win-back

A win-back and a follow-up look like the same gesture and are not. A follow-up happens inside a conversation that is still breathing: he replied recently, the thread has a subject, somebody reopens it at the right moment. A win-back starts from nothing. The subscription lapsed weeks ago, the last message is old, and whatever you write arrives with no context. That difference decides everything: the timing, and whether the message is allowed to sell.

How is a win-back different from a follow-up?

By how cold the thread is, and by whether the message is allowed to carry a price.

	Follow-up	Win-back
State of the thread	Warm, recent reply	Cold, weeks or months old
Delay before sending	Hours to days	Weeks
Goal of the message	Move the conversation on	Get any reply at all
Can it carry an offer	Yes, if the timing is right	No
Who it targets	An active fan	An expired fan or a silent buyer

The row people get wrong is the one about the offer. A win-back that arrives with a price attached is not a win-back, it is a cold pitch to someone who already stopped paying once.

When do you send one?

Later than feels natural, and never the day after the subscription ends: that message reads as an invoice. Weeks is the right order of magnitude, long enough that the lapse is no longer the subject.

Our data, drawn from 42,852 conversations and 1.5 million messages, shows conversion collapsing between 2am and 6am and peaking in the evening. A win-back sent overnight to clear a queue is a wasted message: this is not something you get to send twice.

What goes in the message?

One idea, phrased so that answering is easier than ignoring:

• A personal callback. Something he actually told you months ago. It is the only thing in the message that proves you are not sending it to a list. Our data shows it costs points when it lands at the exact moment of a pitch, which is the most counter-intuitive result in the corpus and the reason a win-back must stay separate from an offer.
• An open question, not a yes/no one. A closed question asked at the moment of closing costs several points of conversion, and the same reflex kills a reopening message: “do you remember me?” can be answered by saying nothing.
• A plain ending. Ending a sales message with an ellipsis is the worst-performing closer we have measured, and they read worse still on a message that is pretending not to sell.

If he replies, the conversation restarts from scratch, not from where it stopped. Pitching before the sixth message drops conversion by roughly a third, and a fan who just came back is the last person to rush.

Why do win-backs fail?

Four reasons, and three of them are impatience:

1. A discount as the opening line. It teaches everyone that leaving is how you get a better price.
2. An identical message to the whole list. He can tell, and a bad first attempt burns the only real chance you had.
3. Repeating it weekly. Spread attempts over months, then stop.
4. No plan for the reply. A win-back that works hands you a live conversation at an unpredictable hour, and if nobody is there to answer it, the whole sequence was theatre.