We stopped looking at signups. We started looking at intent.
For a long time, we tracked the same metrics everyone else does:
Signups
Traffic
Conversion rates
They were useful, but they rarely answered the question that actually mattered:
Who's genuinely interested in becoming a customer?
Two users can both sign up today.
One spends 30 seconds clicking around and disappears forever.
The other explores features, connects their workflow, comes back the next day, and gradually moves toward buying.
Traditional analytics often treat them similarly.
We've been building Autonomy around a different idea: product behavior reveals intent better than acquisition source alone.
Instead of asking "Where did this user come from?", we're asking:
What did they actually do?
Which actions indicate serious evaluation?
Which behaviors consistently lead to conversion?
It's still an evolving model, and we're learning every week.
One thing we've realized is that not every active user has buying intent, and not every quiet user is unqualified. The interesting part is identifying meaningful patterns before the purchase happens.
I'm curious how other founders think about this.
What behavior has turned out to be your strongest signal that someone is likely to become a paying customer?
Would love to hear what you've discovered in your own products.
Really strong direction — this is basically the shift from vanity metrics → behavioral intent systems, and it’s where most SaaS products eventually end up once acquisition stops being the bottleneck.
One pattern that’s been consistently strong in my experience is not a single action, but a sequence of “commitment signals”:
user returns within a short window (signal of unresolved need)
interacts with “core value” feature more than once (not just exploration)
creates or imports real data (this is usually the biggest separator vs curiosity users)
hits friction and still comes back (quiet but high intent signal)
starts trying to fit the product into their workflow (naming, structuring, integrations)
What’s interesting is that intent usually shows up less as a spike and more as increasing resistance tolerance — they’re willing to deal with friction because the outcome matters.
I also like your framing because it naturally leads into a better segmentation model: not “active vs inactive”, but exploring → testing → integrating → depending.
Curious if you’re planning to turn this into a scoring system or keep it more heuristic/qualitative inside Autonomy for now.
I like the idea of commitment signals more than looking for a single activation event.
The point about users pushing through friction really stands out. If someone is willing to spend time importing real data, tweaking settings, or coming back after hitting a roadblock, they're telling you something that a page view or session count never will.
We're leaning toward treating these behaviors as a sequence rather than independent events. The challenge is that the same action can mean different things depending on what happened before it, so context matters as much as the event itself.
For Autonomy, we're exploring ways to combine these behavioral signals into a dynamic intent score instead of relying on fixed rules. Still early, but the goal is to help teams understand why a user is progressing not just what they clicked.
I'm curious have you found any behavioral signal that's been surprisingly predictive across different products, or does it tend to be highly product-specific?