The great enterprise 'ghosting'
A guide to avoiding the SaaS silent treatment
Been here before?
“Hey you seem upset, what’s wrong?”
“...Nothing.”
“Are you sure?”.
“...Yep.”
“Okay, no problem, I’ll stop asking.”
Pro tip: big problem.
I’m seeing the same dynamic play out across software. People and organisations are keeping subscriptions to tools long after they’ve stopped actually using them.
In the consumer world, it’s Netflix. Plenty of people have active subscriptions even though they barely log in. In the enterprise world, it’s things like Illustrator. The license is still there, but more and more of the work is quietly shifting somewhere else.
This is stealth churn.
It’s the gap between paid retention and real product habit. The contract renews but the users don’t.
Not all revenue is created equal
The market used to treat revenue as the ultimate metric for the bright future of SaaS companies.
I’m increasingly convinced it’s a lagging indicator. Not all growth is created equal, especially now that AI is changing where and how work happens.
Yes, an elite sales team can sell more licenses, more adoption “commitments” or more seats. But... if those features aren’t being adopted and integrated into real workflows, or even worse, if they aren’t being used... that’s a problem. A really big one.
What I’m now seeing is the rising importance of feature utilisation, especially around AI features.
It’s no longer enough to say “we’ve shipped the latest and greatest AI feature” (usually a glorified chatbot). The question is... are people actually using it every week to get their work done?
Ultimately, it comes down to three growth metrics:
DAU: daily active users
WAU: weekly active users
MAU: monthly active users
The holy rhyming trifecta. They predict renewal loss long before ARR ever will.
So, how can we judge companies based on this?
Let’s look at Figma
Figma gave investors the metric I want from every B2B AI company.
In Q4 2025, they reported 303.8 million dollars of revenue, up 40 percent year over year.
Great. Revenue led. But then they went a level deeper.
In the same release, weekly active users of Figma Make grew over 70 percent quarter over quarter and over half of Figma’s 100,000-plus customers were building in Make every week.
That’s incredible. That’s a very different story.
Revenue is reported year over year. Make usage is growing quarter over quarter.
Yes, Figma calculates quarterly weekly active user growth using the highest-usage week in each quarter.
So be it. Find me a company without some level of financial engineering.
Even with that caveat, what they’re really showing is habit before full monetisation.
In B2B AI, paid contracts are late. Qualified usage is early.
So I’ve been developing a pretty simple personal test.
In any application you use. Inside a like for like customer cohort, does qualified usage grow ahead of revenue? If not, all those lucrative annual contracts can hide a serious amount of decay.
Basically, a customer can keep paying while the work has already moved somewhere else. It’s the beginning of the end.
The app stays in the budget. The habit leaves the building… for now.
AI-native startups are closing this gap fast. But we’re also seeing a wave of AI-native tools with fragile economics.
When every Tom, Dick, and Harry can ship an “AI-native, agentic clone” of an existing SaaS app, the question becomes... who actually sticks?
On average, AI-native companies at $250,000 dollars-plus ARR have around 40 percent median gross revenue retention and 48 percent median net revenue retention.
They’re acquiring customers and then bleeding them. People are trying the tools. They’re not keeping them.
The AI natives that are winning are far exceeding those statistics and the difference, again, is usage.
Enter Figma Make, again
By Q1 2026, roughly 60 percent of Figma’s 100,000-plus customers used Figma Make weekly, up from “over half” in Q4.
Figma’s net dollar retention reached 139 percent, and Pro teams that bought AI credit add-ons had more seats and more than three times the average ARR of teams that didn’t.
Usage spread through valuable accounts. Revenue followed.
There’s a big “but,” though.
Figma hasn’t disclosed everything.
We don’t know the absolute starting user base behind that “over 70 percent” Make growth figure. We also have to remember the “highest week” calculation.
But even with those caveats, it ties the AI product to weekly behaviour inside large customers. It shows WAU. It connects AI directly to habit.
Be like Figma Make.
Put cohort usage beside cohort revenue.
Show weekly work completed, dormant paid accounts, active-seat penetration, and usage concentration.
Let the two charts sit next to each other and make you uncomfortable.
ARR tells you the customer hasn’t cancelled. It does not tell you that your product is still where the work happens.
So, when you ask “What’s wrong?” and they say “Nothing”... it’s probably something big.
There will be leading indicators. If you’re willing to stare at them, you might actually figure it out in time... and avoid the silent treatment that follows.
Hope you enjoyed reading.
Blake