Win Back Email Campaigns

There is a version of this campaign that every SaaS company runs and almost none of them should. It emails every cancelled account on a 90 day timer, says we have made a lot of improvements, and offers 20 percent off. It reactivates almost nobody and it burns the list. The version that works starts somewhere else entirely: with a database of why people left, and a product team that fixed at least one of the reasons.

Win back only works when you fixed the reason

A churned admin opens your email with one question in mind. Is anything different, or is this the same tool that did not work for us. If you cannot answer that with a specific shipped change, you have nothing to say and the email is an interruption.

This is the part most teams skip. They build the sequence, write the copy, argue about the discount, and never connect any of it to the actual reasons the accounts left. So the email says we have shipped lots of updates, which every churned customer correctly reads as we have not shipped the one you needed.

The prerequisite, then, is a reason database. Every cancellation gets a reason recorded in a structured field, from the cancel flow where you have one, from the CSM where an account has coverage, and from the reason ask email described below. Without that field you cannot segment, and without segmentation this campaign is a drip campaign fired at people who already told you no.

Which churned accounts are worth contacting at all

Five reasons account for most SaaS churn, and only two or three of them are worth a sequence. Being honest about this halves the size of your win back list and roughly doubles its performance.

The champion departure case is worth treating separately because it is not really a win back at all. The account did not reject you, the person who understood you left. That is a new sale to a new person at a familiar company, and it should be routed to sales with the historical usage data attached, not into a marketing sequence. If you already track champion departure as a risk trigger in your churn prevention email campaigns, you have the data to spot this the day it happens.

Accounts that never activated in the first place are the cheapest segment to attempt and the least likely to return. They churned because they never got value, they have no muscle memory of the product working, and nothing about a new email changes that. Attempt them once, with a different offer, and then stop. Fixing the front of the funnel with better activation email sequences returns more than any amount of chasing this segment.

The four email sequence, at 30, 60, 120 and 180 days

Four sends over six months. The spacing is deliberate: it lets a replacement tool fail on its own timeline, which is when most genuine reactivations actually happen.

Two of those four sends go only to a subset. The changelog email needs a matching shipped feature. The migration email needs you to know, or reasonably guess, where they went. Sending all four to everyone is how a well designed sequence turns back into the mailshot.

Email one, the reason ask, from a named human

This is the most valuable email in the sequence and the one that generates no revenue. Its job is to fill the reason database.

Send it from the person who ran the account, or from a founder at smaller companies, using a real personal address. Plain text, no logo, no template, no tracking pixel if your platform lets you disable it. One question. Something close to: you cancelled Acme last month and I would like to know what pushed it over the line, was it the missing Salesforce sync or something else. Naming a likely reason gets more replies than an open question, because it is easier to correct a wrong guess than to compose an explanation.

Expect reply rates well above anything else you send, often several times your normal lifecycle numbers, because the question is genuine and the person has an opinion. Read every reply. This is the highest quality product feedback your company will receive all quarter, and it arrives free.

The changelog email that names the shipped fix

Email two is the one that actually reactivates accounts, and it only exists if engineering shipped something relevant. The structure is short and concrete.

Name the reason they gave. Name the thing you shipped. Give the date. Link to the changelog entry or the docs page, not to the homepage. Offer a way back that does not require a sales call. Something like: you told us in February that the lack of a two way HubSpot sync was the blocker, we shipped it on 4 August, here is the setup doc, and your workspace is still intact if you want to try it.

Keep the discount out of this one. The moment you attach a commercial offer, the email reads as a sales push wearing a product update costume, and the specific detail that made it credible stops doing its job. Save the offer for email four, and consider whether you need it at all.

Migration help beats a discount, almost always

The barrier to returning is not your monthly price. It is that somebody spent three weeks migrating to a competitor, retrained a team, rebuilt reports and reconnected integrations, and coming back means doing all of it again in reverse while explaining the decision to a manager.

So the third email offers to do that work. Free data import, including from named competitors where you have built importers. A person assigned to the migration. Reconnection of the integrations you can see they had, because you still have the record. A restored workspace with historical data intact rather than a blank account.

That offer costs you a few hours of a support engineer’s time and it addresses the real objection. A 20 percent discount costs you margin forever and addresses an objection that, for most accounts, was never the primary one. Where the reason genuinely was price, the right answer is usually a structural one: a smaller tier, usage based pricing, or an annual commitment at a lower monthly equivalent, and that conversation belongs with a human.

Who to suppress, permanently

A win back list is the most dangerous list you own for deliverability. The addresses are older, the people did not opt back in, and the emotional baseline is negative. Spam complaints from churned users hurt the inbox placement of every other email you send, including the dunning email sequences that recover real revenue from active customers.

Suppress permanently, and record the reason so nobody re adds them in a list rebuild:

• Accounts that churned in arrears after a failed payment chase, because they already associate your name with debt collection
• Anyone with an unresolved support escalation, a chargeback, or a public negative review
• Companies that were acquired, wound down, or whose domain no longer resolves
• Anyone who asked for deletion, unsubscribed, or exercised a data request
• Addresses that hard bounced once, with no retry, ever

Run a bounce and validation pass before every send, because a list of 18 month old churned contacts will typically carry a double digit percentage of dead addresses. A list hygiene step costs a few hundred dollars and protects the sending domain your revenue emails depend on. The quarterly lifecycle email audit should check that these suppression rules are still enforced after every platform migration, since suppression lists are the single most common casualty of moving between tools, as anyone who has run a Customer.io vs Braze migration will tell you.

The numbers to expect, and how to report them

Set expectations before you build, because the raw numbers look bad next to any other campaign you run and someone will ask.

Reactivating 2 percent of a 900 account churned list is 18 accounts. At a $9,000 ACV that is $162,000 of recovered ARR for maybe three weeks of build and a few hours of sending each quarter. That is a defensible number, and it is a better argument than an inflated rate.

Report it as recovered ARR and as accounts returned, segmented by churn reason, and track whether returned accounts stay. A reactivated account that churns again within two renewal periods was not a save, it was a delay, and that pattern usually means you reactivated on a discount rather than on a fix.

Build the reason database before you write the sequence

If you do one thing from this page, add a structured churn reason field and make it required on every cancellation, whether that comes from a cancel flow, a CSM note or the reason ask email. Give it five values, not twenty. Backfill the last two quarters by hand if the volume is manageable.

Then wait until engineering ships something that maps to one of those reasons, and send the changelog email to that segment only. That single send, to a few hundred people, will tell you more about whether this programme is worth automating than a full four email build ever would. The wider context, including how the same tone works across activation and retention, sits in the SaaS email marketing cluster, and the annotated examples in win back email examples show what the four sends look like from real companies. If you are still deciding which sends your programme needs at all, start with SaaS email types.