App增长停滞时的诊断框架:四步排查与漏斗联动分析
Your app's growth has stalled: what now?
提供了一套从心态建设到漏斗拆解的完整诊断流程,特别是“预演失败”和“联动效应”视角,创业者可直接用于复盘当前业务瓶颈。
We've all had it: that week when the numbers aren't quite what you hoped. Fewer new subscribers than the week before. Maybe churn creeps up. You tell yourself it's just a blip, that next week will be better.
我们都经历过这种情况:那一周的数据并不如你所愿。新订阅用户比上周少。也许流失率也在悄然上升。你告诉自己这只是一个短暂的小波动,下周情况会好转。
Then one week becomes two, two becomes four, and as the graph keeps drifting down, a low-level panic starts to set in.
然后一周变成两周,两周变成四周,随着图表持续下滑,一种低强度的恐慌感开始蔓延。
What’s happening? What are we doing wrong? Whose fault is it?
到底发生了什么?我们做错了什么?这是谁的错?
Every great growth story I've heard has a stall somewhere in it: a plateau, a setback, a period when nothing seemed to work. It's almost never the clean hockey-stick graph we love to post on LinkedIn. #GrowingInPublic — but not that publicly.
我听过的每一个伟大的增长故事里都有一个停滞期:一个平台期、一次挫折、或是一段似乎什么都不奏效的时期。它几乎从来不是我们在 LinkedIn 上喜欢发布的那种干净的曲棍球棒式增长曲线。#GrowingInPublic——但并没有那么公开。
Growth stalls are normal, but finding the answers for what causes them is tough. I've experienced it both as Head of Growth at Heights and while advising clients, telling myself the same reassuring things I'm about to tell you. So I know how it feels, not just how to fix it.
增长停滞是正常的,但要找到导致停滞的原因却很难。我既在 Heights 担任增长负责人时经历过,也在为客户咨询时经历过,当时我也对自己说着即将告诉你的那些令人安心的话。所以我知道那种感受,而不仅仅是知道如何修复它。
What not to do when your app’s growth stalls
当应用增长停滞时不该做的事
It's tempting to play the blame game. Meta is often the favorite culprit, or perhaps a teammate who isn't pulling their weight (or refilling the coffee). But when you dig into the data, most stalls are actually within your control. Research into why companies stop growing found that only around 13% of growth stalls are driven by external factors. The rest come down to internal decisions — and most companies experience at least one major stall.
互相指责是很诱人的。Meta 往往是首选的替罪羊,或者可能是某个没有尽到责任(或没去 refill 咖啡)的队友。但当你深入分析数据时,大多数停滞实际上是在你的控制范围内的。关于公司为何停止增长的研究发现,只有约 13% 的增长停滞是由外部因素驱动的。其余的都归结为内部决策——而且大多数公司至少会经历一次重大停滞。
The worst response is to become a headless chicken: changing everything at once, chasing every new idea, and hoping something sticks.
最糟糕的反应是像无头苍蝇一样乱撞:一次性改变所有事情,追逐每一个新点子,并指望其中某个能奏效。
Instead, I want to help you work through a growth stall methodically, from diagnosing what's actually causing it to getting growth moving again.
相反,我想帮助你系统地应对增长停滞,从诊断真正的原因到让增长再次动起来。
The growth stall diagnostic
增长停滞诊断
Think of this as growth therapy. My goal is to reassure you, reduce the panic, and leave you with a clear plan rather than a long list of random experiments.
把这当作增长疗法吧。我的目标是让你安心,减少恐慌,并给你留下一个清晰的计划,而不是一长串随机的实验列表。
Over the years, I've developed a simple framework for working through these moments. It comes down to four questions:
多年来,我开发了一个简单的框架来处理这些时刻。它归结为四个问题:
- Is it a blip or a stall?
- Is it within your control?
- Where exactly is growth breaking down?
- What should you do about it?
- 这是一个短暂的小波动还是一个停滞?
- 这在你的控制范围内吗?
- 增长究竟在哪里出现了问题?
- 你应该怎么做?
Most people panic and jump straight to question four. But the real value lies in taking the time to answer the first three questions — that’s what tells you which action will actually work.
大多数人会惊慌失措,直接跳到第四个问题。但真正的价值在于花时间回答前三个问题——这才是告诉你哪些行动真正有效的方法。
Step 1: Is it a blip or a stall?
第一步:这是一个短暂的小波动还是一个停滞?
If we chased every number that dipped for a week, we'd be very fit from all the running around — and completely exhausted.
如果我们追逐每一个下降了一周的数字,我们会因为到处奔波而变得非常“健壮”——并且完全筋疲力尽。
There’s a lot of noise in growth, especially early on when numbers can swing wildly. So before you panic, define your thresholds: what counts as a real miss vs. normal variation?
增长领域充斥着大量噪音,尤其是在早期阶段,数据可能会剧烈波动。因此,在恐慌之前,先定义你的阈值:什么算作真正的失误,什么只是正常波动?
If you typically grow 5–10% week-on-week and occasionally dip, is a 20% drop the point where you investigate? The answer will be different for every business, but having those guardrails is how you separate real signals from seasonality and one-off blips.
如果你通常每周环比增长 5–10%,偶尔出现下滑,那么 20% 的跌幅是否就是你开始调查的临界点?每个企业的答案都不同,但拥有这些护栏机制,才能将真实信号与季节性因素和一次性异常值区分开来。
Seasonality is especially easy to miss. Once you have enough data, patterns start to emerge, and you can look back at a slowdown with more context. I have a client now in their eighth year who told me: “We’ve grown a bit this quarter — not loads — but these aren’t the months we usually grow in, so we’re not worried. We’re using the time to fix things and build our infrastructure.”
季节性因素尤其容易被忽视。一旦积累了足够的数据,模式就会显现,你可以结合更多背景信息来回顾增速放缓的情况。我有一位客户的企业已进入第八年,他告诉我:“这个季度我们略有增长——不算多——但这些通常不是我们增长的月份,所以我们并不担心。我们正利用这段时间修复问题并构建基础设施。”
That perspective is much harder to have when you’re smaller, which is why you need to look for trends over several weeks, and resist reacting to a single bad one. Don’t obsess over daily swings; focus on week-on-week movement.
当企业规模较小时,很难保持这种视角,因此你需要观察数周的趋势,并抵制对单一糟糕数据的反应冲动。不要沉迷于每日的波动;应关注每周环比的变化。
And when you catch yourself thinking, “This doesn’t look like a blip anymore”, treat it as a stall. Bring the team together and approach it like an experiment. Start with a pre-mortem: imagine it’s three months from now and growth is still flat. What went wrong?
当你发现自己想‘这看起来不再是一次性异常了’时,将其视为停滞。召集团队,像进行实验一样去应对它。先从事前验尸(pre-mortem)开始:假设三个月后增长依然停滞,哪里出了问题?
Step 2: Is the blocker within your control?
第二步:阻碍因素是否在可控范围内?
Your circle of control is everything you can actually influence: the internal factors that shape your growth.
你的控制圈包括所有你能实际影响的事物:塑造你增长的内部因素。
Things like:
例如:
- What you’re focusing on
- What your processes look like
- Which channels you’re investing in
- How you’re positioning yourself
- 你关注的重点
- 你的流程状况
- 你投入资源的渠道
- 你的定位方式
There’s usually more within your control than you think. Then there’s everything outside it:
实际上,你在可控范围内的余地通常比你想象的要大。然后是那些超出控制范围的因素:
- Seasonality
- New competitors
- Policy shifts
- Platform changes
- Market saturation
- 季节性
- 新竞争对手
- 政策变化
- 平台变更
- 市场饱和
The first job is to figure out which bucket your stall belongs in.
首要任务是确定你的停滞属于哪一类。
Even when something external is clearly playing a role, the next question is: how do you bring it back into your circle of control? You can’t delete a new competitor, but you can decide how you respond, through your pricing, positioning, product, or messaging.
即使外部因素显然在起作用,下一个问题是:你如何将其拉回你的控制圈?你无法消除新竞争对手,但你可以通过定价、定位、产品或信息传递来决定如何应对。
As the research suggests, external factors are rarely the whole story. Once you’ve ruled them in or out, focus your energy where you have the most leverage: the things you can actually change.
正如研究所表明的那样,外部因素很少是故事的全部。一旦排除了它们或确认了它们的影响,就将精力集中在你拥有最大杠杆效应的地方:那些你真正能改变的事情。
I spoke to a wellness app last year at App Growth Annual that was facing this exact challenge. They told me AI had become an easy replacement with just a few smart prompts. And it wasn’t only a fear; their growth was slowing.
去年在 App Growth Annual 会议上,我与一家健康应用公司进行了交流,他们正面临这一完全相同的挑战。他们告诉我,AI 只需几个智能提示词就能轻易成为替代品。这不仅是担忧,他们的增长确实在放缓。
That can feel completely out of your control. But the more useful question is: what parts of the problem are within your control?
这可能会让你感觉完全失去了控制。但更有用的问题是:问题的哪些部分在你的掌控之中?
The advice I gave them was to go back to their core value proposition and ask:
我给他们的建议是回到核心价值主张,并问自己:
- What value can we offer that AI cannot?
- What other ways can we help users achieve their Job to Be Done?
- What parts of the experience become more valuable when technology gets better, rather than less?
- 我们能提供哪些 AI 无法提供的价值?
- 我们还有哪些其他方式可以帮助用户完成他们‘待办任务’(Job to Be Done)?
- 在哪些体验环节,随着技术进步而非退步,其价值会变得更加显著?
For example, yes, AI can generate a similar wellness plan. But it’s much harder for it to keep users accountable, automatically collect behavioral data and insights, and continuously adjust the plan based on what’s actually happening in someone's life.
例如,是的,AI 可以生成类似的健身计划。但它很难做到让用户保持自律、自动收集行为数据和洞察,并根据某人生活中实际发生的事情持续调整计划。
Competing with AI would be like David vs. Goliath, but they can differentiate their app from AI and double down on the parts of the experience that create deeper, longer-term value.
与 AI 竞争无异于大卫对战歌利亚,但他们可以让自己的应用与 AI 区分开来,并加倍投入那些能创造更深、更长期价值的体验环节。
Step 3: Where exactly is growth breaking down?
第三步:增长究竟是在哪里出现瓶颈的?
This is the step people skip, and it’s the one that matters most. Instead of guessing at a cause, audit your growth the way an outside consultant would: look at every part of your funnel and (just as importantly) at how you’re working. It can feel like you’re slowing down when every instinct is telling you to move faster. I promise you, this is the thing that speeds you up.
这是人们容易跳过的一步,却是最关键的一步。与其猜测原因,不如像外部顾问那样审计你的增长情况:审视漏斗的每一个环节,以及(同样重要的是)你们的工作方式。当所有本能都催促你加速时,这样做可能会让你感觉像是在减速。但我向你保证,这才是真正能让你提速的方法。
Start by getting clear on your North Star metric: the one metric that best defines success for your business. What is it, and what inputs actually drive it?
首先明确你的北极星指标(North Star metric):最能定义你业务成功的那个指标。它是什么?哪些输入因素真正驱动着它?
Then evaluate each area separately, as if you were looking at someone else’s company:
然后,像审视别人的公司一样,分别评估每个领域:
- Acquisition
- Activation and engagement
- Retention and revenue
- Monetization
- Team structure and approach
- 获客(Acquisition)
- 激活与参与度(Activation and engagement)
- 留存与收入(Retention and revenue)
- 变现(Monetization)
- 团队结构与工作方式
For each area, ask:
对于每个领域,问自己:
- What’s working?
- What isn’t?
- How are these pieces influencing each other?
- 什么在起作用?
- 什么不起作用?
- 这些环节是如何相互影响的?
FYI
附注
I include team structure deliberately, because many stalls aren’t caused by a single broken metric. They come from how a team operates: where decisions are made, what gets prioritized, and where a founder is spending their attention.
我特意将团队结构纳入其中,因为许多停滞并非由单一指标的失效引起。它们源于团队的运作方式:决策在哪里做出、什么被优先考虑,以及创始人将注意力集中在何处。
The trap here is auditing each area in isolation and declaring victory the moment one number improves. Ekaterina Gamsriegler, who led growth at coding app Mimo, has a great example of this: LTV had plateaued while acquisition costs kept climbing, the classic growth-stage stall.
这里的陷阱是孤立地审计每个领域,并在其中一个数字改善时就宣布胜利。Mimo(一款编程应用)的增长负责人 Ekaterina Gamsriegler 就有一个很好的例子:LTV(用户终身价值)趋于平缓,而获客成本却不断攀升,这是典型的增长阶段停滞现象。
The solution came from understanding the funnel as a connected system. A paywall change shifted which plans users chose, which changed LTV, which then changed how much the company could afford to spend on acquisition. The result: Mimo increased customer LTV by 65% and reduced paid CAC by 20% in under a year.
解决方案源于将漏斗视为一个相互关联的系统。付费墙的调整改变了用户选择的套餐,进而影响了客户终身价值(LTV),最终改变了公司在获客上可承受的花费。结果:Mimo 在不到一年的时间内将客户 LTV 提升了 65%,并将付费获客成本(CAC)降低了 20%。
Ekaterina’s rule applies here too: before any change, ask three questions:
Ekaterina 的规则同样适用于此:在任何变更之前,先问三个问题:
- What is the first-order effect?
- What are the downstream effects?
- What could quietly get worse as this gets better?
- 一阶效应是什么?
- 下游效应有哪些?
- 随着某方面变好,哪些方面可能会悄然恶化?
The third is about accepting tradeoffs. I've seen price increases give ARPU a big lift, while the number of new customers dropped. Is that a tradeoff you are willing to make?
第三点关乎接受权衡。我曾见过提价使每用户平均收入(ARPU)大幅提升,但新客户数量却下降。这是你愿意做出的权衡吗?
更进一步:量化金融体系
看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力