Automated Dunning: Our First 23 Days and What We Learned

What Changed with Automated Dunning

A member's card fails on the 1st. Nobody notices. They keep training. Three weeks later, someone runs a report, sees the outstanding balance, and now you get to have an awkward conversation with somebody who's been coming in all month and thought they were paid up.

That's the version most gym owners know. And the reason it stings isn't the money on that one invoice—it's that a decent share of those people never sort it out. They stop coming instead. A failed card is where churn starts, and it looks like an accounting problem right up until it turns into an empty spot in the 6am class.

So we built automated dunning into the platform—a structured sequence that chases a failed payment on its own, on a schedule, without anyone remembering to do it. We flipped it on across our tenants in August. Here's what 23 days of production data says, including the bits that are inconvenient.

Key Improvements Noted

Recovery within 7 days went from 34.5% to 50.4%.

That's 15.9 percentage points, or a 46% relative improvement. Within 14 days, it went from 43.1% to 60.0%.

• Before: about 34 got resolved inside a week. 66 were still hanging.
• After: about 50 got resolved inside a week. 50 still hanging.

So roughly 16 more members out of every 100 failures get sorted within seven days instead of drifting. At $50/month dues, that's around $795 a month of membership that didn't quietly lapse—but the dues are the smaller half. The bigger half is 16 people who didn't start the slow fade of "I should probably fix that" into "I guess I quit."

The Problem Is Getting Bigger, Not Smaller

Here's the number that convinced us to stop putting this off. Failed invoices across our platform, by month:

Failed payments scale with your member count. They don't stay a nuisance you handle on a Sunday—they become a job. The recovery rate is the part you control, and it's the part that decides whether growth means more revenue or just more leakage.

Why the Old Way Loses

Most gym software does retry the card. Usually, a couple of automatic attempts a few days apart, and then it gives up and drops the invoice into an outstanding-balance report that someone is supposed to look at.

Two things go wrong with that.

The retry is silent. If the card is expired, or the member switched banks, or the charge tripped a fraud hold, retrying the same number changes nothing. The card isn't going to start working because you asked again. Somebody has to tell the human.

The follow-up depends on a person remembering. Gym owners are coaching at 6am, cleaning at noon, and answering DMs at 9pm. "Check the failed payments report" is a task that loses to literally everything else, every single day. It's not a discipline problem. It's that the job never should have been manual.

Automated dunning fixes the second problem, which unlocks the first—the member actually finds out, in time to do something about it, without you being the one to tell them.

What We'd Tell You Even If You Never Buy Anything From Us

This isn't proprietary. If you're on other software, most of it is still worth doing:

• Reach the member, not just the card. A retry with no message is a coin flip. A retry plus a text that says what happened and how to fix it is a different thing entirely.
• Move fast, then back off. The first few days after a failure are when you get most of what you're going to get. A sequence that front-loads contact and then eases up beats one that nags at a constant rate.
• Make fixing it take one tap. Every extra step—log in, find billing, re-enter a card—sheds people. If updating a card takes more than about thirty seconds on a phone, that's where your recovery rate is going.
• Separate "won't pay" from "can't pay right now." They need different messages. Treating both like collections costs you the second group, and the second group is bigger.
• Track recovery rate as a number you actually look at. Not the dollar amount recovered—the rate, at 7 days and at 14. Dollars go up when you grow. The rate tells you whether the process is any good.

The Part Where We Tell On Ourselves

We'd rather publish the caveats than have you find them.

23 days is not very long. This is early data. It is not a season, and it hasn't been through a January or a holiday period.

Our control group is dirty. About 38% of the "before" invoices already had some dunning activity, because some tenants got the feature early. That inflates the baseline, which means the measured lift is probably understating the real effect—but "our error runs in our favor" is exactly the sentence you should be suspicious of, so treat the 46% as directional rather than banked.

Not every gym improved. Some sites went backwards over the period. Small samples swing hard—a facility with a dozen failed invoices can move twenty points on two members—but we're not going to pretend the line went up everywhere, because it didn't.

30-day recovery isn't meaningful yet. There simply hasn't been enough time for a 30-day window to close on the post-cutover cohort. We'll know more in three months.

Where This Lands

Failed payments are the most boring churn problem in the business and one of the most expensive. Nobody opens a gym because they love chasing declined cards, which is precisely why it doesn't get done.

Early numbers say automating it works, and works meaningfully. We'll republish this in 90 days with a clean control group and a real 30-day window—including if the number comes down.

If your current software drops failed payments into a report and calls it a day, that's worth an afternoon of your attention regardless of whose software you use next. If you're weighing a move, we do the whole migration for you, payment tokens included, so nobody has to re-enter a card.