Last activity, as a timestamp
Every fan record carries when the fan was last active, written alongside the event itself. It is a field you can sort by rather than a position in a thread you have to scroll.
Coming back
Telegram fan retention: the signals that show a fan is still there
Last activity, relationship stage, an availability schedule and a thirty-day access window. What Velvet records about returning, and what it leaves to you.
Retention signalsTelegram fan retentionDirect answer
Velvet records the signals that show whether a fan is still in the relationship: when they were last active, which stage the relationship reached, what they have bought and whether they still hold access. It never concludes on your behalf that a fan is gone, because a quiet month and a finished relationship look identical from the outside.
Retention advice usually arrives as a cadence: message at day three, discount at day fourteen. That is a schedule, not a strategy, and it treats a fan who bought last week the same as one who never has.
The alternative is to read what is actually recorded. Velvet keeps four kinds of signal about whether a fan is still there, and each one supports a different move.
Zero credential sharingChatters never see your Telegram login
Role-based accessEach chatter gets only what they need
Full audit trailEvery message logged, every action tracked
Anti-leak protectionFan data stays in your controlled system
The signals worth reading
The stage the relationship reached
Cold, qualified, active or vip, moved forward automatically by real activity. Where a fan stopped is often more useful than where they are, because it says which step failed.
A live access window
A buyer holds access to what they bought for 30 days. A fan who is inside that window has a reason to reopen the page, which is a different conversation from one who has nothing waiting.
Churn is never automatic
Nothing in Velvet writes off a fan on its own. Marking a relationship as churned is a human decision, because a quiet fan may simply be on holiday.
Reasons you can group by
When a subscriber cancels, the reason comes from a fixed, server-validated list. Twelve creators inventing their own wording produce twelve categories of one; a closed set produces something you can act on.
A published schedule
Recurring shifts give fans an "usually active" signal in their own local time. A fan who knows when you answer has a reason to come back at a specific hour rather than not at all.
Building a win-back routine
- Sort by last activity before writing anything, so the list is ordered by evidence.
- Separate fans who stopped at qualified from fans who stopped at active; those are two different failures.
- Check who is still inside a 30-day access window and give them a reason to open it.
- Group the fans you intend to win back, and keep the group small enough to write to properly.
- Read the cancellation reasons before rewriting an offer, since the fixed set is what makes them comparable.
- Mark a relationship as churned yourself when you have decided it, rather than waiting for a rule to do it.
Questions about retention
Will Velvet tell me which fans are about to leave?
It gives you the signals and stops there. A prediction dressed as a fact is worse than no prediction, and the difference between a quiet fan and a lost one is usually something only the creator knows.
Does a cancelled subscriber lose access immediately?
No. The renewal is switched off and access runs to the end of the period they already paid for, which is also the window in which a win-back message can still land well.
Can automation move a fan backwards?
No. Automated requalification only moves fans forward, and it never touches a stage a person set by hand.
Is a 30-day access window a retention feature?
In practice, yes. A buyer with something still open has a reason to return to the page, and the page is where the next offer lives.
How do I win back a fan without discounting?
Use the signal that explains the gap. A fan who stopped at qualified never had a reason to buy; a fan who stopped after buying had one and it ran out. Those need different messages, not a smaller price.
Read one week of signals
Open a creator account, let a week of conversations accumulate, then sort by last activity and see how much the list already tells you.