Save Offer

While the cancel button was simply the end, nobody separated the reasons people pressed it. Press it, the account closes, next month's invoice stops. With no way to keep anyone, there was no point in asking why.

What Ravi heard from twenty-eight teams that cancelled and signed up again two months later shows what that cost. More than half had not left because they disliked the product. A project had wrapped, there was nothing to use it for over those two months, and paying for an idle account felt wasteful, so they closed it. Given the option to pause billing for a month, none of them would have cancelled at all.

The save offer exists to sort those customers out from the rest. At the final step of cancelling, one more set of terms goes on the table, which separates people who are done with the product from people for whom the current terms stopped working. The second group stays when the terms change.

It sits at the last step of the cancellation flow. After the customer picks a reason, the next screen shows one card matched to that reason. Four cards cover most cases.

1. Temporary discount — half price for two months. This is the card for "too expensive".
2. Billing pause — one or two months with no charge and the account left intact. This is the card for "nothing to use it for right now". Nothing gets discounted, so it costs the company the least of the four.
3. Plan downgrade — moving from a team plan to an individual one, for a team that shrank and no longer wants per-seat pricing.
4. Talk to a person — a 15-minute call on the calendar. This is the card for "I couldn't figure it out" or "it's missing something we need", where a discount does nothing at all.

Cards are split by reason because the wrong card accomplishes nothing. Half price offered to someone leaving over a missing feature buys half-priced dissatisfaction and a second cancellation two months later. Which reasons actually dominate comes from the answers piling up in the cancellation survey.

How much a card is allowed to contain is set by the company's discount policy. Without a written ceiling, different terms go out from different people, and customers on the same plan end up paying different amounts.

The arithmetic compares what was given up against what the retained customer will pay. Give a team paying $30 a month two months at half price and the company forgoes $30; if the team stays ten more months, $270 comes in. The math looks clean, but some of the people who saw the offer would have changed their minds without it. So the real effect gets measured by showing the offer to only half of the traffic and comparing how long each group stays.

There is a limit on frequency too. If the same account presses cancel every six months to collect half price, its price is effectively half. One offer per account per twelve months is the usual rule.

The most common place is the cancellation screen of a small self-serve account. Nobody can staff thousands of teams individually, so the screen puts the terms up instead. One step earlier, a reactivation campaign reaches out to teams that have gone quiet; this card is for the ones who reached the cancellation screen anyway. Reverse that order and discounts go to people who were not leaving.

For enterprise customers on annual contracts, a person does this work rather than a screen. Ahead of renewal, Sofia lowers next year's increase, drops a few seats, or adds a training session. The same idea, with amounts dozens of times larger.

It also shows up ahead of a pricing rework. When more than 20 percent of people on the cancellation screen choose to move down to the individual plan three months running, there is a gap between the two plans. A save offer patches that gap temporarily, so a run like that should hand off to a pricing change.

Two numbers report the result: the share of people who saw the offer and cancelled their cancellation, and the share of those who are still around three months later. On the first number alone, offering half price for every reason wins. Read alongside the second, matching the card to the reason wins.

"Thirty-four teams took the billing pause this month, and 29 of them came back the month after."

A save that cost the company nothing, because no price was cut.

"Don't put half price in front of someone leaving over a missing feature. Route that one to a call."

A card that doesn't match the reason won't work.

"Save rate is 22 percent and I can't tell how much of that is the offer. Let's show it to half."

A suspicion that people who would have stayed anyway are mixed into the number.

At Pagedeck one card appears at the end of the cancellation screen. Of 210 teams that pressed cancel last month, 46 accepted an offer: 31 took the billing pause, 12 took the plan downgrade, and 3 took two months at half price.

What Grace asked about was not the save rate but whether the 3 who took half price were still there three months on. If two of them cancelled in the month the discount ran out, that card did not prevent a cancellation — it moved one two months down the calendar. The answer in that case is not a deeper discount. A team whose budget was cut needs a cheaper plan it can stay on, not eight weeks of half price.

The most common misreading is that a higher save rate means better work, since the count of customers kept is what gets filed as the result. But some of the people who saw the offer would have stayed without it. If 30 of the 210 teams on the cancellation screen were going to change their minds anyway and all 210 get half price, the save rate reads 22 percent while the company has discounted 30 customers it was keeping for free.

The other one is treating the discount as the strongest card, because cutting the price feels like the biggest concession available. Retention by reason usually puts the billing pause ahead of it. To someone with nothing to use the product for, half price is an offer to buy idleness at a discount, while a pause is an invitation to come back when the work returns. Once reaching for the discount first becomes the habit, the same customer stands on the same screen two months later and the company has to find a card stronger than half price.