Dunning

Subscription billing — charging a fixed amount every cycle automatically — created a gap that didn't exist when people paid by hand. The customer has no intention of leaving, and the money still doesn't arrive, because something happened on the card side. The limit was hit that day. The expiration date passed without anyone noticing. A replacement card arrived after a loss report and never got entered.

Left alone, none of that looks different from a cancellation on the company's side. The charge failed, so the account gets locked, and the customer finds out the next time they log in and hit a wall. Being thrown out of a service you were happily using, with no warning, makes some people shop for a competitor instead of re-entering a card.

At scale the gap reaches a visible size. A company at $75,000 in monthly recurring revenue (MRR) — the revenue that comes in every month on repeat — with a 5 percent payment failure rate has $3,750 stuck every month. A large share of that would be solved the same day if the customer simply knew, and it's lost only because there's no channel telling them.

Dunning manufactures those few days. Instead of converting a failure straight into a cancellation, it retries on a set rhythm and uses the interval to reach the customer and give them time to fix the card.

Start by sorting failures by reason. Response codes from the card network fall into two broad groups: reasons that may clear in a few days, like a credit limit hit or a temporary decline, and reasons that will never clear on the same card, like a loss report or a closed account. The first group is worth retrying. The second returns the same answer however many times you ask.

One loop runs in five steps.

1. Send the first notice the moment it fails. Put what happened in the subject line and one link to update the card in the body. Close to half of everything you'll ever recover gets recovered here.
2. Retry on spaced intervals. Something like day 3, day 7, and day 14 is typical. Firing four attempts the day after the failure barely moves the recovery rate and accumulates as suspicious traffic on the card network's side.
3. Choose the day, not just the gap. When the reason is a credit limit, attempts landing just after common payday dates succeed noticeably more often. The same three attempts perform differently depending on when they're made.
4. Keep the service open while attempts run. This is what a grace period is for — the window where the product keeps working despite a failed payment. Seven to fourteen days is common, and the messaging gets more specific as the window closes.
5. Wind down once the last attempt fails. Rather than deleting the account, many companies drop it to read-only. A customer who comes back needs to find their old documents intact for paying again to make any sense.

Email alone isn't the whole channel set. An in-product banner reaches anyone who logs in, which beats email on recovery rate, and services with an app usually mix in one push notification. Hitting the same person through all three at once, though, reads as debt collection.

It fits card-based subscriptions. The company can re-initiate the charge whenever it wants, and each attempt resolves immediately, so a schedule is possible in the first place. It transfers poorly to enterprise customers paying by bank transfer. There's no attempt to re-fire there, and what's owed sits in accounts receivable — money invoiced but not yet collected — bucketed at 30, 60 and 90 days, with someone contacting the customer's finance team directly.

Some share never recovers, and that's worth planning for. When the reason is a cancelled or lost card, the same card is finished. Recovery then depends entirely on getting new card details, so adding channels that reach the customer does more than adding attempts.

Running it too aggressively backfires. An account being charged ten times a month gets flagged as risk by the payment gateway, the service that routes card charges on the company's behalf, and a confused customer who disputes the charge with their bank turns what was a simple failure into a chargeback. Chargeback rates above a threshold raise processing fees, which puts a hard ceiling on how many retries make sense.

"We recovered 62 percent of last month's failed charges through dunning."

Reporting how much the process actually brought back.

"This one's a closed card, so retries are pointless. Just send the re-enrollment notice."

The failure reason is in the unrecoverable group, so outreach replaces retries.

"Stretching the grace period to 14 days lifts recovery, but that's 14 free days."

Naming the cost that comes attached to a longer window.

At Pagedeck, Ravi owns this. Collecting money sounds like finance work, but what actually gets edited is the subject line, the banner copy, and the send time, which makes it closer to growth work. Sofia takes the enterprise failures separately and calls those account owners herself.

The change Ravi made last quarter was one line of copy. Subject lines that opened with "Your payment failed" became "Your card has expired," leading with the reason, and the share resolved on the first email rose noticeably. The action being asked for was already visible before anyone opened the message.

People often read dunning as chasing down a debt, which the word's own history encourages. But the person on the other end isn't refusing to pay. They don't know anything is wrong with their card. Written in a collections tone, it makes someone who did nothing wrong feel accused, and that impression pushes them toward the cancel button instead of the update-card form. The register should sit closer to an incident notice.

The second mistake is treating it as something you switch on once. It runs automatically, so it's easy to assume the configuration does the work from there. What actually separates a strong recovery rate from a weak one is the copy, the send times, and the length of the grace period, far more than the number of attempts anyone configured. Leave those untouched for six months and fixable failures quietly leak into churn month after month.