Saving an at-risk account. There is no single save play.

The tip-off

How did we know an account was going sideways? Sometimes it was the data: our daily email showing logins and usage dropping off — the same signals that fed our health scores. But more typically, it was a conversation. A CSM talking to a customer hears the temperature change before any dashboard shows it. The instrumentation buys you time; the relationship tells you the truth.

"We just aren't using it" is not a reason

Here's where the save actually starts. When an account went lukewarm, the customer would sometimes say: we just aren't using it — and offer no reason why. That's a symptom presenting itself as an explanation, and if you accept it at face value, your only remaining move is a discount. So the CSM stepped in and diagnosed, working through the questions that actually locate the problem:

Does the admin need to be retrained? Does the team? Is the configuration right for how their business actually runs? Did the data import go in clean, or are they working around a messy database? Is the tool embedded in their weekly and daily sales processes — or sitting next to them? Is the champion still there, and still on board? Is it a pricing thing? Did a competitor step in with a better offer?

That's the whole argument against the email-driven save motion, and we learned it by sending the emails. A video tutorial answers a training problem. It does nothing for a champion who left, a sales team that's quietly resisting, or a competitor undercutting you. You can't treat what you haven't diagnosed — and the diagnosis takes a human.

The menu of plays

Once you know the real reason, the play picks itself. It could be a re-training play — the admin, the team, or both, back through the product. A data-import play — a fresh, clean import so the tool reflects reality again. A configuration play — reworking the setup to match how their business actually operates. A process play — the product hasn't been integrated into their weekly sales pipeline routines, so the work is wiring it into the meetings and motions they already run. Or a change-management play: a resistant sales team that fears the tool means their every action will be watched by big brother.

And often it's a combination — several plays run together to make one save. A messy import plus an untrained admin plus a skeptical team isn't three accounts at risk; it's one account with three treatments due.

The commercial tools come last

Notice what hasn't appeared yet: discounts. That's deliberate, and it was the sequence. Once the real reasons were figured out and addressed — then you add the other tools into the mix: discounts, pauses, service concessions, roadmap commitments, doing everything possible to save the account. Our CSMs ran these conversations, and for larger-ARR accounts, usually everything was on the table. But the roadmap commitments in particular were for exceptional cases only.

The order is the insight. A discount applied to an undiagnosed account just defers the churn at a lower price point — the customer still isn't getting value; now they're just paying less to not get it. A discount applied after a retraining play and a clean re-import is a different instrument entirely: it's goodwill on top of a fixed account, and it buys the runway for the fix to take hold.

The account not worth saving

Some churn you stop fighting. For us the category was clear: customers who made unreasonable demands around our product roadmap — who thought we were their custom software vendor and could simply change the product to fit their unique needs. They were never a good fit, and never worth fighting for. They didn't understand what they bought in a SaaS product, and no save play fixes a misunderstanding about what the product is.

That's worth internalizing before your first save conversation, not after your tenth: the goal isn't a 100% save rate. It's keeping the customers who can actually succeed with what you've built.

The number we never tracked

An honest admission to close: we didn't track a save rate per se. We probably should have. The save motion ran on the success team's judgment and the weekly operating machinery — at-risk accounts went onto the IDS list with an owner and a resolution path — but the outcomes never rolled up into a number we reviewed. If I were running it again, saves attempted and saves won would sit on the scorecard next to churn, because a motion this important deserves a number that tells you whether it's working.