Understand · Likely to buy and likely to cancel
Churn prediction and upgrade prediction for SaaS
Atlis predicts which free and trial users are likely to buy and which paying customers are likely to cancel, by comparing them with your own users who did. Every prediction shows its reasons, so you know what to say. It works from day one with common signals and learns from your data as you grow.
Growth, Scale and every 14-day trialUpdated
- Invited a teammate
- Visited pricing twice
- Activated on day one
How it works
How does churn prediction work?
- 1
Atlis compares your users
What people did before they paid, against everyone else, and what paying customers showed in their last 30 days before cancelling, against those who stayed.
- 2
It finds the strongest signals
Actions, key page visits, features people stopped using, a business email, activating fast and more, each with how much more often it led to paying or cancelling.
- 3
It flags the people
Likely to buy and Likely to cancel list people with at least two of those signals, strongest first, with the reasons on their profile.
What you get
What's included in Likely to buy and likely to cancel
What makes people pay
The actions that go with paying in your product, with how many people did each and how much more often they paid.
Likely to buy
Free and trial users active in the last 14 days who have at least two buying signals and look at least 3 times as likely to pay.
What comes before cancelling
What customers who cancelled showed in their last 30 days, such as a cancel page visit or a feature they stopped using.
Likely to cancel
Paying customers with at least two of those signs, including something they did or stopped doing, who look at least 4 times as likely to cancel.
Reasons on every profile
Each person's profile shows why they were flagged and suggests the right email: an upgrade note or a check-in.
Alerts the moment it happens
Owner alerts in the bell, in Slack or by webhook when someone new looks likely to buy or cancel.
Use cases
How founders use it
Upgrade the right free users
Send a personal note to the people who already behave like your paying customers.
Save customers before they leave
Check in with paying customers whose usage looks like those who cancelled, while there's still time.
Learn what drives revenue
See which actions really go with paying, then build onboarding around them.
FAQ
Likely to buy and likely to cancel: questions and answers
Growth, Scale and every 14-day trial.
- Buying predictions learn from your data once 10 people have paid and 10 haven't; until then Atlis uses common buying signals.
- Cancel predictions start once 10 paying customers have cancelled in the last year and 10 have stayed.
- Plan names and CRM fields are left out of cancel signals, because billing changes them when someone cancels.
- Predictions are statistics from your own users, not a language model.
01How does Atlis predict who will cancel?
It compares what paying customers who cancelled showed in their last 30 days with customers who stayed, finds the signs that were at least twice as common among those who cancelled, and flags current customers with at least two of them.
02How does Atlis know who is likely to buy?
It compares people who signed up 2 weeks to a year ago: of those who did each action, how many paid. Until 10 have paid and 10 haven't, it uses common buying signals such as visiting pricing, starting checkout or inviting a teammate.
03Is this AI?
It's a statistical comparison of your own users, worked out once a day, without a language model. Every prediction shows the signals behind it.
04Which plan includes predictions?
Growth and Scale, and every 14-day trial.
See who's about to buy or cancel, this week.
Connect your app in about 10 minutes. Atlis sorts every user, ranks your next best moves and sends the right email for you.
14 days of Growth free, then free forever up to 250 users · No credit card