validating PMF: retention beats surveys every time

localpack

Been chewing on this for a while. The Sean Ellis 'very disappointed' survey is a decent starting point, but honestly it's more of a vibe check than proof. Hitting ₄₀% is a signal, sure, but without hard behavioural data to back it up, it's just noise.

Stuff like Day-30 retention, repeat usage, and organic referrals - that's where the rubber meets the road. For B2B especially, one company loving your tool isn't enough. You need to see that pattern repeat across a few different customer profiles before you start dumping cash into ads.

People say they love things all the time, then never use them again. Seen it in my own ad accounts: we'd scale based on happy survey responses, only to watch churn spike three months later. Now the rule is: no budget until month-3 retention above 40% across multiple segments - no nudging, no re-engagement emails.

With ad attribution getting messier every quarter, first-party retention and activation data is carrying way more weight. Cohort behaviour is doing the heavy lifting that channel data used to do.

The signal that made me confident enough to scale? When we saw 60%+ month-3 retention across three different customer profiles without any automations. That's when I knew we could safely turn on the ad tap.

tiktokguru

For me, it was all about actual behaviour over words. 📊 A couple of segments hit roughly 60% retention by month three with zero nudging - way more reliable than surveys. People say they love something, then ghost. 👻

communityfirst

Nothing beats that moment when retention just clicks into place and growth feels effortless. From my work with e-com brands, the clearest PMF signal is when you're not the one pushing the flywheel anymore - customers bring each other in, activation curves smooth out, and churn stops being about "not useful" and starts being about timing or budget.

Surveys hint, but cohort behaviour tells the truth. The moment you see referrals happening without any incentive, that's when you know you've built something people can't stop talking about. Scaling then becomes about clearing the path, not forcing the pace.

wordweaver

Oh, I've watched this happen in the wild. Founders end up with a focus group of professional opinion-givers rather than actual early adopters desperate for a fix. It's like asking your mates what they think of your new haircut-nice, but useless for validation.

24fire

Glad someone pointed this out. In B2B especially, that first happy client feels like validation, but more often than not it's just a lucky alignment - maybe your product solved a niche problem their team happened to have, or your sales process accidentally matched their buying cycle. The real signal comes when you see the same pattern of adoption across different segments: the scrappy five-person startup, the fifty-person mid-market team, and the enterprise with procurement layers. That's when you stop wondering if you built a feature and start knowing you built a platform. It's like building a bridge - one successful crossing doesn't prove the structure, you need different loads, different weather, different speeds. Same principle in product-market fit.

localpack

Month 3 retention without any in-app nudges is the only metric I actually trust. Surveys are pure vanity - people tell you what makes them look good. But that unprompted usage pattern? That's the real signal. Ran this playbook on a $500k/month client: stopped chasing survey NPS, started watching login frequency in weeks 8-12. That told us exactly when to scale the ad spend.

wordweaver

Duct-tape signals never lie. Love hearing what people's back-alley workarounds actually reveal about real demand.