1. What is a loyalty card?
A loyalty card is a card issued by a business to a returning customer that identifies that customer and records their accumulated credit — stamps, points, visits or spend — toward a defined reward. A loyalty programme is the set of rules the card operates under: what earns credit, how much is needed, what the reward is and when it expires. The card is the artefact; the programme is the logic. A business can change the programme without changing the card, and with a digital card it can change both after the customer already has it.
How does a loyalty card work?
Four steps, in the same order whether the card is plastic or digital.
- Enrolment. The card is handed over at the counter, or saved to a phone by scanning a QR code.
- Identification. The customer presents it at checkout — handed over, scanned from the screen, or on some platforms tapped via NFC.
- Accrual. The business records the qualifying event and the balance rises.
- Redemption. At the threshold the reward is claimed and the counter resets, or points are spent down.
The difference that matters is what happens between steps 3 and 4. On a paper card, nothing: it sits in a wallet until the next visit. A card held in a mobile wallet pass can be updated remotely in that gap — a corrected balance, a bonus, an expiry warning — and, within platform limits, can surface that change on the lock screen.
Loyalty Card at a Glance
Not every card carries every element. A ten-stamp coffee card may hold nothing but a stamp count; a tiered retail card may hold all seven.
| Element | Purpose |
|---|---|
| Customer identifier | Links this card to a person or account so activity can be attributed. Without it you have a discount, not a loyalty programme. |
| Reward mechanism | The rule that converts behaviour into credit — one stamp per visit, one point per unit of currency, a tier boundary at a spend level. |
| Progress | The current balance, and ideally the distance still to travel. Visible progress is what makes the card motivating rather than administrative. |
| Offer or reward | What the customer is working toward, stated in terms they can picture. "A free 12oz latte" beats "a reward". |
| Barcode or QR code | Where applicable — the scannable link between the physical checkout and the digital record. |
| Expiry | Optional, and consequential. Expiry protects the liability on your balance sheet and can prompt a return visit; it also reliably annoys customers who lose credit they earned. |
| Updates | Digital only. The ability to change the card's contents after issue — balance corrections, new offers, tier upgrades — without reissuing it. |
2. The 5 main types of loyalty cards
Stamp or punch cards (buy a set number, get one free), points-based cards (earn points per unit spent, redeem against a catalogue), tiered cards (status levels that unlock better benefits), spend-based or threshold cards (a reward at a cumulative spend level), and coalition or partner cards (one card earning across several businesses). They differ in what they reward: stamp cards reward frequency, points reward spend, tiers reward long-term value, thresholds reward basket size, and coalition cards reward breadth.
1. Stamp / punch cards
- How it works
- One stamp per qualifying purchase or visit. Reach the threshold, get the reward, reset.
- Best for
- High-frequency, low-value, low-variance purchases where every transaction is roughly interchangeable.
- Example
- Buy nine coffees, get the tenth free.
- Advantages
- Instantly understood, no exchange rate, visible progress.
- Limitations
- Treats a $3 espresso and a $9 brunch identically — rewards frequency, ignores value.
- Best KPI
- Visit frequency per enrolled customer.
- Ideal business
- Coffee shops, bakeries, quick-service food, car washes, barbers, nail salons.
2. Points-based cards
- How it works
- Earn a defined number of points per unit of currency spent; redeem points against rewards at a stated rate.
- Best for
- Variable basket sizes, where rewarding a large purchase the same as a small one would be indefensible.
- Example
- 1 point per $1; 200 points redeems for $10 off.
- Advantages
- Scales fairly with spend; supports partial redemption and bonus multipliers.
- Limitations
- Customers must learn an exchange rate; unredeemed points are a liability; slow to feel rewarding at low spend.
- Best KPI
- Average order value among members versus non-members, and points redemption rate.
- Ideal business
- Retail, e-commerce, restaurants with a wide price range, pharmacies, general merchandise.
3. Tiered loyalty cards
- How it works
- Customers move between named status levels based on cumulative spend, visits or membership length. Each tier unlocks better benefits.
- Best for
- Businesses where a minority of customers generate a majority of revenue and are worth treating differently.
- Example
- Silver, Gold and Platinum members receive 5%, 10% and 15% back plus escalating perks.
- Advantages
- An aspirational goal beyond the next reward; status is cheap to grant and disproportionately valued.
- Limitations
- Needs scale and spread; demotion is a real relationship risk; harder to explain and administer.
- Best KPI
- Share of revenue from the top tier, and tier retention rate year over year.
- Ideal business
- Hotels, gyms, premium retail, subscription and membership businesses.
4. Spend-based / threshold cards
- How it works
- A reward unlocks at a cumulative spend figure rather than a visit count. No exchange rate to learn — just a target.
- Best for
- Businesses that want a larger basket rather than a more frequent one.
- Example
- Spend $200 across the season, receive $25 off your next order.
- Advantages
- Simpler than points, still value-sensitive, and easy to price against a known spend.
- Limitations
- Too high reads as unreachable; too low is just a discount. Little visible progress until the end.
- Best KPI
- Average order value and the share of members who cross the threshold.
- Ideal business
- DTC brands, Shopify stores, home and furniture retail, garden and hardware, B2B trade counters.
5. Coalition / partner loyalty cards
- How it works
- One card earns and redeems across several independent businesses that share a scheme.
- Best for
- Businesses too infrequent on their own to sustain a programme, but complementary to neighbours.
- Example
- A high-street or shopping-centre scheme where a card earns at any participating merchant.
- Advantages
- Borrows frequency from partners, shares acquisition cost, reaches customers you could not.
- Limitations
- Shared governance, data and reputation; settlement between partners is hard.
- Best KPI
- Cross-merchant redemption rate — how many members earn at one partner and redeem at another.
- Ideal business
- Shopping centres, business improvement districts, market halls, complementary local independents.
Loyalty Card Type Matrix
| Type | Rewards | Needs | Customer effort to understand | Admin complexity | Reward cost predictability | Primary KPI |
|---|---|---|---|---|---|---|
| Stamp / punch | Frequency | High visit rate | Very low | Low | High | Visit frequency |
| Points | Spend | Variable baskets | Medium | Medium | Medium | Average order value |
| Tiered | Long-term value | Scale and spread | Medium | High | Low | Top-tier revenue share |
| Spend threshold | Basket size | Known seasonal spend | Low | Low | High | Threshold crossing rate |
| Coalition | Breadth | Willing partners | Medium | Very high | Low | Cross-merchant redemption |
3. Stamp vs points vs tiered: which should you choose?
Choose the mechanic from how your customers already behave, not from what looks modern. Frequent, similar-sized purchases suit a stamp or punch card. Variable spending suits points. High-value customers you want to keep for years suit tiers. A push for bigger baskets suits a spend threshold. And if your customers simply do not visit often enough for any of these to accumulate, a coalition scheme with neighbouring businesses is the honest answer — or no programme at all.
The most common error is choosing a mechanic because a competitor uses it. A points programme in a business where customers visit four times a year and spend $30 a visit — at a typical rate of one point per dollar and a 200-point threshold — takes well over a year to deliver a first reward, by which point they have forgotten they joined. The mechanic has to fit the purchase interval you actually observe.
Which Loyalty Card Should You Choose?
- Frequent, low-value visits → stamp or punch card
- Variable spending → points
- High-value repeat customers → tiers
- You want bigger baskets, not more visits → spend threshold
- Multiple businesses or partners → coalition
Set the threshold from the purchase interval, not from a round number. If your customers visit roughly weekly, a ten-stamp card is a ten-week commitment — long, but survivable. If they visit monthly, the same card is a ten-month commitment, and almost nobody finishes. A workable rule of thumb: the reward should be reachable within about two to three months of ordinary behaviour. Divide by your observed interval and set the threshold accordingly, then check what that reward actually costs you at that frequency before you print anything.
4. Benefits of loyalty cards
A loyalty card does three things a discount cannot: it identifies the customer, so their behaviour can be measured rather than guessed at; it creates a reason to return that is worth more the more they have already invested in it; and it establishes a first-party relationship the business owns rather than rents from a platform. The benefits that follow — repeat purchases, higher frequency, better retention — depend entirely on whether the mechanic fits the business.
The strongest evidence for why loyalty cards change behaviour comes from experimental research, and it is worth knowing because it tells you how to design the card, not merely that you should have one.
Progress accelerates behaviour, and the perception of progress is enough. Kivetz, Urminsky and Zheng ran a field experiment in a real café reward programme. Their published abstract reports that "participants in a real café reward program purchase coffee more frequently the closer they are to earning a free coffee," and — the finding with direct design consequences — that "the illusion of progress toward the goal induces purchase acceleration (e.g., customers who receive a 12-stamp coffee card with 2 preexisting 'bonus' stamps complete the 10 required purchases faster than customers who receive a 'regular' 10-stamp card)." They also report that "a stronger tendency to accelerate toward the goal predicts greater retention and faster reengagement in the program."
A convergent finding from a separate study. Published the same year and studying a different mechanism, Nunes and Drèze describe what they call the endowed progress effect: "people provided with artificial advancement toward a goal exhibit greater persistence toward reaching the goal. By converting a task requiring eight steps into a task requiring 10 steps but with two steps already complete, the task is reframed as one that has been undertaken and incomplete rather than not yet begun. This increases the likelihood of task completion and decreases completion time."
The instruction that follows is unusually specific: issue the card with credit already on it. A ten-stamp card handed over with two stamps applied is not the same offer as an eight-stamp card, even though the customer buys eight coffees either way — and the published research indicates the first performs better. This is one of very few loyalty-design decisions with peer-reviewed evidence behind it.
What a loyalty card gives the business
| Benefit | Mechanism | Only if… |
|---|---|---|
| Repeat purchases | An unfinished card is an unclaimed asset the customer already owns part of | The threshold is reachable in ordinary behaviour |
| Higher purchase frequency | Goal-gradient acceleration near the reward (see the research above) | Progress is visible to the customer |
| Customer identification | Transactions attach to a person instead of a till | The card is actually presented, not just held |
| First-party customer data | You own the relationship rather than renting it from an ad platform | Enrolment captures something you can contact them with |
| Segmentation and personalisation | Behaviour becomes visible: who lapsed, who accelerated, who never started | You look at the data at all |
| Promotion delivery | The card becomes a channel, not just a counter | The card is digital — a plastic card cannot carry a message you send after issuing it |
| Win-back opportunity | A lapsed member with a half-finished card is the easiest person to bring back | You can reach them, and you actually try |
| Customer lifetime value | The compound result of frequency, retention and basket size | You measure it against a comparison group — see Section 13 |
Loyalty marketing is full of confident percentages — "a 5% increase in retention raises profits by 25–95%", "it costs five times more to acquire a customer than keep one" — that circulate without traceable methodology or with sourcing that dead-ends in another blog post. We have deliberately not used them. The two studies cited above are peer-reviewed, linked to their DOIs, and describe experiments you can read. If a number matters to your business case, calculate it from your own transaction data rather than importing someone else's.
5. Digital vs plastic loyalty cards
Not universally. Digital cards win decisively on updates, reminders, measurement and cost per card, because the record lives on a server rather than in a customer's wallet. Plastic still wins where the transaction is fast and hands-on, where customers do not want to unlock a phone at the counter, and where a physical object carries brand weight. The honest answer for many high-street businesses is both: a digital card as the primary record, with plastic available for the customers who prefer it.
| Dimension | Plastic / physical | Digital / wallet |
|---|---|---|
| Customer convenience | Instant at the counter; one more thing to carry and lose | Always on a phone they already have; needs an unlock and a screen |
| Distribution | Print, ship, stock, hand over in person | QR code, link, email or receipt — no physical inventory |
| Replacement | Reprint and reissue; lost balance is usually lost | Re-add from a link; the balance lives on the server |
| Updates after issue | Not possible | Possible — the defining advantage |
| Notifications | None | Possible, within platform limits — see Section 7 |
| Data captured | Whatever the till records, if anything | Enrolment, every scan, timing, redemption |
| Analytics | Manual at best | Native — cohorts, lapse, redemption rates |
| Marginal cost per card | Printing, per unit, forever | Near zero after the platform cost |
| Branding | Tangible; a well-made card has real presence | Screen-sized and templated by the platform |
| Scalability | Constrained by print runs and shipping | Effectively unconstrained |
| Operational complexity | Low to set up, high to change | Higher to set up, low to change |
| Failure mode | Card left at home or lost | Flat battery, poor lighting on the scanner, customer declines to fetch phone |
Where the queue is fast and the transaction is under a minute, asking a customer to unlock a phone and find a pass can be slower than a card and a stamp. Where the clientele skews away from smartphone use, digital enrolment will simply fail. And where the card itself is part of the product — a members' club, a premium salon, a gym fob — the physical object does work that a screen does not. None of this is an argument against digital; it is an argument for knowing which of your customers each format serves.
6. What is a digital loyalty card?
A digital loyalty card is a loyalty card whose record is stored digitally and presented on a customer's phone rather than printed on plastic. The term covers three structurally different things: a card inside a business's own mobile app, which the customer must download and open; a card in a web account they log into; and a wallet-based card — a pass saved to Apple Wallet or Google Wallet, which uses the wallet app the customer already has. All three are "digital loyalty cards"; their adoption economics differ sharply.
The distinction matters more than the shared label suggests.
| Dimension | Card in a business's own app | Card in a web account | Wallet pass |
|---|---|---|---|
| Customer must first… | Find, download and open an app | Create an account and log in | Tap "Add to Wallet" once |
| Lives in | Your app | A browser | Apple Wallet or Google Wallet |
| Adoption friction | High | Medium | Low |
7. Loyalty cards in Apple Wallet and Google Wallet
Yes, on both — but through different systems that behave differently. Apple calls the loyalty format a "store card" pass, one of five pass styles in its Wallet Passes framework, and Apple's documentation describes the store card page as covering how to "Construct a digital pass for store loyalty and gift cards." Google Wallet issues loyalty cards through its own API using a two-part model of classes and objects. Both support scannable barcodes and both can be updated remotely after issue. Google additionally publishes an explicit numeric cap on notification-triggering messages, which Apple does not state in that form.
Apple Wallet
Apple's Wallet Passes documentation describes passes as "digital representations of information that previously might have been distributed on paper, plastic, or some other physical item," and states that with the framework you can "Create boarding passes, event tickets, store cards, coupons, and generic passes." Loyalty cards use the store card style. Primary documentation · checked 24 August 2026
Apple's store card documentation states: "Use store card passes for discount, gift, and store loyalty cards. Setting the store card pass style provides data that Wallet displays automatically, such as store points, rewards balance, terms and conditions, and other helpful information." On scanning: "Wallet supports 2D barcodes using QR, PDF417, Aztec, Code128, Code 39, Codabar, EAN-13, and Interleaved 2 of 5 (ITF) formats," and passes can also work with NFC readers, in which case "the pass doesn't need a barcode." Primary documentation · checked 24 August 2026
Apple documents three distribution routes: "Add a pass from an app or App Clip," "Provide a download on a web page for one pass or a bundle containing multiple passes," and "Send a pass as an attachment in an email." Most independent businesses use the second — a QR code at the counter leading to an Add to Apple Wallet button. Updates work by "distributing a new version of the pass with the same pass identifier and serial number," optionally driven by a web service you host. Primary documentation · checked 24 August 2026
Google Wallet
Google's model differs in structure. Its loyalty documentation defines a Passes Class as "a shared template that is used to create one or more passes you will issue to your users," and a Passes Object as something that "specifies the unique details of a specific pass that is issued to a specific user." The consequence is operationally significant: "Changes made to a Passes Class instance will propagate immediately across all Passes Object instances that reference it" — one template edit changes every card issued from it. Primary documentation · checked 24 August 2026
Google also publishes a numeric limit Apple does not state in the same form. Its push notification documentation states: "You may send a maximum of 3 messages that trigger a push notification in a 24 hour period." The same page states the equivalent cap for updates — "You may send a maximum of 3 updates that trigger a push notification in a 24 hour period" — and adds that "Google may throttle your push notification delivery quota if it deems you are spamming your users." Primary documentation · checked 24 August 2026
They share a customer-facing idea — a card in the phone's wallet — and diverge underneath in how passes are created, distributed, updated and notified. Platform behaviour also changes without notice. Verify anything load-bearing in the current documentation before building on it; the links above go directly to the pages quoted, all read on 24 August 2026.
The notification cap is a design constraint, not an inconvenience. Three notification-triggering messages in any rolling 24 hours is not a limit most loyalty programmes will hit — but it does mean wallet cards cannot be operated like an email list, and it makes the question "does this message report a real change in the customer's own state?" the right editorial filter. A balance update, a reward earned, an expiry approaching: those earn a slot. A generic promotion does not.
8. How to start a loyalty card program
Eleven steps, in this order: define the business goal as a number you already track; identify the customer behaviour that would move it; choose the reward mechanic that fits that behaviour; price the reward before designing anything; design the card; choose plastic, digital or wallet; build an enrolment process that takes under thirty seconds; launch to a small group first; promote it at the point of purchase; measure against a comparison group; and optimise one variable at a time. The order matters, and in our experience the programmes that fail are the ones that started at step five.
1Define the business goal
State it as a number you already measure and a date. "Raise midweek covers from 40% of weekend covers to 55% by December" is a goal. "Improve customer loyalty" is a mood. If you cannot name the number, you will not be able to tell later whether the programme worked.
2Identify the customer behaviour
Which specific behaviour, if it changed, would move that number? One more visit a month from existing customers is a different problem from a larger basket on the visits you already get — and they call for different mechanics. Get this from your transaction data, not from intuition.
3Choose the reward mechanic
Use the decision tree in Section 3. Frequency behaviour → stamps. Spend behaviour → points or threshold. Long-horizon value → tiers. Choose one. Programmes that run two mechanics at once usually confuse customers into using neither.
4Set the economics
Before designing anything, work out what the reward costs at the frequency you expect — Section 12 covers how to budget it. Then ask the harder question: how many extra visits does one reward have to generate to pay for itself?
In that illustrative case the bar is low: the card has to generate about half an extra visit for every reward it gives away. Rerun it with your own numbers before launch — at a 25% gross margin the same free drink costs $4.88 against $1.63 of margin, and needs roughly 3.0 extra visits to break even. That is a completely different proposition, and it is the reason to build the model before you print anything. Illustrative example — not a customer result
5Design the card
The card has one job at the counter: make the customer's progress obvious in under a second. Brand it, but do not let branding crowd out the balance. Section 9 covers this in detail.
6Choose physical, digital or wallet
Stage two of the decision tree. If the balance needs to change between visits — reminders, bonuses, expiry, tier moves — the card must be digital, because a plastic card cannot be changed after it leaves your hand. If your counter cannot reliably scan a phone screen, run both.
7Create the enrolment process
Enrolment is where a programme is most likely to lose people who were willing to join. Target under thirty seconds and under three taps: a QR code at the till, one screen, add to wallet. Every field you ask for at this moment costs you enrolments — collect the rest later, once they have a reason to trust you with it. PushNotice editorial analysis
8Launch
Start with one location, one shift or one segment before going wide. You are testing the operational reality — whether staff remember to offer it, whether the scanner works in your lighting, whether the flow survives a queue — not the concept. Fix what breaks while it is cheap to fix.
9Promote
The single highest-yield promotion is a staff member saying one sentence at the point of payment. After that: signage at the till, a line on the receipt, the QR code on the table or the bag, and your existing email list. Loyalty enrolment is an in-person sale far more often than a digital one.
10Measure
Hold out a comparison group from the start — a randomly chosen share of eligible customers who are not enrolled, or at minimum a clean pre-period baseline. Section 13 explains why that group is the whole ball game, and sets out the metrics.
11Optimise
Change one variable at a time and give it long enough to show. The threshold, the reward, the enrolment prompt and the reminder timing are the four levers worth testing; the artwork almost never is. Of those four, the threshold is where we would start. PushNotice editorial analysis
9. How to design a loyalty card that customers actually use
Three things, in order: the customer can tell at a glance how close they are to the reward; the reward is worth the effort and reachable in ordinary behaviour; and presenting the card takes no thought at the counter. Everything else — artwork, tiers, personalisation — is secondary to those three, and no amount of it rescues a card that fails any of them.
Design here means the mechanic and the moment, not the artwork. A beautiful card with an unreachable threshold is a worse product than a plain one with a threshold that fits.
| Decision | Get it right by… | Failure mode |
|---|---|---|
| Simple reward | Naming one concrete thing: "a free 12oz latte" | "Rewards" that nobody can picture go unclaimed |
| Clear value | Making the reward obviously worth the visits it costs | A 5% discount for ten visits reads as an insult |
| Visible progress | Showing the balance and the distance remaining | Hidden progress removes the visible cue the goal-gradient effect depends on |
| Achievable threshold | Dividing by the observed purchase interval (Section 3) | Enrolment without completion — the most common failure |
| Reward frequency | Aiming for a first reward within about two to three months | Too slow and they forget; too fast and it is just a discount |
| Brand consistency | Using the same name, logo and colours as the shopfront | An unrecognised card in a wallet gets deleted |
| Minimal friction | One scan, no login, no app switching | Staff stop offering it because it slows the queue |
| Useful updates | Messaging only real changes in the customer's own state | Promotional noise gets the pass deleted, permanently |
| Expiration strategy | Deciding deliberately, and stating it on the card | Silently expired balances turn advocates into complainants |
| Stated terms | Putting the rules where the customer can find them | Disputes at the till, in front of the queue |
Loyalty Card Design Checklist
- Clear reward — one concrete item or amount, named in the customer's language
- Simple mechanic — explainable in one sentence, with no exchange rate to memorise
- Easy enrolment — under thirty seconds, under three taps, no app download
- Visible progress — current balance and remaining distance both shown
- Relevant offers — tied to what this customer actually buys
- Clear terms — expiry, exclusions and reset rules stated on the card itself
- Mobile accessibility — legible at arm's length, scannable in your lighting, works on a cracked screen
- Useful reminders — only when the customer's own state has changed
- Measurement — a comparison group defined before launch, not after
The single best-evidenced design choice in this article: issue the card partly complete. Kivetz, Urminsky and Zheng found that a 12-stamp card carrying two bonus stamps was completed faster than a plain 10-stamp card requiring the same ten purchases, and Nunes and Drèze independently found that a ten-step task with two steps pre-completed beats an eight-step task. Both are cited in full in Section 4. The customer buys the same number of coffees; the card performs better.
10. Loyalty card reward ideas
Reward the behaviour you want more of, not the behaviour you already get. If you need frequency, reward visits. If you need larger baskets, reward spend. If you need reactivation, reward the return itself. The most common mistake is rewarding what customers were going to do anyway, which converts margin into goodwill without changing anything.
| Mechanic | Best business | Behaviour it targets | Main risk | How to implement |
|---|---|---|---|---|
| Buy 9, get 1 free | Café, quick service | Visit frequency | Rewards frequency regardless of spend | Stamp card, one stamp per visit, threshold from interval |
| Spend threshold reward | Retail, DTC | Basket size | Unreachable target is ignored entirely | Cumulative spend counter with a stated deadline |
| Birthday reward | Restaurants, salons, spas | A visit in a low month | Needs a date at enrolment, which costs enrolments | Ask for month and day only; trigger a week ahead |
| VIP / top-tier reward | Hotels, gyms, premium retail | Retaining your best customers | Demotion damages the relationship | Annual qualification with a soft landing on the way down |
| Double points day | Retail, e-commerce | Shifting demand into a quiet period | Trains customers to wait for the multiplier | Irregular, unannounced in advance, capped |
| Visit milestone | Gyms, studios, clinics | Sustained attendance | Milestones that are too far apart demotivate | Named milestones at 10, 25, 50 visits |
| Referral reward | Any local business | Acquisition through existing members | Gaming, and rewarding referrals that would have happened | Reward on the referee's first paid transaction only |
| Tier upgrade | Membership businesses | Crossing a value threshold | Complexity for customers who never approach it | Show the next tier and its distance on the card |
| Anniversary reward | Subscription, membership | Renewal at the decision point | Arrives after the decision if timed badly | Trigger before renewal, not on the date |
| Surprise reward | Cafés, restaurants, retail | Delight and word of mouth | Unpredictable cost; cannot be counted on | Small, occasional, unannounced, budget-capped |
| Early access | DTC, retail, events | Perceived status at near-zero cost | Nothing to access means nothing to offer | 24–48 hours ahead of the public for members |
| Member-only offer | Retail, hospitality | Making membership feel worth holding | Becomes an ordinary discount if run constantly | Genuinely exclusive, genuinely time-boxed |
| Seasonal reward | Retail, garden, hospitality | Smoothing a predictable trough | Discounting a period that was fine anyway | Target the trough you can see in last year's data |
| Product-specific reward | Retail, food, beauty | Trial of a specific line | Cannibalises a product they already buy | Reward with the new line, not with a discount |
| Frequency streak | Gyms, studios, cafés | Habit formation | One missed week destroys the motivation | Allow one forgiveness per streak |
| Win-back reward | Any repeat business | Reactivating a lapsed member | Teaches customers that lapsing pays | One-time, time-limited, only after a real lapse window |
Two of these deserve more caution than they usually get. A win-back reward that runs permanently is an instruction to your best customers to stop visiting for a month. A double points day that runs on a schedule teaches people to postpone purchases until it comes round. Both work when they are rare and unpredictable, and both quietly erode margin when they become fixtures.
11. Loyalty cards by business type
For a coffee shop, a stamp card — visits are frequent, baskets are similar, and progress is easy to show. For a restaurant, it depends on the format: a quick-service or lunch-led restaurant behaves like a café and suits stamps, while a full-service restaurant with a wide bill range and monthly visits suits points or a spend threshold. The determining variable in both cases is how often a typical customer returns, not the category name.
| Business | Best card type | Best reward | Example | Primary KPI | Wallet opportunity |
|---|---|---|---|---|---|
| Coffee shops | Stamp | Free drink | Buy 9, get 1 — issued as an 11-stamp card with 2 already applied | Visits per member per month | Strong — daily habit, phone already in hand |
| Restaurants | Points or threshold | Amount off a future bill | Spend $150 over the season, get $20 off | Visit interval and average bill | Moderate — reminders between long gaps |
| Retail stores | Points | Redeemable credit | 1 point per $1; 200 points = $10 | Average order value vs non-members | Strong — offers and balance updates |
| Salons | Stamp | Free or discounted service | Every 6th cut free | Rebooking rate | Strong — appointment reminders on the same card |
| Gyms | Tiered | Status perks, guest passes | Tier by months of continuous membership | Churn rate and renewal rate | Very strong — expiry and renewal alerts |
| Fitness studios | Stamp | Class milestone or streak | Attend 10 classes, get the 11th free | Class attendance per member | Very strong — class-pack balance on the card |
| Spas | Spend threshold | Treatment upgrade | Spend $400 in a year, receive an upgrade | Annual spend per client | Moderate — long intervals need reminders |
| Hotels | Tiered | Room upgrade, late checkout | Nights-based tiers with escalating perks | Repeat stay rate | Moderate — competing with brand programmes |
| Shopify brands | Points or threshold | Store credit, free shipping | Spend $200, unlock free shipping for a year | Repeat purchase rate | Moderate — the card extends an online relationship offline |
| DTC brands | Tiered or threshold | Early access, member pricing | Tiers by lifetime spend with drop access | Customer lifetime value by cohort | Strong — a branded card as an owned channel |
| Membership businesses | Tiered | Access itself | A membership card carrying status and expiry | Renewal rate | Very strong — the card is the product |
12. How much does a loyalty card program cost?
Total cost has four parts: setup (design, printing or platform configuration), software (a monthly platform fee, or development time if you build it), operations (staff time at the counter and campaign management) and — the part most businesses underestimate — the rewards themselves. Reward cost is the only line that grows as the programme succeeds, so it should be budgeted as a percentage of participating revenue rather than a fixed monthly figure.
| Cost line | Applies to | Behaves like | Commonly underestimated because… |
|---|---|---|---|
| Card design | Both | One-off | Redesigns are cheap digitally, expensive in print |
| Printing | Plastic only | Per unit, recurring | Reprints for damaged, lost and reissued cards add up |
| Software / platform | Digital | Monthly, often tiered by members | Pricing tiers step up as the member base grows |
| POS integration | Digital | One-off, sometimes ongoing | Depends entirely on which POS you run |
| Wallet infrastructure | Wallet cards | One-off setup, then platform | Apple requires a developer account and signing certificates |
| Campaign management | Digital | Staff hours, recurring | Someone has to decide what to send and when |
| Rewards | Both | Variable — scales with success | Budgeted as a fixed line, then blows through it |
| Enrolment / acquisition | Both | Staff time, signage, incentives | The launch incentive is often the largest single cost |
| Administration | Both | Recurring | Disputes, corrections and expiry queries take real time |
If your programme doubles its members and they all redeem, your reward cost doubles. That is the programme working, and it is also the single most common budgeting failure in loyalty. Model it as a percentage of the revenue passing through the programme, set a ceiling, and know in advance what you will change — threshold, reward value or expiry — if you reach it.
PushNotice's own pricing is published on the PushNotice site and is deliberately not restated here, so this page cannot go out of date against it. Compare any platform on total cost of ownership rather than headline subscription: setup, integration, per-member scaling and the staff time to run it.
13. How to measure loyalty card performance
Measure incremental contribution against a comparison group, minus every cost including the rewards actually redeemed. Enrolments, scans and redemptions are leading indicators — they tell you the programme is functioning, not that it is working. The business outcome is whether enrolled customers behave differently from comparable customers who are not enrolled, and without a comparison group you cannot separate your programme from your season.
| Metric | Definition | Type | Needs a comparison group? |
|---|---|---|---|
| Enrolment | Cards issued in the period | Leading indicator | No |
| Activation | Share of enrolled cards used at least once | Leading indicator | No |
| Active members | Members with a scan in the last 30 or 90 days | Leading indicator | No |
| Reward redemption rate | Rewards claimed ÷ rewards earned | Leading indicator | No |
| Repeat purchase rate | Share of members buying again in a window | Outcome | Yes |
| Purchase frequency | Transactions per member per period | Outcome | Yes |
| Average order value | Mean basket among members | Outcome | Yes |
| Customer retention | Share of members still active a period later | Outcome | Yes |
| Churn | Members who lapse past your defined window | Outcome | Yes |
| Win-back rate | Lapsed members who return after a prompt | Outcome | Yes |
| Customer lifetime value | Expected gross contribution over the relationship | Outcome | Yes |
| Revenue from members | Revenue attributable to enrolled customers | Descriptive only | Yes — alone it proves nothing |
| Incremental revenue | Member revenue minus comparable non-member revenue | Outcome | Yes |
"Members spend more than non-members" is not evidence that the programme works. Your best customers join loyalty programmes — that is why they are your best customers. A programme that did nothing at all would still produce that result. The only way to know whether the card changed behaviour is to compare enrolled customers against similar customers who were not enrolled, or against their own pre-enrolment baseline. Everything else describes selection, not causation.
14. Common loyalty card mistakes
Most fail on arithmetic rather than execution: the threshold does not fit the purchase interval, so customers enrol and never finish. The next most common causes are treating loyalty as a permanent discount, sending messages nobody asked for until the card is deleted, and never defining what success would look like — which makes it impossible to tell a failing programme from a slow one.
| Problem | Why it happens | Better approach |
|---|---|---|
| Reward too difficult | The threshold was chosen as a round number, not derived from the purchase interval | Divide by observed frequency; target a first reward within about two to three months |
| Reward too easy | Fear of low participation leads to over-generosity at launch | Price the reward against margin first; generosity is easy to add later and painful to withdraw |
| Unclear terms | Terms were written after launch, if at all | State expiry, exclusions and reset rules on the card itself before issuing one |
| No real customer value | The reward was chosen for what it costs the business, not what it means to the customer | Offer something customers already want; a 5% discount is not a reward |
| Too much complexity | Multiple mechanics bolted together over time | One mechanic, explainable in a single sentence at the counter |
| No segmentation | Every member gets the same message regardless of behaviour | Separate at minimum: new, active, at-risk and lapsed |
| No follow-up | The programme is treated as infrastructure rather than as something operated | Message on real state changes — reward earned, expiry near, tier reached |
| Too many notifications | The card is mistaken for a promotional channel | Respect the platform caps and a stricter editorial one; a deleted pass never comes back |
| Poor enrolment | Sign-up asks for data the business wants rather than data it needs | Ask only for what you need at the counter; collect the rest after the first reward |
| No measurement | No comparison group was defined before launch, so nothing can be attributed | Hold out a random share of eligible customers from day one |
| Ignoring inactive members | Reporting focuses on active members because the numbers look better | Define a lapse window and run a deliberate, time-limited win-back |
| Mechanic vs frequency mismatch | The mechanic was copied from a competitor with different economics | Choose from your own data using Section 3 |
| Loyalty treated as discounting | Margin pressure turns the programme into a standing promotion | Reward behaviour change, not the purchase; if it rewards what they would have done anyway, it is a discount |
15. Loyalty cards vs loyalty apps
Only if the app does substantially more than loyalty. A dedicated app carries the highest adoption friction of any option — the customer has to find it, download it, create an account and remember to open it — and that cost is only worth paying when ordering, booking, content or account management justify it. If the card is the whole product, a wallet pass delivers the same customer-facing outcome without an install.
| Dimension | Dedicated mobile app | Digital card in a web account | Wallet pass |
|---|---|---|---|
| Installation friction | High — search, download, account, permissions | Medium — account creation | Low — one tap |
| Customer access at the till | Open app, navigate | Log in, find page | Swipe up from the lock screen |
| Updates after issue | Yes, via the app | Yes, on next load | Yes, pushed to the pass |
| Notifications | Yes, effectively uncapped by the platform | Email or web push only | Yes, capped by the platform |
| Development cost | Highest | Moderate | Lowest |
| Ongoing maintenance | Two app stores, OS updates, review cycles | One codebase | Platform handles rendering |
| Room for other features | Extensive | Moderate | Minimal — it is a card |
| Right when… | Ordering, booking, content or accounts justify the install | The relationship is already online | Loyalty is the whole job |
Apps are not the wrong answer — they are the wrong first answer. A coffee chain with mobile ordering, stored payment and store locators has ample reason to ask for an install, and the loyalty card belongs inside that app. An independent café asking customers to download an app for a stamp card is asking for a great deal in exchange for very little. The question is not which technology is better; it is whether you have enough to offer to justify the install. A fuller comparison is here.
16. How PushNotice fits into a digital loyalty card strategy
Everything above is true regardless of which tool you use, and the decision framework works just as well if you conclude that plastic cards or your existing POS module is the right answer. This section says where PushNotice fits, so you can judge it against the alternatives.
PushNotice is a platform for creating and running wallet-based customer engagement — loyalty cards, stamp cards, membership cards and coupons issued as passes to Apple Wallet and Google Wallet, without building a mobile app. It is a practical implementation option for the "digital / wallet" branch of the decision tree in Section 3: businesses that have decided the card needs to change after it is issued, and do not want to ship an app to make that possible.
What it does not do is replace your point of sale, your CRM, your email platform or your SMS provider. Those systems hold different data and do different jobs, and a wallet loyalty card sits alongside them rather than instead of them. If you are evaluating whether a wallet-based card belongs in your stack at all, what wallet marketing is is the better starting point, and the POS loyalty guide covers how it interacts with the till.
Frequently asked questions
Basics
What is a loyalty card?
A loyalty card is a physical or digital card that identifies a returning customer and records their progress toward a reward. It holds a customer identifier, a balance of stamps, points, visits or spend, and the reward those add up to. The customer presents it at checkout, the business adds credit, and at the threshold the reward is claimed.
How does a loyalty card work?
The customer enrols and receives the card, presents it at checkout, the business records the qualifying purchase and the balance rises, and at the threshold the reward is claimed. With a digital card the balance can also be changed remotely between visits — which is what makes reminders, bonuses and expiry warnings possible.
What are the five types of loyalty cards?
Stamp or punch cards, points-based cards, tiered cards, spend-based or threshold cards, and coalition or partner cards. They differ in what they reward: frequency, spend, long-term value, basket size and breadth respectively. Section 2 compares all five.
What is the difference between a loyalty card and a loyalty program?
The card is the artefact the customer holds; the programme is the set of rules it runs under — what earns credit, how much is needed, what the reward is and when it expires. You can change the programme without changing the card, and with a digital card you can change both after issue.
Do loyalty cards actually work?
They change behaviour when the mechanic fits the purchase pattern. Peer-reviewed field research found customers in a real café programme bought more frequently the closer they came to the reward, and that cards issued with bonus credit already applied were completed faster than equivalent cards started from zero (Kivetz, Urminsky & Zheng, 2006). A card whose threshold cannot be reached in ordinary behaviour does not work, however well it is designed.
Choosing a card type
Which type of loyalty card should I choose?
Choose from observed purchase frequency. Weekly or more suits a stamp card; every two to six weeks suits points if baskets vary and a spend threshold if they do not; quarterly or less with high customer value suits tiers. If nobody visits often enough to accumulate anything, consider a coalition scheme — or no programme.
What is the best loyalty card for a coffee shop?
A stamp card. Visits are frequent, baskets are similar, and progress is easy to display. Set the threshold from your observed visit interval rather than defaulting to ten, and consider issuing the card with a stamp or two already on it — the research in Section 4 supports it.
What is the best loyalty card for a restaurant?
It depends on format. A quick-service or lunch-led restaurant behaves like a café and suits a stamp card. A full-service restaurant with a wide bill range and monthly visits suits points or a spend threshold, because a stamp card would reward a $20 lunch and a $120 dinner identically.
How many stamps should a loyalty card have?
Enough that the reward is reachable within about two to three months of ordinary behaviour. Divide by your observed visit interval: weekly visitors finish a ten-stamp card in ten weeks, but monthly visitors would need ten months, and almost none will. The right number is derived, not chosen.
Can I run more than one loyalty mechanic at once?
You can, but it usually reduces participation. Customers who cannot explain the programme in one sentence tend not to engage with any part of it. Pick one mechanic, run it long enough to measure, and add complexity only once the simple version works.
Digital and wallet cards
What is a digital loyalty card?
A loyalty card whose record is stored digitally and presented on a phone. It covers three different things: a card inside a business's own mobile app, which requires a download; a card in a web account the customer logs into; and a wallet-based card saved to Apple Wallet or Google Wallet, which uses an app they already have. Their adoption economics differ sharply — see Section 6.
Are digital loyalty cards better than plastic?
Not universally. Digital wins on updates, reminders, measurement and marginal cost. Plastic still wins where transactions are fast and hands-on, where customers will not unlock a phone at the counter, and where the physical object carries brand weight. Many high-street businesses are best served by both.
Can a loyalty card be added to Apple Wallet?
Yes. Apple's Wallet Passes framework supports five pass styles and loyalty cards use the store card style, which Apple describes as constructing "a digital pass for store loyalty and gift cards" and which automatically displays data such as "store points, rewards balance, terms and conditions". Customers usually add one by scanning a QR code leading to an Add to Apple Wallet button.
Can a loyalty card be added to Google Wallet?
Yes, through the Google Wallet API, which uses a two-part model: a Passes Class is "a shared template that is used to create one or more passes you will issue to your users", and a Passes Object "specifies the unique details of a specific pass that is issued to a specific user". Editing the class propagates immediately across every object referencing it.
Do Apple Wallet and Google Wallet work the same way?
No. They present a similar idea but differ underneath in how passes are created, distributed, updated and notified. Google publishes an explicit cap of three notification-triggering messages per 24 hours; Apple does not state a limit in that form. Verify anything load-bearing against current documentation before building on it.
Do I need an app to run a digital loyalty card?
No. A wallet pass runs inside Apple Wallet or Google Wallet, which customers already have, so there is nothing to install. An app only earns its adoption cost when it does substantially more than loyalty. Section 15 covers when that threshold is met.
How do customers scan a digital loyalty card?
Usually by presenting a barcode or QR code on screen. Apple states Wallet supports 2D barcodes in "QR, PDF417, Aztec, Code128, Code 39, Codabar, EAN-13, and Interleaved 2 of 5 (ITF) formats", and that passes can also work with NFC readers, in which case no barcode is needed. NFC availability depends on your platform and hardware.
Launching and running a programme
How do I start a loyalty card program?
Define the goal as a number you already track, identify the behaviour that would move it, choose the mechanic that fits, and price the reward — all before designing anything. Then design the card, choose the format, build enrolment, launch small, promote at the point of payment, measure against a comparison group and optimise one variable at a time. Section 8 covers all eleven steps.
How much does a loyalty card program cost?
Setup, software, operations and rewards. The first three are reasonably predictable; the fourth is not, because reward cost scales with success. Budget rewards as a percentage of revenue passing through the programme rather than a fixed monthly line, and set a ceiling with a plan for what changes if you reach it.
How do you measure loyalty card ROI?
Incremental contribution against a comparison group, minus every cost including rewards actually redeemed. "Members spend more than non-members" is not evidence — your best customers join programmes anyway, so a programme that did nothing would produce that result too. Compare enrolled customers against similar customers who are not enrolled.
Should loyalty points or stamps expire?
A genuine trade-off. Expiry limits accumulated liability and can prompt a return before a deadline, but customers who lose credit they worked for do not forget it. If you use expiry, state it plainly on the card and warn people beforehand — a digital card can do that automatically, a plastic one cannot.
How do I get customers to sign up for a loyalty card?
The highest-yield method is a staff member saying one sentence at the point of payment. After that: a QR code at the till, a line on the receipt, signage where people wait, and your existing email list. Keep enrolment under thirty seconds and ask for as little as possible — every extra field costs sign-ups.
Sources & methodology
How this guide was written
Platform claims about Apple Wallet and Google Wallet are quoted from each platform's own developer documentation, read on 24 August 2026, and linked at the point of the claim rather than summarised second-hand. Behavioural claims are drawn from two peer-reviewed studies, both linked to their DOIs. Everything else — the decision tree, the type matrix, the reward mechanic matrix, the KPI framework, the cost categories and the business-type recommendations — is PushNotice editorial analysis, labelled as such where it appears, and reflects reasoning about the mechanics of each option rather than measured results.
What this guide deliberately does not contain: no invented statistics, no customer case studies, no uplift percentages, and none of the widely circulated loyalty figures whose sourcing dead-ends in other marketing blogs. Where a number would matter to your business case, the guide tells you to calculate it from your own transaction data instead.
- Primary documentation Quoted directly from Apple or Google developer documentation, with a date and a link.
- Academic research Peer-reviewed published research, cited with authors, journal, year and DOI.
- PushNotice editorial analysis Our reasoning and frameworks. Useful, arguable, and not a measured finding.
- Illustrative example A worked scenario built to show a method. Not a customer result.
External sources
| Source | Type | Used for |
|---|---|---|
| Apple — Wallet Passes | Primary documentation | The five pass styles; the definition of a pass |
| Apple — Creating a store card pass | Primary documentation | Store cards for loyalty; automatically displayed data; supported barcode formats; NFC |
| Apple — Distributing and updating a pass | Primary documentation | The three distribution routes; how updates are delivered |
| Google — How classes and objects work | Primary documentation | Passes Class and Passes Object definitions; propagation of class changes |
| Google — Trigger push notifications | Primary documentation | The three-per-24-hours notification cap and throttling statement |
| Kivetz, R., Urminsky, O. & Zheng, Y. (2006). "The Goal-Gradient Hypothesis Resurrected." Journal of Marketing Research, 43(1), 39–58. | Academic research | Purchase acceleration near a reward; the 12-stamp card with two bonus stamps; retention effects |
| Nunes, J. C. & Drèze, X. (2006). "The Endowed Progress Effect." Journal of Consumer Research, 32(4), 504–512. | Academic research | Artificial advancement toward a goal increases persistence and reduces completion time |
Apple and Google revise their wallet documentation without notice, and capabilities available today may be added, changed or removed. Every platform statement here carries the date it was checked. Before you build anything that depends on one, verify it at the source.
About the author
Continue reading
Works through the goal, behaviour, mechanic and reward economics — steps 1 to 4 of the launch framework.
Open the planner →Derives a stamp or spend threshold from your observed purchase interval and margin.
Open the calculator →Models incremental contribution against reward cost before you commit to a threshold.
Open the calculator →The operational checklist for steps 5 to 9 — design, format, enrolment, launch and promotion.
Open the checklist →