feature flags, unlimited experiments, and cloud or self-hosted deployment, while Pro starts at $40 per seat per month. SP015, SP016
CP020 Harness documentation shows Split has been folded into Harness Feature Management and Experimentation, with free-plan onboarding, flag targeting, experimentation, and release monitoring now framed inside Harness FME. SP017
CP021 Google Analytics still positions itself as a free way to understand the customer journey and improve marketing ROI through integrations across Google advertising and publisher tools. SP030
CP022 VWO positions itself as an end-to-end experimentation platform centered on journey optimization, unified customer data, and integrations rather than a full developer product stack. SP029
CP023 Forrester describes feature management and experimentation as a combined capability layer that now spans both software delivery and product management. SP028
CP024 Public vendor pages now show real category convergence: Mixpanel, Amplitude, Statsig, GrowthBook, and PostHog each advertise some combination of analytics, experimentation, and feature controls rather than a single narrow job. SP002, SP003, SP013, SP016, SP019
CP025 Cotera's hands-on comparison says PostHog can replace LaunchDarkly and FullStory alongside core analytics, cutting two vendors and about $3,000 per month in the reviewed stack. SP024
CP026 The same Cotera review says PostHog is weaker for non-technical product managers because its broader interface and HogQL workflow assume more technical comfort than Amplitude. SP024
CP027 Startupik says Amplitude is usually the strongest fit for enterprise teams because of analytics depth, governance, and cross-team scalability. SP026
CP028 Startupik says Mixpanel is often easiest for non-technical teams and PostHog is strongest for developers. SP026
CP029 Fungies describes Heap as the zero-setup autocapture option and PostHog as the open-source, self-hostable bundle that combines flags, replay, experiments, surveys, and HogQL. SP025
CP030 Techno Pulse frames Amplitude as enterprise-growth oriented, Mixpanel as product-led SaaS friendly, PostHog as dev-focused for startups, and Heap as a fit for teams that do not want manual tracking. SP027
CP031 GA4 remains the status-quo default for many buyers because the official product is free and tightly tied to the Google marketing ecosystem, even though it is not marketed as a unified replay-flags-experiments suite. SP030, SP022
CP032 Open-source and self-hosted deployment are meaningful differentiators for PostHog and GrowthBook, but they also reduce hard vendor lock-in because the buyer can keep running the stack outside the vendor cloud. SP018, SP015, SP016
CP033 PostHog's integrated product breadth means a buyer can cover analytics, replay, feature flags, experimentation, surveys, CDP-style routing, and warehouse workflows inside one contract and one login. SP019, SP020, SP021
CP034 Best-of-breed alternatives still preserve segment advantages: Fullstory in replay-rich behavioral analysis, LaunchDarkly in runtime control and approvals, and Amplitude in PM-friendly analysis and governance. SP008, SP010, SP026, SP024
CP035 Public pricing transparency is materially better at PostHog, Mixpanel, Statsig, and GrowthBook than at Fullstory or the higher enterprise tiers of Amplitude and LaunchDarkly. SP018, SP001, SP012, SP015, SP009, SP003, SP010
CP036 Free-entry economics are becoming table stakes because PostHog, Mixpanel, Statsig, GrowthBook, Google Analytics, Heap, and LaunchDarkly all publish some free tier, free start, or free trial path. SP018, SP001, SP012, SP015, SP030, SP005, SP010
CP037 Distribution power still favors incumbents with ecosystem defaults: GA4 via Google's ads stack, LaunchDarkly via enterprise release governance, and Amplitude via PM or executive reporting workflows. SP030, SP010, SP026
CP038 Multi-homing remains feasible because analytics, replay, feature management, and experimentation are still sold separately by specialists and also bundled by suites, so buyers can mix layers instead of accepting one locked stack. SP008, SP010, SP029, SP024, SP025
CP039 PostHog's real competitive risk is not a single mirror-image rival but many credible combinations: analytics-first suites, replay specialists, feature-management incumbents, optimization platforms, and internal or warehouse-native builds. SP024, SP025, SP028, SP029, SP030
CP040 Forrester's market framing implies rising commoditization pressure because feature management and experimentation are no longer isolated niches that only one vendor category owns. SP028, SP013
CP041 PostHog's moat is strongest when a buyer values integrated developer workflow, open-source or self-hosted control, and transparent usage pricing at the same time. SP018, SP019, SP021, SP024
CP042 PostHog's moat is weaker when the buyer prioritizes non-technical self-serve analytics, incumbent enterprise governance, or specialized replay depth over integrated breadth. SP024, SP026, SP008, SP010
CP043 Heap's attachment to Contentsquare and Split's migration into Harness show that adjacent categories are also consolidating into broader digital-experience and DevOps platforms. SP006, SP017
CP044 Statsig's homepage highlights an Amplitude partnership, showing that analytics and experimentation budgets can overlap through coopetition as well as direct rivalry. SP013
CP045 GrowthBook explicitly pitches running more experiments at lower cost and with unlimited traffic, reinforcing price-led pressure on closed incumbents in feature management and experimentation. SP016, SP015
CI001 PostHog says more than 90% of companies use the product for free. SI001, SI034
CI002 PostHog says no credit card is required to get started. SI001, SI034
CI003 PostHog publishes separate free-tier and overage meters across analytics, replay, feature flags, surveys, warehouse, pipelines, AI observability, AI, workflows, and logs. SI001, SI002
CI004 Published analytics overages start after the first 1 million free events. SI001, SI034
CI005 Published analytics overage rates step down from $0.0000500 per event to $0.0000090 per event at higher scale. SI001
CI006 Published session replay overage rates step down from $0.0050 to $0.0015 per recording after the free allowance. SI001
CI007 Published feature-flag overage rates step down from $0.000100 to $0.000010 per request after the free allowance. SI001
CI008 PostHog's Product Analytics docs say the product is billed by captured event volume rather than seats. SI034
CI009 PostHog's estimating-costs docs say default local evaluation can be billed as 10 feature-flag requests every 30 seconds per running instance. SI032
CI010 PostHog's billing-limit docs say additional ingestion stops once a product crosses the user-set cap. SI030, SI031
CI011 PostHog's billing-limit docs say data above the cap is lost rather than stored for later billing. SI030, SI031
CI012 PostHog's billing FAQ says early-stage startups can get up to $50,000 in credits. SI031
CI013 PostHog's billing FAQ says nonprofits can discuss discount options with sales after signing up. SI031
CI014 PostHog's billing handbook says the company supports coupons, custom price tiers, flat-first-tier plans, and flat up-front no-metering plans. SI033
CI015 PostHog's billing handbook shows discounts change effective volume purchased against a billing limit. SI033
CI016 PostHog's about page says the company tries to match the cheapest major competitor for each product. SI005
CI017 PostHog's about page says the company covers costs with razor-thin margins and makes up for it with scale. SI005
CI018 PostHog's careers page says revenue is over $50 million a year. SI006
CI019 PostHog's careers page says the company is default alive. SI006, SI022
CI020 PostHog's handbook future page says management wants to hit $100 million of annual revenue by the end of 2026. SI008
CI021 PostHog's handbook future page says the company needs about 7% monthly revenue growth to reach that 2026 target. SI008
CI022 Y Combinator's company profile says PostHog has been averaging about 10% monthly revenue growth. SI022
CI023 PostHog's about page says the platform is used by 190254+ teams. SI005
CI024 PostHog's pricing page says the company has over 60,000 customers. SI001
CI025 PostHog's Series D post says over 176k companies had signed up by June 2025. SI007
CI026 PostHog's careers page says the company has 200+ people. SI006
CI027 PostHog's careers page says its small teams are looking to add 25 team members. SI006
CI028 Y Combinator said PostHog let it collect 30% more data than Google Analytics. SI009
CI029 Y Combinator said the direct Slack support channel gave it direct access to PostHog engineers. SI009
CI030 Y Combinator said a six-week experiment produced 40% more messages than the control group. SI009
CI031 Y Combinator said the same experiment produced 35% more accepted requests than the control group. SI009
CI032 Hasura said PostHog-driven onboarding changes improved conversion by 10-20%. SI010
CI033 Hasura said usage expanded from engineering into UX and marketing teams after the initial onboarding analysis use case. SI010
CI034 PostHog's about and pricing pages both frame onboarding as transparent and usable without talking to sales. SI005, SI001
CI035 PostHog's pricing FAQ says the best way to estimate cost is often to sign up for free and observe projected billing after a few days. SI031, SI032
CI036 PostHog's estimating-costs docs publish example monthly event-per-MAU heuristics, including 87 events per MAU for a B2B PostHog example. SI032
CI037 PostHog's Series D post says the company raised $70 million of primary capital at a $920 million valuation in June 2025. SI007, SI019, SI020
CI038 PostHog's Series D post says only around $10 million of that financing was primary because employee liquidity was a major goal. SI007
CI039 PostHog's Series D post says employees could sell up to 20% of vested shares in the expanded liquidity program. SI007
CI040 PostHog's careers page says the company held its first employee secondary in 2024. SI006
CI041 PostHog's careers page says the company executed its first tender offer in 2025. SI006
CI042 Economic Times and Entrepreneur both reported a $75 million round at a $1.4 billion valuation in September 2025. SI019, SI020, SI021
CI043 Sacra estimates PostHog reached about $57.5 million of ARR in February 2026. SI017
CI044 PostHog's billing handbook says the Billing Service is the source of truth for product plans and entitlements. SI033
CI045 PostHog's billing handbook says Stripe is the source of truth for customer records, invoices, and payments. SI033
CI046 PostHog's billing handbook says credit-based contracts typically carry a 30-day invoice due date. SI033
CI047 PostHog's billing handbook says upfront contract payments are made via bank transfer rather than checks. SI033
CI048 PostHog's billing handbook says pay-as-you-go accounts receive four automated payment attempts before further attempts stop. SI033
CI049 PostHog's billing handbook says three consecutive missed payment periods can trigger a requirement for three months of advance payment. SI033
CI050 PostHog's billing handbook says non-payment can lead to access suspension or reversion to the free tier and its usage limits. SI033
CI051 Atlassian's FY2025 annual report reports $5.215 billion of revenue. SI040
CI052 Atlassian's FY2025 annual report reports an 83% gross margin. SI040
CI053 Atlassian's FY2025 annual report reports research and development spending equal to 51% of revenue. SI040
CI054 Atlassian's FY2025 annual report reports marketing and sales spending equal to 22% of revenue. SI040
CI055 Atlassian's FY2025 annual report reports $1.4 billion of free cash flow. SI040
CI056 Atlassian says its product-led philosophy emphasizes self-service entry while sales focuses on expanding larger enterprise relationships. SI040
CI057 Datadog's 2025 Form 10-K reports $3.427 billion of revenue. SI044
CI058 Datadog's 2025 Form 10-K reports $2.740 billion of gross profit. SI044
CI059 Datadog's 2025 Form 10-K implies roughly 80% gross margin. SI044
CI060 Datadog's 2025 Form 10-K implies sales and marketing intensity of about 28% of revenue. SI044
CI061 Datadog's 2025 Form 10-K implies research and development intensity of about 45% of revenue. SI044
CI062 Datadog's 2025 Form 10-K reports $401.3 million of cash and cash equivalents plus $4.073 billion of marketable securities. SI044
CI063 Datadog's 2025 Form 10-K reports $914.7 million of free cash flow. SI044
CI064 Datadog's 2025 Form 10-K says substantially all revenue comes from subscription software sales. SI044
CI065 Datadog's 2025 Form 10-K reports $3.461 billion of remaining performance obligations at year-end 2025. SI044
CI066 Atlassian annual-reports pages and Datadog investor-relations pages show that public peers disclose audited annual reports and continuing quarterly updates. SI035, SI042
CI067 The reviewed public PostHog pack does not disclose current cash, monthly burn, or runway. SI001, SI006, SI007, SI008
CI068 The reviewed public PostHog pack does not disclose gross margin, net revenue retention, or CAC payback. SI001, SI005, SI006, SI008
CI069 Open-source and self-host positioning is public, but the reviewed pack does not disclose the cloud-versus-self-hosted revenue mix. SI002, SI004, SI005
CI070 Public materials show both self-serve metering and negotiated contract constructs, so realized price per customer cannot be inferred from list pricing alone. SI001, SI031, SI033
CI071 PostHog's terms and DPA provide public contracting surfaces, but they do not add audited financial statements or balance-sheet detail. SI012, SI013
CE001 Product OS publicly bundles product analytics, web analytics, session replay, feature flags, A/B testing, surveys, and a data warehouse in one platform. SE001, SE021
CE002 PostHog’s public repo and YC profile extend that scope to error tracking, CDP/data pipelines, LLM observability, and an AI product assistant. SE021, SE024, SE027
CE003 Product OS says easy client and server SDKs, including posthog-js, autocapture frontend events so teams do not have to manually instrument every simple interaction. SE001, SE030
CE004 The public repo describes product analytics as event-based analytics that can be analyzed with visualizations or SQL. SE021, SE024
CE005 Feature flag docs say flags work for users, groups, or percentages of traffic and underpin safe rollouts, A/B testing, and remote configuration. SE003
CE006 The current feature-flag docs list phased rollouts, kill switches, targeting, remote config, and beta programs as standard use cases. SE003
CE007 Session replay docs position replay as a tool for diagnosing UI issues, improving support, analyzing user friction, and investigating performance with network monitoring. SE005, SE015
CE008 PostHog’s CDP page says events can update user records or trigger workflows in other products as data moves through the stack. SE007
CE009 The CDP transformation layer supports enrichment, property mapping, validation, PII scrubbing, and event filtering before events are stored. SE007
CE010 Product OS says data pipelines can send data to monitoring, marketing automation, sales, and support tools, while SQL grants unrestricted custom analysis of PostHog data. SE001, SE007
CE011 The pricing page exposes monthly free-tier units across analytics, replay, feature flags, surveys, data warehouse, data pipelines, AI observability, PostHog AI, workflows, and logs, and says more than 90% of companies use PostHog for free. SE008, SE021
CE012 The architecture overview names Django for the web app/API, Rust services for capture, flag evaluation, and replay ingestion, Kafka as the message bus, Celery plus Temporal plus Dagster for workers, a Node.js CDP worker, and ClickHouse/PostgreSQL/Redis/blob storage as core components. SE016
CE013 The cloud infrastructure diagram places application services in AWS EKS and shows self-managed ClickHouse on EC2 alongside Aurora PostgreSQL, Redis or Valkey, Kafka or WarpStream, and S3. SE016
CE014 ClickHouse is documented as PostHog’s main analytics backend and ingests via Kafka rather than direct inserts to improve resilience toward outages. SE016, SE017
CE015 The ClickHouse docs describe a sharded design that uses Kafka-engine tables, materialized views, and distributed tables in the ingest path. SE017
CE016 The public architecture flow shows client apps and SDKs sending events and recordings into capture services before data fans out into storage and export paths. SE016
CE017 The JavaScript web docs support installation by HTML snippet or package manager and point to framework-specific guides for Next.js, React, Vue, Angular, Astro, Remix, and Svelte. SE030
CE018 The JavaScript web library lazy-loads extensions such as surveys or the replay recorder by default and also offers slim or no-external bundle options for CSP, Electron, and other constrained environments. SE030
CE019 Current JS extension bundles explicitly cover Feature Flags, Session Replay, Analytics, Error Tracking, Surveys, Experiments, Site apps, Tracing, Toolbar, Logs, Conversations, and an all-in bundle. SE030
CE020 JS docs say PostHog can follow users across a marketing site and product app with a cross-domain cookie and supports replay triggers plus full capture opt-outs. SE030, SE010
CE021 PostHog supports multiple named JS instances at the same time but warns teams to configure autocapture carefully to avoid sending the same events twice. SE030
CE022 Server-side local evaluation is currently available in the Node, Ruby, Go, Python, C#/.NET, PHP, Java, and Rust SDKs. SE004, SE031
CE023 Local evaluation replaces per-check /flags calls with background /flags/definitions fetches, which the docs position as faster and more cost-effective for high-traffic services. SE004
CE024 Local evaluation requires the caller to supply all properties used in release conditions and to keep the feature-flags secure API key secret, while edge or stateless environments should use shared external cache or remote evaluation. SE004
CE025 Self-host docs say self-hosted PostHog is the same product as Cloud but operators manage deployment, scaling, URLs, and risk themselves, with a free MIT Docker Compose hobby deploy offered as the standard path. SE019, SE024
CE026 Self-hosted PostHog does not use tagged releases and comes without support guarantees for behavior on customer infrastructure. SE019, SE021
CE027 The public README recommends PostHog Cloud as the fastest and most reliable path and says open-source deployments should scale to roughly 100k events per month before teams migrate. SE021, SE024
CE028 New deployments of PostHog’s paid open-source product using Kubernetes are no longer supported. SE019
CE029 PostHog’s release handbook says new products and features move through Setting up, Alpha, Beta, and GA, with initial planning and alpha development happening behind a feature flag. SE009, SE003
CE030 The same release guidance says feature previews and coming-soon items are exposed at the user level so individuals can opt in or register interest before wider launch. SE009
CE031 MCP docs say PostHog offers a free MCP server at mcp.posthog.com/mcp and automatically routes authenticated users to the correct US or EU data region. SE018, SE014
CE032 The PostHog wizard can install the MCP server directly into PostHog Code, Cursor, Claude Code, Claude Desktop, Codex, VS Code, and Zed. SE018
CE033 MCP documentation supports OAuth or project-scoped personal API keys and explicitly warns users to review tool calls because LLM workflows are exposed to prompt injection risk. SE018
CE034 PostHog’s GA4 comparison uses the MCP server and SQL query builder as proof that the product is built for developers rather than primarily for marketers. SE014, SE018
CE035 The npm page shows posthog-js version 1.376.0 was published two days before the run date with about 6.97 million weekly downloads, 333 dependents, and 1,157 versions. SE023, SE025
CE036 PyPI shows posthog 7.15.3 uploaded on 2026-05-21 with trusted publishing provenance and Python 3.10 through 3.14 support. SE031
CE037 The public posthog.com repository says the website, handbook, roadmap, API docs, and related product surfaces are maintained in a public repository treated like a product. SE022
CE038 The privacy policy covers both hosted services and self-managed installations and says PostHog automatically collects usage information from self-managed instances while offering an opt-out path. SE010, SE019
CE039 The privacy policy says hosted customer data is stored and processed in the United States or in Germany for EU-hosted cloud customers and names AWS and Google Cloud Platform as cloud infrastructure providers. SE010, SE011
CE040 The public DPA page says a legally binding countersigned DPA must be generated inside the PostHog app rather than relying on the public preview page alone. SE011
CE041 PostHog’s comparison pages say the product offers EU-hosted cloud and positions itself as SOC 2 certified, GDPR-ready, and HIPAA-ready with BAAs available on platform packages. SE014, SE015
CE042 The security advisory page says there are no active advisories today, but PSA-2025-00001 was a resolved medium issue where an overly permissive SQL-editor table exposed query text from unrelated teams. SE012
CE043 That 2025 advisory was mitigated by removing access, adding a team_id field to the query-log table, auditing other tables, and planning automated tests to ensure new tables include team_id where appropriate. SE012
CE044 The public post-mortem index preserves multiple 2025-2026 reliability caveats, including workflow wait-until-condition failures, logs data loss, feature-flags cache degradation, a replay SDK fetch-wrapper incident, and feature-flag outages. SE013
CE045 Self-host docs explicitly ask operators whether they are willing to accept the risk of potential data loss because self-hosted instances carry no guarantees. SE019, SE013
CE046 Forrester says feature management and experimentation are increasingly serving different personas and use cases even when vendors continue to package them together. SE029
CE047 Mixpanel argues that AI has become the front door to data and that analytics increasingly has to help teams decide what to do next, not just explain what happened. SE028
CE048 Y Combinator describes PostHog as one platform for product analytics, session replays, feature flags, experimentation, LLM observability, and a SQL data warehouse with one-click imports. SE027
CE049 Sacra says self-hosting can help address privacy concerns but also adds user complexity and can complicate monetization of the cloud offering. SE026
CE050 PostHog’s GA4 comparison says reverse-proxying events through a company’s own domain can reduce interception by tracking blockers. SE014
CE051 The public roadmap URL was live at the run date, but the retained fetch exposed only a loading shell rather than readable upcoming items, so specific near-term roadmap detail remains inaccessible in the audited public pack. SE032
CU001 PostHog's about page says the product is used by 190254+ teams. SU012
CU002 The same about page also says over 190254+ customers and just under a quarter of a million engineers use PostHog. SU012
CU003 PostHog's pricing page separately says the company has over 60000 customers. SU011
CU004 PostHog's pricing page says more than 90% of companies use PostHog for free. SU011
CU005 The free plan requires no credit card, allows one project, and allows unlimited team members. SU011
CU006 PostHog says there are no minimums or annual commitments, although it can offer annual commitments as part of an enterprise agreement. SU011
CU007 PostHog publicly positions itself as a 10+ product suite with separate usage meters that can expand with customer adoption. SU011, SU013
CU008 The customer landing page shows named accounts using a wide mix of analytics, replay, flags, experiments, surveys, CDP, warehouse, AI, and error-tracking surfaces. SU001
CU009 PostHog says 65% of every Y Combinator batch uses its products. SU012
CU010 Y Combinator describes PostHog as a current production tool for Startup School, the YC Startup Library, and Co-Founder Matching. SU002
CU011 Y Combinator says Google Analytics dropped 30% of its user data due to adblockers or third-party-cookie issues. SU002
CU012 Y Combinator says a PostHog experiment increased messages sent by 40% for a six-week stale-profile treatment group. SU002
CU013 Y Combinator says the same experiment produced 35% more accepted requests and therefore 35% more matches than the control group. SU002
CU014 Hasura says it started using PostHog in 2021 because Google Analytics was too broad for deeper product and UX analysis. SU003
CU015 Hasura says funnel changes informed by PostHog improved onboarding conversion by 10-20%. SU003
CU016 Hasura says PostHog is now used across its blogs, website, and broader product flows rather than for one isolated use case. SU003
CU017 Supabase says its pre-PostHog data stack was fragmented across Plausible, BigQuery, and internal tools. SU004
CU018 Supabase says it deployed PostHog server-side through the Node SDK. SU004
CU019 Supabase says every team can use the same product data through built-in analyses, SQL, and PostHog AI. SU004
CU020 Supabase says PostHog helped it detect AI-builder acquisition signals early and turn those signals into partnerships. SU004
CU021 Supabase says PostHog helped it 10X weekly new user acquisition. SU004
CU022 Phantom says it trialed the open-source version first and then rolled PostHog out fully after validation. SU005
CU023 Phantom says it uses PostHog daily to monitor DAU, swap volumes, stake volumes, and internal dashboards. SU005
CU024 Phantom says PostHog helped trigger infrastructure changes that cut failure rates by 90% and that feature flags now keep failure rates at 1% or below. SU005
CU025 Phantom says it grew from a private beta with zero users to more than a million users and was adding almost 100000 new users every week. SU005
CU026 ElevenLabs says it uses product analytics, feature flags, session replay, surveys, and G2-review prompts as one launch workflow. SU006
CU027 ElevenLabs says it tracks conversion, retention, repeat visits, and weekly retention across personas. SU006
CU028 ElevenLabs says it rolled an annual pricing experiment out to 100% of users. SU006
CU029 Lovable says it started using PostHog very early and now relies on feature flags, experiments, and AI observability to debug its agent loop. SU007
CU030 Lovable also says it runs two other LLM observability and analytics tools alongside PostHog. SU007
CU031 Lovable says PostHog shipped an LLM playground requested by the team in less than a month. SU007
CU032 Arena says it has 5M+ monthly users generating millions of comparisons each month. SU008
CU033 Arena says everything it ships is experimented on and that PostHog is the source of truth for company performance. SU008
CU034 Arena says users spend an average of 19 minutes on leaderboard pages. SU008
CU035 Arena says its event volume increased 19× over the prior six months. SU008
CU036 Arena says retention and returning behavior are north-star metrics for its repeat-visit platform. SU008
CU037 Exa says it centralized product analytics into PostHog after operating a scattered analytics stack and still has modules such as surveys and revenue analytics left to adopt. SU009
CU038 ResearchGate says it uses PostHog to test product changes for over 25M users and across hundreds of millions of sessions. SU010
CU039 ResearchGate says its scale puts it into custom enterprise packages and that PostHog provided more responsive expert support than many larger vendors. SU010
CU040 ResearchGate says it has been testing feed algorithms for a year and that PostHog enables rapid autonomous iteration by data scientists. SU010
CU041 G2's archived review page shows PostHog at 4.5/5 across 950 reviews. SU014
CU042 A visible G2 review praises PostHog's customizable dashboards, cohorts, heat maps, and helpful support. SU014
CU043 A visible G2 review criticizes PostHog for frequent crashes, many incident emails, and confusing documentation for custom event data. SU014
CU044 Sacra says PostHog's usage-based pricing enables broad adoption within engineering teams and expansion as customers increase event volume and adopt additional products. SU015
CU045 Mixpanel says activation, stickiness, and retention are more durable growth signals than raw acquisition volume in 2026 digital analytics. SU016
CU046 SlashData estimates the global developer population at 48.4 million as of Q3 2025. SU017
CU047 Y Combinator describes itself as a funder of early-stage startups that invests $500k in select groups four times a year. SU018
CU048 Hasura describes itself as GraphQL and data-delivery infrastructure loved by developers. SU019
CU049 Supabase describes itself as the Postgres development platform. SU020
CU050 Phantom describes itself as a money app and crypto product trusted by 20+ million users. SU021
CU051 ElevenLabs describes itself as an AI voice generator and voice-agents platform. SU022
CU052 Lovable describes itself as an AI app builder for coding apps and websites quickly. SU023
CU053 Exa describes itself as a web-search API built for AI, and ResearchGate describes itself as a network to find and share research for 25 million scientists. SU024, SU025
CU054 The accessible named-customer proof set is concentrated in startup, AI, developer-tooling, crypto, and scientific-network products rather than in traditional non-technical verticals. SU002, SU003, SU004, SU005, SU006, SU007, SU008, SU009, SU010, SU018, SU019, SU020, SU021, SU022, SU023, SU024, SU025
CU055 Most accessible named stories read as production deployments because they describe current daily use, year-long testing, or core operational workflows rather than time-boxed pilots. SU002, SU003, SU004, SU005, SU006, SU007, SU008, SU009, SU010
CU056 PostHog's public customer counts are not directly comparable because the reviewed official pages separately use teams, customers, engineers, and free companies as units of account. SU011, SU012
CU057 Public expansion evidence is stronger on multi-product and usage growth than on disclosed account-count growth or retention math. SU007, SU008, SU009, SU010, SU011, SU013, SU015
CU058 Public proof narrows sharply from top-of-funnel claims to named evidence: 190254+ teams on the about page, 60000+ customers on pricing, 950 G2 reviews, and nine accessible case-study pages in this run. SU002, SU003, SU004, SU005, SU006, SU007, SU008, SU009, SU010, SU011, SU012, SU014
CU059 Entry procurement friction appears low because PostHog offers a no-card free plan with unlimited team members and no minimum commitment, but enterprise-grade support appears later at custom-package scale. SU010, SU011
CU060 Public customer economics remain opaque because no reviewed source discloses NRR, GRR, logo churn, top-customer share, or a paid-versus-free customer breakout. SU011, SU012, SU014, SU015
CR001 PostHog's privacy policy says PostHog is the data controller for the processing operations described in that policy. SR001
CR002 PostHog's security page says that for PostHog Cloud the customer is the data controller and PostHog is the data processor. SR012, SR014
CR003 PostHog's security page says self-hosting customers are both processor and controller because they operate their own instance. SR012
CR004 PostHog's DPA says PostHog participates in the EU-U.S. DPF, the UK Extension, and the Swiss-U.S. DPF. SR003, SR016
CR005 The Data Privacy Framework participant page lists PostHog as active under the EU-U.S. DPF, UK Extension, and Swiss-U.S. DPF, with original certification date 2024-05-08 and next certification due 2027-03-10. SR016, SR001
CR006 PostHog's privacy policy says international transfers may also rely on SCCs and unresolved DPF complaints can go to JAMS. SR001, SR003
CR007 PostHog's security guidance says customers can choose EU or US AWS hosting and that UK/EU/EEA to US transfers rely on SCCs when required. SR012, SR001
CR008 PostHog's privacy docs say customers are responsible for the data they collect and for deciding whether their PostHog use complies with regulations. SR014, SR001
CR009 PostHog's privacy docs say project tokens that start with phc_ can be public but personal API keys that start with phx_ should not be public. SR014
CR010 PostHog says it can provide a Business Associate Agreement for HIPAA-compliant cloud use. SR014, SR012
CR011 PostHog's security page says the company is SOC 2 Type II compliant and requires MFA, with YubiKeys for certain infrastructure accounts. SR012
CR012 PostHog's security advisories page says there are currently no active security advisories or CVEs. SR005
CR013 PostHog's security advisories page discloses PSA-2025-00001, a medium-severity issue where SQL editor users could see query text from unrelated teams. SR005
CR014 The same advisory says no usage was observed in EU cloud, US query-log history only went back to 2025-07-03, and the earlier US window could not be fully confirmed. SR005
CR015 PostHog says it fixed PSA-2025-00001 by removing the table, adding team_id, auditing similar tables, and planning automated tests. SR005
CR016 PostHog's terms cap aggregate liability at the greater of $1,000 or one year of fees paid. SR002
CR017 PostHog's terms require customers to ensure customer data is collected, processed, transferred, and used in compliance with applicable data protection laws. SR002, SR001
CR018 PostHog says it only publishes a public post-mortem when an incident causes permanent data impact, customer disruption, or extended unavailability. SR004
CR019 PostHog's public post-mortem list names workflow failures, logs data loss, feature flags cache degradation, replay SDK failure, Shai-Hulud, persons migration, feature flags recurring outages, surveys SDK bug, and a feature flags service outage. SR004
CR020 An official status incident says the US Logs product suffered confirmed customer data loss up to 2026-02-16 21:00 UTC. SR015, SR004
CR021 The same incident says only the last three days of logs backups were available for backfill and that events and replays lived in a more mature separate cluster. SR015
CR022 PostHog's Shai-Hulud post-mortem says malicious npm versions were live from 04:11 UTC until PostHog identified and removed them by 09:30 UTC on 2025-11-24. SR013, SR024
CR023 PostHog says the Shai-Hulud attack path affected npm-distributed JavaScript SDKs rather than the browser script tag. SR013, SR024
CR024 PostHog says the attacker exploited a pull_request_target workflow that checked out PR-controlled code, stole a bot PAT, and then stole other GitHub secrets including the npm token. SR013, SR024, SR025
CR025 PostHog says the malicious packages used a preinstall script to scan for credentials and exfiltrate them before propagating via npm tokens. SR013, SR024, SR025
CR026 PostHog says it responded to Shai-Hulud by moving package release workflows to Trusted Publisher, tightening workflow review, switching to pnpm 10, and reworking GitHub secret management. SR013, SR024
CR027 iLert says the Shai-Hulud window from malicious publish to initial containment was roughly 5 hours 19 minutes. SR024
CR028 Daily Security Review describes the Shai-Hulud incident as caused by a CI/CD automation flaw and frames it as PostHog's most severe security incident. SR025, SR013
CR029 OpenCVE lists CVE-2025-1520 as a high-severity (CVSS 8.0) PostHog ClickHouse Table Functions SQL injection / remote code execution issue updated 2025-08-07. SR017
CR030 OpenCVE says exploitation of CVE-2025-1520 requires authentication but can allow arbitrary code execution on affected installations. SR017
CR031 StatusGator says PostHog was operational on 2026-05-24 and that the last officially acknowledged outage was on 2026-05-21. SR018
CR032 EagleStatus lists recent May 2026 updates affecting App, Error Tracking Ingestion Lag, REST API query endpoints, and multiple components. SR023
CR033 StatusGator pages show separate App, Workflows, and Logs status histories since May 2023, evidencing a multi-component operational surface. SR018, SR019, SR020, SR021
CR034 StatusSight said there were no active incidents at fetch time, which lowers immediate outage concern but does not erase recent incident history. SR022
CR035 PostHog's security advisories say the company recommends Cloud because OSS/self-host is not suitable for use at scale and older K8s deployments were sunset. SR005, SR008
CR036 PostHog's pricing page says more than 90% of companies use PostHog for free. SR007
CR037 PostHog prices usage across analytics, replay, feature flags, experiments, warehouse, pipelines, AI observability, workflows, and logs, while letting customers set billing limits. SR007
CR038 PostHog's product, Product OS, and CDP pages position the product as a broad platform that can ingest, transform, and send data across many surfaces and hundreds of tools. SR006, SR008, SR009
CR039 PostHog's careers page says the company has 200+ people. SR010
CR040 PostHog's careers page advertises open roles and says the company takes exceptional people when they come along. SR010
CR041 PostHog's future page says the mid-term goal is $100M ARR by 2026. SR011
CR042 Sacra estimates PostHog reached about $57.5M ARR in February 2026 and says growth remained high but was decelerating. SR027
CR043 The gap between a public $100M ARR goal and a February 2026 outside ARR estimate around $57.5M implies material execution pressure over the rest of 2026. SR011, SR027
CR044 G2 review text praises flexibility and support but says frequent crashes during loading and many incident emails were irritating. SR026
CR045 G2 review text also says some behavior is ambiguous or poorly documented, citing replay and webview capture issues. SR026
CR046 4 Day Week describes PostHog as remote-first, async, and meeting-light. SR028
CR047 Forrester says feature management is increasingly developer-centric while experimentation is increasingly product or marketing-centric, creating role-split pressure on bundled vendors. SR030
CR048 Tracxn says PostHog has 2,641 active competitors and a current valuation around $1.4B. SR029
CR049 PostHog's DPA says AI features may use subprocessors depending on the enabled services. SR003
CR050 PostHog's security page says a sub-processor list is maintained as part of the DPA and kept to a strict minimum. SR012, SR003
CR051 PostHog's documented mitigations are real—SOC 2, MFA, EU hosting choice, BAAs, DPF/SCC mechanisms, incident transparency, and post-incident release hardening. SR012, SR014, SR001, SR003, SR013
CR052 The highest residual legal and privacy risk is a misfit between PostHog's controls and how customers actually instrument, transfer, or expose data. SR014, SR001, SR003
CR053 GitHub Actions plus npm publishing is now a proven critical dependency rather than a hypothetical supply-chain risk. SR013, SR024, SR025
CR054 The logs incident suggests newer product surfaces can carry materially weaker backup and recovery characteristics than mature core analytics clusters. SR015, SR004
CR055 PLG monetization risk is material because the company says 90%+ of companies are free while product breadth, hiring, and reliability expectations keep expanding. SR007, SR010, SR011, SR027
CR056 Suite breadth creates partner and integration risk because PostHog positions CDP and Product OS around moving data between hundreds of tools and surfaces. SR009, SR008, SR006
CR057 The most important thesis-break events would be another confirmed cross-tenant exposure, a core-data loss incident outside the newer logs product, or a repeat supply-chain compromise affecting customer environments. SR005, SR015, SR013
CR058 The most monitorable public KPIs are acknowledged outage cadence, new post-mortems or advisories, DPF certification continuity, review complaints about crashes, and progress toward the 2026 revenue goal. SR004, SR005, SR016, SR026, SR011, SR027
CR059 Cloud-first guidance and self-host-at-scale warnings create enterprise-fit risk for buyers that want maximum data control without taking on the operational burden themselves. SR005, SR012, SR014
CR060 PostHog's unusual transparency is a trust asset, but it also proves residual exposure across privacy, security, and reliability rather than hiding it. SR004, SR005, SR013, SR015
CV001 Independent 2025 coverage says PostHog raised $75 million at a $1.4 billion valuation in a round led by Peak XV. SV011, SV012
CV002 PostHog's official Series D post says it raised $70 million in primary capital at a $920 million valuation led by Stripe. SV003
CV003 The move from a $920 million valuation to a reported $1.4 billion valuation is roughly a 52% step-up within 2025. SV003, SV011, SV012
CV004 PostHog's careers page says company revenue is over $50 million a year. SV008
CV005 PostHog's handbook says the company wants to hit $100 million in annual revenue by the end of 2026. SV004
CV006 Sacra estimates that PostHog reached about $57.5 million of ARR in February 2026, up roughly 99% year over year. SV009
CV007 PostHog's pricing page says more than 90% of companies use the product for free. SV001
CV008 PostHog monetizes with usage-based pricing and publishes product-specific meters and volume discounts instead of seat-based list pricing. SV001
CV009 Y Combinator's company profile says PostHog has been averaging about 10% monthly revenue growth and is default alive. SV014
CV010 PostHog's about page says the platform is used by more than 190,254 teams. SV005
CV011 PostHog's about page says the company already sells 10 or more paid products. SV005
CV012 PostHog's products page positions analytics, replay, feature flags, experiments, surveys, warehouse, pipelines, AI observability, AI, workflows, and logs inside one Product OS. SV002
CV013 Peak XV's portfolio page says it partnered with PostHog in 2025. SV013
CV014 Contrary frames PostHog's commercial thesis around consolidating several product tools into one developer workflow and data stack. SV010
CV015 Grand View Research estimates the global product analytics market was $14.81 billion in 2023 and is growing at a 19.8% CAGR through 2030. SV015
CV016 Expert Market Research estimates the product analytics market reached $12.03 billion in 2025 and can grow at a 15.10% CAGR through 2035. SV016
CV017 Mordor Intelligence estimates the product analytics market will rise from $13.04 billion in 2026 to $25.73 billion by 2031 at a 14.55% CAGR. SV017
CV018 Mixpanel says digital analytics in 2026 has become AI-first and that the key product benchmarks are acquisition, engagement, stickiness, and retention. SV018
CV019 Forrester says feature management and experimentation span both software delivery and product management, which supports the strategic logic of an integrated suite. SV019
CV020 SlashData says there are 48.4 million developers around the world. SV020
CV021 The combination of transparent pricing and customer case studies supports a product-led expansion motion rather than a pure top-down seat-sales model. SV001, SV006, SV007
CV022 Hasura says PostHog-backed onboarding changes improved conversion by 10% to 20%. SV007
CV023 Y Combinator says PostHog experiments generated 40% more messages and 35% more accepted requests. SV006
CV024 Using only the official >$50 million revenue floor, the reported $1.4 billion round implies a current revenue multiple no better than roughly 28x. SV008, SV011, SV012
CV025 Using Sacra's $57.5 million ARR estimate, the same $1.4 billion round implies roughly a 24x ARR multiple. SV009, SV011, SV012
CV026 PostHog's current floor-based or estimate-based multiple is above Datadog's public 18.3x ARR multiple. SV009, SV011, SV012, SV030
CV027 SaaSValuation.io shows Datadog at about 18.3x ARR in Q1 2026 with 79.2% gross margin and a small positive operating margin. SV030
CV028 Datadog's 2025 Form 10-K and 2026 Q1 Form 10-Q show FY2025 revenue of $3.4272 billion and Q1 2026 revenue of $1.0064 billion. SV026, SV027
CV029 Datadog's Q1 2026 filing shows $797.2 million of gross profit on $1.0064 billion of revenue, implying roughly 79% gross margin. SV027, SV030
CV030 Atlassian's 2025 Form 10-K and 2026 Q3 Form 10-Q show FY2025 revenue of $5.2153 billion and nine-month FY2026 revenue of $4.8058 billion. SV024, SV025
CV031 Atlassian filings show gross margin staying in roughly the 83% to 85% range across FY2025 and FY2026 year to date. SV024, SV025
CV032 Combining Atlassian's public market cap with current annualized revenue implies only a low-single-digit revenue multiple, far below PostHog's private round on disclosed denominators. SV025, SV030
CV033 Amplitude's 2025 Form 10-K and 2026 Q1 Form 10-Q show FY2025 revenue of $343.2 million and Q1 2026 revenue of $93.5 million. SV028, SV029
CV034 Amplitude filings show gross margin in roughly the 73% to 74% range while the company remains lossmaking. SV028, SV029
CV035 Amplitude's market cap and current quarterly revenue imply about a 2.3x revenue multiple. SV029, SV030
CV036 Multiples.vc says public software valuations in May 2026 are highly segmented by infrastructure, vertical, and horizontal categories. SV031
CV037 Multiples.vc's May 2026 view shows all-SaaS averages around 10.4x even though the sector spread is wide. SV031
CV038 Livmo says public SaaS entered 2026 around 6x to 7x EV to revenue while private SaaS traded about 3x to 7x ARR with a 4.5x median. SV032
CV039 Livmo says 7x to 9x private ARR is usually reserved for companies above 50 on Rule of 40 and above 120% NRR, while 10x to 12x is for exceptional outliers. SV032
CV040 Reviewed public sources still do not disclose PostHog's NRR, gross margin, burn, exact cap table, or liquidation preferences. SV001, SV003, SV004, SV008, SV011, SV012
CV041 The Series D post and careers page both signal employee liquidity or tender activity, so investors cannot assume every disclosed financing dollar extended runway as pure primary capital. SV003, SV008
CV042 PostHog's post-mortem archive and Shai-Hulud incident show trust and reliability risks that can widen discount rates and delay IPO readiness. SV021, SV022
CV043 The public comp quality gap matters because Datadog, Atlassian, and Amplitude publish audited quarterly filings while PostHog offers only a revenue floor, a target, and one external ARR estimate. SV008, SV009, SV023, SV024, SV026, SV028
CV044 A conservative bear case that reaches only $75 million to $90 million of 2027 revenue and clears at 5x to 7x implies roughly $0.4 billion to $0.6 billion of value. SV004, SV008, SV032
CV045 A base case that reaches $100 million to $120 million of 2027 revenue and clears at 7x to 10x implies roughly $0.7 billion to $1.2 billion of value. SV004, SV008, SV031, SV032
CV046 A bull case that reaches $130 million to $160 million of 2027 revenue and clears at 10x to 14x implies roughly $1.3 billion to $2.2 billion of value. SV004, SV009, SV031, SV032
CV047 The latest $1.4 billion valuation already sits near the low end of the bull case rather than the base case under current public-market discipline. SV011, SV012, SV031, SV032
CV048 The price-sensitive recommendation is RESEARCH-MORE and investors should not chase fresh primary capital at the $1.4 billion headline until denominator and term data are disclosed. SV001, SV008, SV011, SV012, SV032
CV049 Confidence should be medium and risk should be high because the product and market case are real but the underwriting variables that sustain premium multiples remain private. SV009, SV021, SV022, SV032
CV050 Exit readiness is emerging but not public-ready because a credible IPO narrative still requires sustained $100 million-plus revenue, retention disclosure, margin disclosure, and cleaner reliability signaling. SV004, SV008, SV021, SV022, SV023
CV051 Thesis-break triggers include another material trust event, missing the 2026 revenue target by a wide margin, or any next round that reprices materially below the current headline. SV004, SV021, SV022, SV032
CV052 The first diligence gate is the exact cap table, security type, liquidation preference stack, and primary-versus-secondary split for both 2025 rounds. SV003, SV008, SV011, SV012
CV053 The second diligence gate is cohort-level NRR, gross churn, free-to-paid conversion, and product attach rates across the multi-product suite. SV001, SV006, SV007, SV009, SV014
CV054 The third diligence gate is product-level cloud gross margin, self-host versus cloud mix, and infrastructure cost by workload. SV001, SV002, SV026, SV028
CV055 Transparent pricing, broad product breadth, and heavy free-tier usage make PostHog more comparable to a PLG dev-tools platform than to a seat-licensed enterprise application vendor. SV001, SV002, SV005, SV014
CV056 Official disclosures bracket PostHog between more than $50 million of current annual revenue and a $100 million target by the end of 2026. SV004, SV008
CV057 That official revenue bracket means investors still need material growth before current public-market multiples comfortably support the latest private valuation. SV004, SV008, SV032
CV058 Because more than 90% of companies remain free, premium valuation support depends on conversion and expansion quality rather than raw logo count alone. SV001, SV005
来源
编号出版方标题引文
SO001 PostHog About PostHog
SO002 PostHog Team
SO003 PostHog Products
SO004 PostHog Pricing
SO005 PostHog Customers
SO006 PostHog How we got here
SO007 PostHog Future
SO008 PostHog Why does PostHog exist? Our mission and strategy
SO009 PostHog Careers
SO010 PostHog PostHog raises a series D (and a small C)
SO011 PostHog Privacy policy
SO012 PostHog Data processing agreement preview
SO013 PostHog Terms
SO014 GitHub PostHog GitHub repository
SO015 PostHog Product analytics documentation
SO016 Y Combinator PostHog company profile
SO017 Sacra PostHog revenue, valuation & funding
SO018 Contrary Research PostHog Business Breakdown & Founding Story
SO019 4dayweek.io PostHog remote jobs and careers
SO020 Economic Times SaaS startup PostHog turns unicorn after $75 million fundraise led by Peak XV
SO021 Peak XV PostHog | Peak XV
SO022 1984 Ventures Portfolio
SO023 Entrepreneur India PostHog becomes unicorn with USD 75 Mn funding led by Peak XV
SO024 PostHog Public post-mortems
SO025 PostHog Security advisories
SM001 PostHog Products
SM002 PostHog Pricing
SM003 PostHog Customers
SM004 PostHog How Y Combinator gathers 30% more data with PostHog than Google Analytics
SM005 PostHog Hasura customer story
SM006 PostHog PostHog vs Google Analytics 4
SM007 PostHog Product OS documentation
SM008 PostHog CDP
SM009 Contrary Research PostHog Business Breakdown & Founding Story
SM010 Mixpanel 2026 State of Digital Analytics: Benchmarks, analysis, and recommendations
SM011 Mixpanel What is product experimentation? A complete guide for 2026
SM012 Amplitude What is product analytics? A data-backed guide
SM013 Grand View Research Product Analytics Market Size, Share & Growth Report, 2030
SM014 Expert Market Research Product Analytics Market
SM015 Mordor Intelligence Product Analytics Market Analysis
SM016 StartUs Insights Data Analytics Market Report 2025
SM017 SlashData Developer population sizing
SM018 JetBrains State of Developer Ecosystem 2025
SM019 dbt Labs 2025 State of Analytics Engineering Report
SM020 Forrester Feature management and experimentation — an evolving market
SM021 VWO Solutions
SM022 Statsig Customers
SM023 Heap Product Analytics Buyer's Guide
SM024 Optimizely Reports
SM025 Atlassian SEC filings
SP001 Mixpanel Mixpanel Pricing: Find Your Plan & Get Started | Mixpanel
SP002 Mixpanel Mixpanel: AI Digital Analytics Platform for Product Teams
SP003 Amplitude Amplitude Pricing Options | Fast, Intelligent Customer Behavior Insights with Affordable Pricing Plans
SP004 Amplitude What Is Product Analytics? A Data-Backed Guide
SP005 Heap Pricing
SP006 Heap Heap - Better Insights. Faster. | Heap
SP007 Heap Product Analytics Buyer's Guide
SP008 Fullstory The Behavioral Data Platform
SP009 Fullstory Plans & Packages | Find the Right Fullstory Plan for You
SP010 LaunchDarkly Pricing | LaunchDarkly
SP011 LaunchDarkly Platform Overview | LaunchDarkly
SP012 Statsig Statsig | The modern product development platform
SP013 Statsig Statsig | The modern product development platform
SP014 Statsig Statsig is the best, say our customers
SP015 GrowthBook Predictable Pricing – Free Tiers, Enterprise Plans | GrowthBook
SP016 GrowthBook GrowthBook | Experimentation, Feature Flags &amp; Product Analytics Platform
SP017 Harness Start Feature Management & Experimentation (FME) in Harness | Harness Developer Hub
SP018 PostHog Statsig | The modern product development platform
SP019 PostHog Product OS – PostHog
SP020 PostHog CDP sources & destinations
SP021 PostHog Product OS - Docs - PostHog
SP022 PostHog PostHog vs Google Analytics 4 in-depth tool comparison
SP023 PostHog The best FullStory alternatives & competitors, compared - PostHog
SP024 Cotera Cotera Where PostHog fell short for Elena was the non-technical experience.
SP025 Fungies.io 10 Best Product Analytics Tools for SaaS in 2026: Complete Comparison - Fungies.io
SP026 Startupik Best Product Analytics Tools Compared (Amplitude vs Mixpanel vs PostHog) - Startupik | Startup magazine
SP027 Techno Pulse Best AI Product Analytics Tools in 2026: Amplitude vs Mixpanel vs PostHog vs Heap
SP028 Forrester Feature Management And Experimentation — An Evolving Market Feature management and experimentation is a broad set of capabilities that spans both software delivery and product management.
SP029 VWO VWO | Digital Experience Optimization
SP030 Google Analytics Tools & Solutions for Your Business - Google Analytics
SI001 PostHog PostHog pricing – Transparent, usage-based, generous free tier
SI002 PostHog Product OS – PostHog
SI003 PostHog CDP sources & destinations
SI004 PostHog Product OS - Docs - PostHog
SI005 PostHog About PostHog
SI006 PostHog Careers - PostHog
SI007 PostHog PostHog raises a series D (and a small C) - PostHog
SI008 PostHog Future - Handbook - PostHog
SI009 PostHog How Y Combinator used PostHog experiments to boost engagement by 40% - PostHog
SI010 PostHog How Hasura improved conversion rates by 10-20% with PostHog - PostHog
SI012 PostHog Terms, PostHog style
SI013 PostHog DPA
SI017 Sacra PostHog revenue, valuation & funding Sacra estimates that PostHog hit $57.5M in annual recurring revenue (ARR) in February 2026.
SI019 The Economic Times SaaS startup PostHog turns unicorn after $75 million fundraise led by Peak XV
SI020 Entrepreneur India PostHog becomes unicorn with USD 75 Mn funding led by Peak XV
SI021 Peak XV PostHog
SI022 Y Combinator PostHog: The single platform to analyze, test, observe, and deploy new features | Y Combinator
SI030 PostHog Docs Billing limits and alerts - Docs - PostHog If you exceed the billing limit you set, your additional data is lost forever.
SI031 PostHog Docs Common questions about billing - Docs - PostHog
SI032 PostHog Docs Estimating usage & costs - Docs - PostHog
SI033 PostHog Handbook Billing - Handbook - PostHog After three consecutive missed payment periods, the customer must provide advance payment covering three months of service based on their typical usage before account access is restored.
SI034 PostHog Docs Product Analytics pricing - Docs - PostHog
SI035 Atlassian Investor Relations Atlassian - Financials - Annual reports
SI040 Atlassian Atlassian FY2025 Annual Report on Form 10-K
SI042 Datadog Investor Relations | Datadog
SI044 Securities and Exchange Commission Datadog, Inc. Form 10-K for fiscal year ended December 31, 2025
SE001 PostHog Product OS - Docs - PostHog
SE002 PostHog Product analytics - Documentation - PostHog
SE003 PostHog Feature flags - Docs - PostHog
SE004 PostHog Server-side local evaluation - Docs - PostHog
SE005 PostHog Session replay - Docs - PostHog
SE006 PostHog Data warehouse - Docs - PostHog
SE007 PostHog CDP sources & destinations
SE008 PostHog PostHog pricing – Transparent, usage-based, generous free tier
SE009 PostHog Releasing new products and features - Handbook - PostHog
SE010 PostHog Privacy policy, PostHog style
SE011 PostHog DPA
SE012 PostHog Security advisories - Handbook - PostHog
SE013 PostHog Public post-mortems - Handbook - PostHog
SE014 PostHog PostHog vs Google Analytics 4 in-depth tool comparison
SE015 PostHog The best FullStory alternatives & competitors, compared - PostHog
SE016 PostHog PostHog's architecture - Docs - PostHog
SE017 PostHog ClickHouse - Docs - PostHog
SE018 PostHog Model Context Protocol (MCP) - Docs - PostHog
SE019 PostHog Self-host PostHog - Docs - PostHog
SE020 PostHog Install PostHog - Docs - PostHog
SE021 GitHub GitHub - PostHog/posthog
SE022 GitHub GitHub - PostHog/posthog.com
SE023 npm posthog-js
SE024 GitHub Raw PostHog/posthog README.md
SE025 npm registry posthog-js package metadata
SE026 Sacra PostHog revenue, valuation & funding
SE027 Y Combinator PostHog: The single platform to analyze, test, observe, and deploy new features | Y Combinator
SE028 Mixpanel 2026 State of Digital Analytics: Benchmarks, analysis, and recommendations | Signals & Stories
SE029 Forrester Feature Management And Experimentation — An Evolving Market
SE030 PostHog JavaScript web - Docs - PostHog
SE031 PyPI posthog
SE032 PostHog Roadmap – PostHog
SU001 PostHog customers.mdx – PostHog
SU002 PostHog How Y Combinator used PostHog experiments to boost engagement by 40% - PostHog We recently used it to improve our matching algorithm... users in the 6-week group sent 40% more messages than the control group.
SU003 PostHog How Hasura improved conversion rates by 10-20% with PostHog - PostHog
SU004 PostHog How Supabase 10Xed with the help of PostHog - PostHog As a result—with AI builders as an important piece of the puzzle—we have already 10Xed our weekly user acquisition.
SU005 PostHog How Phantom enhanced its product and cut failure rates by 90% - PostHog
SU006 PostHog How ElevenLabs uses every tool PostHog has to launch new features - PostHog
SU007 PostHog How Lovable builds better agents with AI Observability and experimentation - PostHog
SU008 PostHog How Arena uses PostHog to ship without bias at the AI frontier - PostHog Over the past six months alone, event volume increased 19×.
SU009 PostHog Why Exa loves PostHog AI - PostHog
SU010 PostHog How ResearchGate tests product changes for over 25M users - PostHog We have hundreds of millions of pageviews per month.
SU011 PostHog PostHog pricing – Transparent, usage-based, generous free tier Our generous free tier means more than 90% of companies use PostHog for free.
SU012 PostHog About PostHog Since then, we've grown far beyond analytics ... used by 190254+ teams.
SU013 PostHog Product OS – PostHog
SU014 G2 The G2 on PostHog The frequent crashes during loading can be quite irritating at times, and I also receive a large number of incident emails.
SU015 Sacra PostHog revenue, valuation & funding
SU016 Mixpanel 2026 State of Digital Analytics: Benchmarks, analysis, and recommendations | Signals & Stories
SU017 SlashData Developer Population Sizing | SlashData Software Developer Insights & Research
SU018 Y Combinator Y Combinator
SU019 Hasura Hasura: Creator of PromptQL, Data Delivery Network & GraphQL Engine
SU020 Supabase Supabase | The Postgres Development Platform.
SU021 Phantom Phantom: The money app that'll take you places
SU022 ElevenLabs Free AI Voice Generator & Voice Agents Platform | ElevenLabs
SU023 Lovable AI App Builder | Vibe Code Apps & Websites with AI, Fast
SU024 Exa Exa
SU025 ResearchGate ResearchGate | Find and share research
SR001 PostHog Privacy policy, PostHog style Posthog complies with the EU-U.S. Data Privacy Framework, the UK Extension to the EU-U.S. Data Privacy Framework, and the Swiss-U.S. Data Privacy Framework.
SR002 PostHog Terms, PostHog style IN NO EVENT ... WILL NOT EXCEED, IN THE AGGREGATE, THE GREATER OF (i) ONE THOUSAND DOLLARS ($1,000), OR (ii) THE FEES PAID TO POSTHOG HEREUNDER IN ONE YEAR PERIOD ENDING ON THE DATE THAT A CLAIM OR DEMAND IS FIRST ASSERTED.
SR003 PostHog DPA Processor confirms that it participates in the EU-US Data Privacy Framework, the UK Extension to this Framework and the Swiss-U.S. Data Privacy Framework.
SR004 PostHog Public post-mortems - Handbook - PostHog We publish a public post-mortem when an incident results in permanent impact on user data ... or result in extended unavailability of PostHog services for customers.
SR005 PostHog Security advisories - Handbook - PostHog Currently, there are no active security advisories or CVEs. All is well.
SR006 PostHog Product OS – PostHog
SR007 PostHog PostHog pricing – Transparent, usage-based, generous free tier Our generous free tier means more than 90% of companies use PostHog for free.
SR008 PostHog Product OS - Docs - PostHog
SR009 PostHog CDP sources & destinations PostHog's CDP makes it easy to transform events as they arrive, and sync them over to other services that you use to run your business.
SR010 PostHog Careers - PostHog Starting a job at a company of 200+ people can be intimidating!
SR011 PostHog Future - Handbook - PostHog TL;DR: Mid term, it's $100 million ARR by 2026, working backwards from there.
SR012 PostHog Security & Privacy - Handbook - PostHog If a customer is using PostHog Cloud, then PostHog is acting as Data Processor and the customer is the Data Controller.
SR013 PostHog Post-mortem of Shai-Hulud attack on November 24th, 2025 - PostHog By 9:30 AM UTC, we had identified the malicious packages, deleted them, and revoked the tokens used to publish them.
SR014 PostHog Privacy compliance - Docs - PostHog Yes, we can provide a Business Associate Agreement (BAA) to enable HIPAA-compliant usage of PostHog Cloud.
SR015 PostHog Logs data loss in US Cloud - PostHog Status Regretfully, we have confirmed data loss for all customer logs for the new Logs product in the US cloud region up until 16th February 21:00 UTC.
SR016 Data Privacy Framework Data Privacy Framework participant detail for PostHog EU-U.S. Data Privacy Framework : Active ... UK Extension to the EU-U.S. Data Privacy Framework : Active ... Swiss-U.S. Data Privacy Framework : Active
SR017 OpenCVE Posthog CVEs and Security Vulnerabilities CVE-2025-1520 ... 8.0 High ... SQL Injection Remote Code Execution Vulnerability.
SR018 StatusGator PostHog Status. Check if PostHog is down or having an outage. | StatusGator The last officially acknowledged outage was on May 21, 2026.
SR019 StatusGator PostHog App Status. Check if PostHog App is down or having an outage. | StatusGator
SR020 StatusGator PostHog Workflows Status. Check if PostHog Workflows is down or having an outage. | StatusGator
SR021 StatusGator PostHog Logs Status. Check if PostHog Logs is down or having an outage. | StatusGator
SR022 StatusSight PostHog Status: check for PostHog outages and issues - StatusSight PostHog status is Operational ... Active Incidents No active incidents
SR023 EagleStatus PostHog Status. Check if PostHog is down or having issues. | EagleStatus PostHog status is UP ... Recent PostHog events May 21, 2026, 3:00 PM ... PostHog / US Cloud / App
SR024 iLert Posthog: npm installs risked secret exfiltration for 5 hours TTD (Time to Detect): ~5h19m (04:11 → 09:30 UTC).
SR025 Daily Security Review PostHog Hit by Shai-Hulud 2.0 npm Worm Through CI/CD Automation Flaw this breach was facilitated by an automation flaw in the continuous integration and delivery (CI/CD) workflow
SR026 G2 PostHog Reviews 2026: Details, Pricing, & Features | G2 The frequent crashes during loading can be quite irritating at times, and I also receive a large number of incident emails.
SR027 Sacra PostHog revenue, valuation & funding Sacra estimates that PostHog hit $57.5M in annual recurring revenue (ARR) in February 2026.
SR028 4 Day Week PostHog Remote Jobs & Careers - Flexible Hours
SR029 Tracxn PostHog The company has 2641 active competitors ... with a current valuation of $1.4B.
SR030 Forrester Feature Management And Experimentation — An Evolving Market feature flags are the domain of developers and experimentation is the domain of product and marketing
SV001 PostHog PostHog pricing – Transparent, usage-based, generous free tier
SV002 PostHog Product OS – PostHog
SV003 PostHog PostHog raises a series D (and a small C)
SV004 PostHog Future - Handbook - PostHog
SV005 PostHog About PostHog
SV006 PostHog How Y Combinator used PostHog experiments to boost engagement by 40%
SV007 PostHog How Hasura improved conversion rates by 10-20% with PostHog
SV008 PostHog Careers - PostHog
SV009 Sacra PostHog revenue, valuation & funding
SV010 Contrary Research Report: PostHog Business Breakdown & Founding Story
SV011 The Economic Times SaaS startup PostHog turns unicorn after $75 million fundraise led by Peak XV
SV012 Entrepreneur India PostHog Becomes Unicorn With USD 75 Mn Funding Led by Peak XV
SV013 Peak XV Partners PostHog
SV014 Y Combinator PostHog: The single platform to analyze, test, observe, and deploy new features
SV015 Grand View Research Product Analytics Market Size, Share & Growth Report, 2030
SV016 Expert Market Research Product Analytics Market Size, Share, Analysis | Report 2035
SV017 Mordor Intelligence Product Analytics Market Size, Competitive Landscape, Trends 2026–2031
SV018 Mixpanel 2026 State of Digital Analytics: Benchmarks, analysis, and recommendations
SV019 Forrester Feature Management And Experimentation — An Evolving Market
SV020 SlashData Developer Population Sizing
SV021 PostHog Public post-mortems - Handbook - PostHog
SV022 PostHog Post-mortem of Shai-Hulud attack on November 24th, 2025
SV023 Atlassian Atlassian - Financials - Quarterly results
SV024 Securities and Exchange Commission Atlassian 2025 Form 10-K
SV025 Securities and Exchange Commission Atlassian 2026 Q3 Form 10-Q
SV026 Securities and Exchange Commission Datadog 2025 Form 10-K
SV027 Securities and Exchange Commission Datadog 2026 Q1 Form 10-Q
SV028 Securities and Exchange Commission Amplitude 2025 Form 10-K
SV029 Securities and Exchange Commission Amplitude 2026 Q1 Form 10-Q
SV030 SaaSValuation.io Public SaaS Multiples | Valuation Benchmarks
SV031 Multiples.vc Public Software Valuation Multiples — May 2026
SV032 Livmo SaaS Valuation Multiples 2026: 3x to 12x ARR Data