There's a reckoning happening in product-led growth.
Last week, Jay Stansell published a piece in Product Coalition called "The PLG Mirage" that traced the arc of PLG enthusiasm from peak hype to today's more honest conversation. His observation was sharp: "The earlier the piece, the more certain the author sounded. By 2023, everyone was hedging."
That hedging has turned into a full-on strategic pivot for many teams. The original PLG promise—that great products sell themselves, that if sales and marketing disappeared you'd still grow—is colliding with a stubborn reality: knowing that users are leaving isn't the same as knowing why.
And it turns out, the "why" is everything.
The Dashboard Paradox
Here's what a typical PLG team sees in their customer success platform: login frequency dropping, feature adoption declining, session duration shrinking. Red flags everywhere. The health scores are cratering.
But what do those metrics actually tell you?
They tell you that something is wrong. They don't tell you what. And by the time usage metrics turn red, you're often looking at a decision the customer made weeks or months ago. You're measuring the symptoms of churn, not its causes.
This is the dashboard paradox that's driving a growing number of customer success leaders to rethink their entire approach. According to recent industry analysis, 72% of CS leaders now identify "understanding the reasons behind churn" as their top-three challenge. Not tracking churn. Not measuring churn. Understanding it.
That gap—between seeing the what and knowing the why—is where billions of dollars in recurring revenue quietly disappear.
Why PLG Teams Are Especially Vulnerable
Product-led growth was supposed to be the antidote to bloated sales organizations and enterprise complexity. Let the product do the selling. Let users self-serve. Build something so good that it spreads virally.
And for a handful of companies, it worked exactly that way. Slack. Zoom. Dropbox. Figma. These became the poster children for an entire movement.
But there's a survivor bias problem here that the industry is finally starting to acknowledge. As PLG founder Guy Barner wrote in Product Coalition: "Even though we went with PLG, I still don't think it's the best choice for most startups."
His reasoning is instructive. PLG requires you to nail everything at once: lead generation, website conversion, onboarding, product quality, and a freemium model that actually converts. With sales-led, you need a good product and good salespeople. The margin for error in PLG is razor-thin.
And here's the kicker—a stat from Noa Ganot that should give every early-stage PLG team pause: research shows that up to $10M ARR, it's much easier to succeed without PLG than with it. After $10M ARR, PLG wins big. But most PLG advice is aimed at teams nowhere near that threshold.
This creates a perfect storm. Teams optimized for self-serve motion often have minimal customer touchpoints. They're swimming in usage data but starving for context. They can tell you exactly how many times a user clicked on Feature X, but they have no idea what that user said about Feature X in their last support ticket, or what they complained about to their team in Slack.
The Shift to Qualitative Intelligence
The smartest CS teams are making a fundamental shift: from lagging indicators (usage metrics that show decline after it happens) to leading indicators (qualitative signals from conversations that predict churn before it shows up in dashboards).
This isn't just theory. The customer success platform market is projected to reach $3.1–3.5 billion by 2028, and a new category of tools is emerging that combines traditional health scoring with what's being called "conversation intelligence"—the ability to analyze unstructured data from calls, tickets, and messages to surface what customers are actually thinking and feeling.
The logic is simple. Usage metrics are outputs. Customer conversations are inputs. If you want to understand churn before it happens, you need to look at the inputs.
Consider the difference:
Usage metrics tell you: User hasn't logged in for 14 days.
Conversation intelligence tells you: User mentioned in three separate support tickets that they're frustrated with the reporting feature, asked about competitors, and expressed concern about upcoming contract renewal.
One of these is a red number on a dashboard. The other is a story you can actually act on.
What Research Shows About Preventable Churn
Here's a stat that should change how you think about retention: research indicates that up to 85% of churn is preventable—if you catch the signals early enough and respond appropriately.
Eighty-five percent. That's not a small optimization. That's the difference between a business that compounds and one that's constantly replacing lost revenue.
The challenge is that those signals are often scattered across dozens of touchpoints: support tickets, sales calls, renewal conversations, in-app feedback, NPS responses, social media mentions, community forums. No single human can process all of that. And traditional CS platforms weren't built to synthesize unstructured qualitative data.
This is why AI-powered customer insights have moved from "nice to have" to "strategic necessity." According to recent HubSpot research, 62% of customer service specialists say AI helps them understand buyers better, and 78% report it helps them focus on the most crucial aspects of their job.
The irony is rich. PLG was supposed to let the product do the selling without human intervention. But sustainable PLG requires deep human understanding at scale—exactly the kind of understanding that AI makes possible.
Beyond Dashboards: What Actually Works
So what does effective retention look like in 2026?
It's not about choosing between quantitative and qualitative. It's about integrating both into a coherent picture of customer health that actually predicts behavior, not just describes it.
Listen to every channel. Your customers are telling you why they're frustrated, what they need, and what would make them stay. But they're telling you in support tickets, in sales calls, in Slack messages, in NPS responses. If you're only looking at usage data, you're ignoring 90% of the signal.
Look for patterns, not just metrics. A single complaint about your reporting feature is noise. Ten customers mentioning the same issue across different channels is a pattern. Twenty customers connecting that issue to their consideration of competitors is a leading indicator of churn. AI excels at surfacing these patterns from unstructured data at scale.
Act on insight, not intuition. The gap between insight and action is where most retention programs fail. Understanding that enterprise customers are frustrated with your onboarding is useless if it takes three months for that insight to reach the product team. The organizations winning on retention are building closed-loop systems where customer insights flow directly into product prioritization.
Measure what matters. Stop celebrating metrics that feel good but don't predict outcomes. Net Promoter Score is famous for being gamed and poorly correlated with actual retention. Usage metrics tell you the past, not the future. The metrics that matter are the ones that give you time to act.
The Hybrid Future
Here's the uncomfortable truth that the PLG movement is finally accepting: pure product-led growth—the dream of a product so perfect it needs no human intervention—was always a mirage for most companies.
As Mart Objartel noted in Product Coalition, companies shift between strategies all the time, "sometimes subtly, akin to the proverbial frog boiling." The danger isn't picking the wrong strategy. It's changing direction without building the insights infrastructure to support the new approach.
What actually works for most teams is a hybrid. A product good enough that users want to try it. Self-serve onboarding that doesn't require hand-holding. But also—critically—a continuous feedback loop that ensures you understand what customers are experiencing, not just what they're clicking.
Microsoft Teams didn't beat Slack on product experience. It won by being bundled with Office 365. The PLG darling lost to enterprise distribution. But here's the thing: Slack could have seen it coming. The signals were there in customer conversations, in the enterprise accounts discussing procurement constraints, in the IT teams asking about compliance features that Slack hadn't prioritized.
The winners in the next phase of SaaS won't be the teams with the prettiest dashboards or the most sophisticated usage tracking. They'll be the teams who actually understand their customers—who can synthesize signals from every touchpoint and turn that understanding into action before it's too late.
The Bottom Line
The PLG mirage is fading. What's emerging in its place is a more honest conversation about what sustainable growth actually requires: not just great products, but great understanding of the humans using those products.
Usage metrics will always have their place. But they're the end of the story, not the beginning. If you want to predict churn, prevent attrition, and build the kind of retention that compounds over years, you need to start where your customers start: with what they're actually saying.
The tools to do this at scale finally exist. The question is whether your team will embrace them—or keep watching the dashboards while customers quietly walk away.
Building a product-led business and struggling to understand why customers churn? Pelin helps product teams synthesize customer feedback from every channel—support tickets, sales calls, NPS responses, and more—into actionable insights that predict churn before it shows up in your usage metrics. See how it works →