Do the count that actually matters, not the revenue chart. How many paying customers did you add this month, and how many did you lose. If the second number is bigger than the first, your business is shrinking, and it can be shrinking while the revenue line still climbs, because the customers who stayed expanded enough to paper over the ones who left. Founders read the revenue chart and feel fine. The customer count is the one telling the truth.
Put a number on it before the quarter closes
Say you added eight new paying customers this month and lost eleven. Net, you have three fewer paying customers than when the month started, even if the eleven who left were smaller accounts than the eight who joined and the total revenue figure still ticks up slightly. Customer count and revenue can point in opposite directions for a surprisingly long stretch before the mismatch becomes impossible to explain away, and by the time it is impossible to explain away, you have lost months you could have spent fixing whatever is actually driving people out the door. Run this count monthly, on its own, separate from the revenue chart, and treat a negative number as the more urgent of the two.
Say this plainly before anything else
Churn is almost always a signal about the product or about who bought it, not about the pipeline that brought them in. If people are leaving broadly, across segments, for a mix of reasons that do not cluster, that is a product and support problem, and the fix lives there: what breaks, what is missing, what the onboarding fails to explain. No acquisition tool, including the one described further down, treats that kind of churn. Selling harder into a leaking bucket produces a bigger leak, faster.
Two reasons a customer leaves, and they are not the same reason
The first reason is that they never got the value they were promised. That is squarely a product issue: something did not work as expected, or the setup was too heavy for what they needed, or a competing priority made the product easy to drop. Fix it by fixing the product, and by talking to the people it happened to about specifically where it fell apart for them.
The second reason is different and gets misdiagnosed constantly: the customer got exactly what the product does, used it correctly, and still was never going to stay, because they were never the right buyer in the first place. This one is not a product failure. It is a targeting failure that happened months earlier, at the moment someone was closed who fit the deal you needed that week rather than the customer your product is actually built for.
The mistake that shows up downstream as a retention number
Picture two versions of the same quarter. In one, outbound goes to anyone who might plausibly say yes, and ten deals close. In the other, outbound goes only to a tightly defined segment: a specific size, a specific role, a specific situation that makes your product actually solve something they have. Fewer deals close, but the customers who do close are the ones the product was built to keep. Three months later the first group is churning hard and the second is barely moving, and the founder running the first group is looking at a churn dashboard trying to find a retention fix for a targeting problem.
If your churn concentrates in one kind of customer, a specific size band, a specific use case, a specific way they found you, that concentration is the diagnosis. The fix is narrowing who gets contacted going forward, not building a save campaign for the ones already walking out the door.
The list nobody works: people who already left
Here is the one place this genuinely becomes an outreach question, and it is worth separating clearly from everything above. A customer who churned six months ago is not a stranger. You know their name, what they used, roughly why they left if anyone asked at the time, and whether the reason still applies. Most founders never look at that list again, because the account is already gone and closed lists feel closed. Some of those reasons stop applying: a missing feature ships, a price changes, the account grows into a size where the product fits better than it did the first time.
Writing back to that list is real outbound aimed at known people instead of cold names, and it is the one part of a churn problem that a tool built for finding and reaching buyers can honestly help with. Rocketship's worker can be pointed at a list of lapsed accounts instead of a fresh market, write to each one personally from your own inbox referencing why the timing might be different now, and read what comes back so a cautious reply gets handled carefully rather than dropped into a generic sequence. It will not tell you which churned accounts are worth the approach. That judgment, which reasons for leaving were fixable and which were not, is still entirely yours to make before anyone gets written to.
Sort the list before you write to anyone
Not every departed account belongs on this list. Split them into two piles first: the ones who left because of something you can now honestly say is different, and the ones who left because your product was simply never going to be a fit for how they work. Writing to the second pile wastes their patience and yours. Writing to the first pile, with a specific, true reason attached to the message, is one of the highest-converting things a founder can send, because you already know their exact problem and they already know you exist. Skipping the sort and writing to the whole list at once is how a legitimately good move starts to feel, and read, like spam.
What not to buy
If churn is broad, spread evenly across the kind of customer you have, and not explained by anything specific, do not respond by buying a sales or outreach product. That spend treats a symptom on the wrong side of the business. Fix the product, fix onboarding, or fix who you are targeting going forward, and only then look at the list of people already gone as a small, separate, legitimate thing worth one more conversation.