Cancel Save-Offer Matrix: Offer by Reason

A cancel flow that shows every departing subscriber the same discount is not a save strategy. It is a coupon dispenser, and subscribers learn to use it that way. The fix is matching: read the reason the subscriber gives for leaving, and answer that reason with the one offer type that actually addresses it. This matrix maps the seven cancel reasons subscribers give most often to the best first save offer, the message angle that makes the offer land, and — just as important — the situations where you should not make an offer at all.

Swipe Again's cancel deflection works with three canonical offer types: Discount, Pause, and Service. Every row in this matrix resolves to one of those three, because a save offer that cannot be fulfilled by a real lever is just copy. For the mechanics of how a deflection flow is built and measured, see cancel deflection and what cancel deflection is. This page is the narrower tool: which offer, for which reason, said which way — and when to stand down.

The three offers you actually have

• Discount — a price concession: money off or a reduced rate for a defined stretch. Discount is the strongest lever and the most abused one. It answers exactly one problem: the price is the obstacle. Used on any other reason, it trains subscribers to complain their way to a cheaper plan and does nothing about why they were leaving.
• Pause — the subscription stops, billing stops, and the account, history, and settings wait intact for a defined return. Pause answers timing problems: not using it right now, life getting in the way, a season ending. It preserves the relationship at zero revenue today instead of ending it permanently.
• Service — a human or hands-on fix: support outreach, onboarding help, a setup call, account adjustments, or a person who owns the subscriber's problem until it is resolved. Service answers capability problems — the product is not working for them, they never got it set up, something is broken — where the subscriber needs competence, not a coupon.

One rule governs the whole matrix: one offer, matched to the stated reason, presented once. Stacking offers — discount plus pause plus a service call on the same screen — reads as desperation, devalues every lever, and makes your save data unreadable, because you will never know which offer did the work.

The cancel save-offer matrix

Cancel reason	Best save offer	Message angle	When NOT to offer
Too expensive	Discount	Name the value they already got, then make the price survivable: a reduced rate for a defined period, stated plainly, with the regular price and the return date visible.	Do not offer if they have already taken a save discount recently, or if they are closing the card or account entirely — a cheaper price does not fix a payment method that is going away.
Not using it	Pause	Zero usage is a timing and habit problem, not a price problem. Offer to stop billing until a date they pick, with everything saved, and offer Service help getting value on return if they never got started.	Do not offer a Discount to a non-user — a cheaper unused subscription is still unused. Do not Pause if they say they will never use it again; that is a real cancel.
Missing feature	Service	First check whether the feature exists and they missed it — that is a Service save: show them, set it up, walk them through it. If it genuinely does not exist, say so honestly, and offer a Pause only against a real, dated roadmap item.	Never promise a feature that is not committed and dated. A save bought with a fictional roadmap converts one clean cancel into one angry cancel plus a refund request later.
Switching to a competitor	Discount	Compete on the total picture first: what they have already built, stored, or configured with you, and what switching costs them in setup and history. A short Discount can bridge a genuine price gap while they compare.	Do not offer if they name a capability you truly do not have and need — you will lose them in a month anyway, after paying a discount for the privilege. Let that cancel go cleanly and note the reason.
Life circumstance	Pause	Job loss, illness, a move, a new baby, a tight season: the subscriber is not rejecting the product, their life is full. Pause with a generous, clearly stated return date and no billing in between. Lead with empathy, keep it short.	Do not Discount someone in a hard circumstance as the first move — it reads as tone-deaf haggling. Do not offer anything if they ask you not to; in a hardship cancel, the fastest, kindest exit is what brings people back later.
Technical problems	Service	Acknowledge the specific failure, name a person or team who will own it, and give a concrete next step and timeframe. If the problem burned a paid period, pair the fix with a Service credit for the lost time.	Do not offer a Discount instead of a fix — discounting a broken experience pays the subscriber to keep suffering it. Do not offer Service you cannot deliver; an unfulfilled fix promise is worse than the bug.
Just browsing / pausing	Pause	Some subscribers click cancel to see what happens, or genuinely only want a break. Make Pause the obvious, one-click answer: stop now, resume on the date they choose, nothing lost. Keep the tone light — this subscriber is not angry.	Do not lead with a Discount here. This is the most expensive habit in cancel flows: paying a price concession to subscribers who were only ever asking for a pause.

Reason by reason: the rows operators get wrong

Not using it → Pause

Split this reason before offering: never got started, or used to use it and drifted? Never-started is a Service problem wearing a Pause costume — they need onboarding, and a pause that ends with the same confusion ends in a cancel. Drifted is a true Pause case: habits lapsed, the season changed, the project ended. Either way, do not reach for Discount. Price was never the complaint, and cutting it leaves no usage exactly where you found it.

Missing feature → Service

If the capability exists, the save is education: a Service touch that shows the subscriber where it lives or turns it on for them. If it does not exist, honesty is the whole play. A Pause is legitimate only against a genuinely committed feature with a real timeframe you can state in writing. Vague roadmap language buys a few weeks and spends trust you will want if that subscriber ever returns. For the reasons behind cancel clicks in depth, see why subscribers churn.

Life circumstance → Pause, with a light touch

This row is where tone does the saving. The subscriber's problem is real, temporary, and not about you. A Pause that stops billing cleanly, preserves everything, and sets a realistic return date answers the actual situation. Keep the offer short and human. And if a subscriber in a hard moment asks to just cancel, complete it graciously and fast — the goodwill from an easy exit during a bad month is what brings people back when life settles.

Just browsing / pausing → Pause, and do not overpay

This row looks harmless and quietly costs the most. Many cancel clicks are exploratory — the subscriber wants to see their options, check renewal terms, or simply take a break. They were never asking for money. A clean, one-click Pause with a subscriber-chosen return date resolves the whole interaction. Leading with Discount here gives away margin to subscribers who told you price was fine. Save the Discount lever for the row where price is the stated problem.

When no offer fits: the clean exit

Some cancels should complete: the subscriber is closing the business, no longer needs the category, is unable to continue, or has clearly said they want out. In those cases the matrix has one entry — a fast, dignified cancel with no obstacle course. That is not surrender; it is data hygiene and reputation at once. A subscriber who leaves easily can return; one who had to fight a flow tells people about the fight. Record the reason, complete the cancel, and keep a simple return path open.

The bottom line

Match the offer to the reason: Discount only where price is the stated obstacle, Pause where timing or circumstance is the problem, Service where capability or reliability failed. One offer, presented once, with an honest "when not to offer" rule on every row. Swipe Again's cancel deflection is built on exactly these three offer types and is included in the same $179/mo flat price as its failed-payment recovery, with a 14-day trial and the guarantee in its exact form: "Recover $179 in 30 days, guaranteed — or Swipe Again is free the next 30 days." The matrix works whether or not you use our software — audit your flow against it quarterly. Offers drift; reasons do not.

Frequently asked questions

What are the three save offer types in a cancel flow?

The three canonical save offers are Discount, a defined price concession; Pause, a billing stop with the account and its history preserved for a chosen return date; and Service, a hands-on fix such as support outreach, onboarding help, or a named person owning the subscriber's problem. Every cancel reason in this matrix maps to one of these three.

Why not offer a discount to every cancelling subscriber?

Because a discount only answers a price problem. Offered for other reasons, it does not fix why the subscriber is leaving, it trains subscribers to click cancel to get a cheaper rate, and it spends your strongest lever where a pause or a service fix would have worked for free. The most expensive example is the subscriber who only wanted a pause and got paid a discount instead.

What should I do when a subscriber cancels because of a missing feature?

First check whether the feature exists and the subscriber missed it — if so, a Service save that shows or enables it is the right offer. If the feature genuinely does not exist, be honest. Offer a Pause only against a committed feature with a real timeframe you can state in writing, and never promise an undated roadmap item as a save.

When should I not make a save offer at all?

When the subscriber's situation cannot be changed by an offer: closing the business, leaving the category entirely, a hardship where they have asked to simply cancel, or a second save attempt after a previous offer already failed. In those cases a fast, clean cancel protects the relationship and the possibility of a later return better than any offer.

How do I know if my save offers are working?

Judge a save by whether the subscriber is still active after the offer period ends, not by whether they clicked accept in the flow. A discount that delays a cancel by one cycle is a delayed cancel. Track saves by reason and offer type separately, so you can see which rows of the matrix hold subscribers and which rows are just postponing the exit.