Points are a liability, not a currency
Every point you issue is a promise to sell something later, at a price you fixed today, in a unit you print yourself. Most programs manage the issue price with real rigour, assert the retire price in a deck, and then report the breakage as a win.
A point is not money you handed someone. It is a promise to sell them something later, at a price you fixed today, denominated in a unit you print yourself. Accounting treats that correctly. Almost nobody else in the building does.
Under IFRS 15 and ASC 606, points awarded alongside a purchase are generally a material right and a separate performance obligation. Part of the transaction price gets allocated to them and sits on the balance sheet as a contract liability until the points are redeemed or expire. The finance team already knows the truth: this is a debt, and the unit it is denominated in has a value you set by decree.
Then the org chart takes over. The liability sits with finance. The exchange rate sits with marketing. The thing that actually retires the liability, which is a trip somebody takes, sits with a partnerships team who negotiated it as a feature launch. Three owners, one price, and nobody accountable for the spread between them.
Two prices, and only one of them gets managed
Every point has two prices.
The issue price is what it costs to create. Card programs model this carefully against interchange, fees and funding cost, because issuance looks like acquisition spend and lands next to a campaign in the plan. Retailers model it as a discount. Whatever the mechanism, the issue side gets analysts.
The retire price is what it costs you to make a point go away with the customer feeling good about it. That number is rarely computed per redemption. It is asserted: a cent a point, a cent and a quarter, whatever the last deck said.
The gap between those two prices is the entire economics of the program. Most programs actively manage one end of it and treat the other as a constant of nature.
A fixed exchange rate is a standing offer to be selected against
Take the common design: points redeem at a fixed rate against the retail price of whatever the customer picks. It sounds neutral and fair. It is neither, because it is not a price. It is an option you wrote, handed out for free, and never repriced. The strike is frozen. The underlying is not.
Travel cash prices move constantly with season, demand, fare class, length of stay and day of week. Your redemption rate does not move at all. So the real value of that option swings continuously, in a direction your customers can observe and you cannot control. Every redemption forum on the internet is a distributed optimiser searching for the exact moments when your fixed rate is most generous relative to cash.
That is adverse selection, and it is not a defect in customer behaviour. You published the arbitrage. By construction, the redemptions you receive are skewed toward the cases where your rate is worst for you. A program whose cost per retired point keeps drifting above plan has not been unlucky. It has been correctly exploited by people reading the terms you wrote.
The instinctive fix is a worse rate. That is the wrong move twice over: it punishes the majority who were never optimising, and it turns the points into a coupon. The real fix is to let the rate carry information, which means pricing redemption against what a redemption actually costs you to retire.
Breakage is a failure that gets reported as income
Expected breakage is recognised as revenue in proportion to the pattern of redemption, and there is nothing improper about that. It is real, it is audited, and it is often material.
It is also the program telling you it failed, in the one language the P&L rewards.
Three things go wrong when breakage becomes a load-bearing line. It is negatively correlated with the engagement the program exists to create, so the better the program works the worse that line performs. It is fragile, because expiry rules are exactly the sort of thing regulators and competitors change, and when they move the earnings vanish in a quarter. And it hides the diagnostic underneath: points expire because spending them was not worth the effort, and that is a product finding, not a windfall.
Why travel is the exit that is genuinely different
There are only three ways a point leaves your balance sheet, and they are not variations on a theme.
Cash-like redemptions are the honest baseline. A statement credit or a gift card costs you roughly the face value you promised, so the spread is near zero by design. You have converted a loyalty program into a rebate with extra steps and a mobile app.
Travel is structurally different, and the reason has nothing to do with wanderlust. Travel is bought at net rates and perceived at retail. A hotel night procured on a negotiated rate and presented against the public price delivers value the customer can verify on another tab of their browser, at a cost to you that is materially lower. That gap is the margin. It is the only exit where the number leaving the balance sheet and the number leaving the bank account are genuinely different quantities.
Breadth is what makes that mechanic hold at scale, and it is the part programs consistently underestimate. A redemption catalogue that covers a shortlist of cities converts a thin slice of the outstanding balance and leaves the rest to expire. Depth of supply is what turns "points are useful" from a slogan into a fact for the customer flying to a second-tier city in October: 2M+ properties, 517+ carriers including 140+ low-cost operators that never appear in a GDS, across 190+ countries. Coverage is not a vanity number here. It is the redemption surface, and it sets the ceiling on how much liability you can retire well.
Four ways programs price this wrong
| What the program does | Why it looks safe | What it actually is |
|---|---|---|
| One exchange rate across every product | Simple to explain and easy to market | A written option with a fixed strike on a floating price |
| Prices redemption off the retail screen price | It matches what the customer sees | A margin that moves with someone else's discounting calendar |
| Treats expected breakage as program health | Recognised, audited, genuinely revenue | A measurement of how little the points were worth using |
| Sources travel through one retail channel | One contract, fast to launch | No visibility of cost to retire, so no way to price the spread |
The fourth row is the one that surprises people, because it looks like an integration choice. It is not. If your travel supply arrives as a bundled retail price from a single reseller, you cannot compute cost to retire per redemption, which means you cannot price the retire side at all. You are running a treasury function blind. That is why redemption sourcing belongs to whoever owns the program's margin, not to whoever owns the app screen. The same argument, applied to the wider embedded travel decision, sits in the tab is not the product.
Measure the exit, not the redemption rate
Redemption rate is a participation statistic. It tells you that points left, not what they cost. Four numbers do the actual work.
- Cost to retire per point, by product. Real cash out, per redemption, split by flights, hotels and everything else. If you cannot produce this, that is the finding.
- Liability velocity by exit. How fast issued points reach an exit, and through which of the three doors. A program with slow velocity and rising breakage is not stable. It is accumulating an obligation and calling the delay a profit.
- Spread per retired point. Fair value released, less cash cost, less servicing cost. The last term is the one everyone omits, and it is the subject of the next section.
- Selection skew. Compare the realised value per point achieved by your heaviest redeemers against the median. A wide gap is the market pricing your option for you.
The cost travel adds back
Here is the part that belongs in the same paragraph as the margin, because leaving it out is how programs book a good number in month one and discover the real one in month six.
A gift card never gets rescheduled. A redeemed trip is fulfilled weeks later by third parties who did not sign your brand guidelines, and every schedule change, cancellation and refund in it arrives at your support desk wearing your logo. Travel redemption converts a financial liability into an operating one. The spread is real, and it is net of an operation you have to actually run: order state, disruption handling, refunds that trace back to the points they consumed. What that operation has to look like is the whole argument in post-booking is the trust product, and the strategic version of the question, whether travel belongs in the product at all, is in the travel readiness framework.
Programs that skip this step do not fail at the accounting. They fail at the second disruption, when the customer who spent years of accumulated balance on an anniversary trip discovers that the brand which took the points has no view of the booking.
Where this leaves the program
Points that never get spent are a debt with a marketing budget attached. Points spent on a rebate are a discount you made complicated. Points spent on a trip are the only version where the customer receives something they can describe to another person and the program keeps a spread it can defend in a board meeting.
So price the retire side like a treasury function rather than a brochure. Vary the rate by what redemption costs. Buy supply where you can see the net rate. Count servicing as cost of goods sold, because it is. And stop reading breakage as a result.
The question worth asking in the next program review is not what the redemption rate was. It is this: what did the last ten thousand retired points cost in cash, who chose those redemptions, and knowing what that selection looked like, would you write the same option again this morning?
