Winback

There used to be a stretch where churned customers sat on a list and nothing happened to them. When the quarterly target counts new contracts only, the 80 accounts that left last year appear on nobody's list, because they've been filed as a number already gone.

Those 80 are in better shape than a stranger, though. They've used the product, so nothing needs explaining, and they know for themselves what worked and what fell short. A new account means starting over: getting them to hear about you, getting them to trust you, getting a budget approved. Here, most of that has already happened once.

What makes the difference is having the reason on record. If a customer left because permissions couldn't be split by department, and that capability shipped six months later, there is exactly one thing to say to them. It's a different kind of outreach from listing features at somebody new. The letter more or less writes itself, which is the opposite of the usual problem with cold outreach.

Winback treats that situation as a campaign. It reads the churned list as a list of qualified candidates rather than as a closed ledger.

It starts with choosing the list, and not sending to everyone is what decides whether it works. Six steps.

1. Build the list. Attach three things to each account: when they left, how much they were paying, and why they left. If the churn reason field is empty the campaign has nothing to say, and the work becomes building a process for recording reasons instead. That sounds like a detour, and it's the difference between a campaign and a mass email.
2. Keep only reversible reasons. Drop the ones whose company folded or whose team dissolved. What stays are accounts that left over a missing capability that now exists, or over price when the plans have since changed. Going from 80 accounts to 26 is normal.
3. Write down what changed in the meantime. This is a message about what's different, not a message asking after someone. "Please come back" and "the permission setting you asked for shipped" read as completely different letters.
4. Decide who sends and through what. Larger accounts get a person reaching out directly; smaller ones get one email. The staffed route only works at low volume, which sets the list size again.
5. Lower the doorway back in. If the account and its data are still there, make them resumable. When returning means redoing the migration from scratch, willingness isn't enough. Keep discounts as the last resort. A discount is the fastest way to get a reply and the most expensive way to get a customer, because the price it sets tends to stay set.
6. Send once and stop. No reply means the next opportunity. Send the same person a third time and you lose the ability to send at all. No version of this campaign gets better with volume.

Sales handles the accounts that were paying the most and a marketer handles the rest by email. The product side supplies what changed, written as sentences. Without that list, step 3 has no content and what goes out is a letter asking after someone. Inside lifecycle marketing, which decides what to say based on the stage a customer is in, this campaign owns the last stage.

Timing turns on three things.

1. Not right after they leave. The complaint is still fresh and no sentence reads well. Three months is a floor.
2. The best moment is the day the reason they left got resolved. That's an event rather than a date on a calendar, which means the sorting in step 2 also sets the send date. Ship the fix and then wait for the next campaign calendar, and the only moment when the message is obviously true has passed.
3. For customers on annual contracts, two months before their current tool renews. That's the only window where looking at something else is realistic.

Two cases argue against sending. When the reason was something on the customer's side, and when the last conversation ended badly. The second needs more time, and reaching out brings that impression back with it.

Results get read as reply rate and actual contracts signed. If 26 went out and 3 came back, count what those 3 started paying again before counting the reply rate. Set that against the cost of acquiring one new account and it becomes clear whether to keep running this. Reply rate on its own flatters the campaign, since a polite no counts as a reply.

"Of the 80 that left last year, I kept the 26 whose reason was a missing feature."

Nobody sent to everyone; only the reversible ones stayed on the list.

"Skip the come-back-to-us line. Just write what changed since."

Reach out on the strength of a fact rather than a request.

"They left a month ago. Not now."

Too early and the same sentence doesn't land.

At Pagedeck, Sofia reviews this list every quarter. Of 80 accounts lost last year, 26 had left over department-level permissions, and she contacted only those 26 after the capability shipped. The other 54 got nothing.

The content came down to one line Mia supplied: the permission setting you asked for is in, and your existing material is still there. Ravi rewrote the same line as the email for smaller accounts, then sorted the replies from both routes by reason again. Three of the 26 came back, and two of them at a larger size than when they left, because the permission work only mattered to accounts that had grown.

The most common mistake is treating winback as a discount exercise. Cutting the price feels like it should bring people back, and it does lift reply rate. But a customer who returns on a discount is anchored to that price. Try to move them back to list a year later and the same account leaves for the same reason, and now they've left twice, which makes a third approach very hard. For a customer who left over price, changing the plan is the answer. For one who left over a missing feature, a discount answers a question they didn't ask — and teaches the account that leaving is how you negotiate.

The second is assuming that sending to everyone who left is free upside. The list already exists, so more sends should mean more returns. In practice, mixing in irreversible reasons pushes the reply rate down, and mail repeatedly sent to uninterested people gets classified as spam, which damages delivery for the messages that come after. Shrinking the list is the real work of this campaign.