Telco Loyalty Programs: Telecom Rewards & Retention

1. Why telecom providers lose customers — and what a loyalty program can and cannot fix

Start with the numbers operators actually report, because they are more interesting than the folklore.

Where customers actually go — and how few of them go anywhere

The switching data is the part most retention strategies get backwards. In the UK, Ofcom's Switching Tracker found 16% of mobile customers changed provider in the previous twelve months in 2025, a rate Ofcom describes as stable and consistent across pre-pay and pay-monthly. The other 84% stayed. And when Ofcom asked why, the answers were not about price satisfaction:

• 71% agreed that "even if I could save money by getting a different deal, I am happy enough with my current deal" — up from 67% in 2023.
• 65% agreed that any savings would be "too low to make it worth spending time looking for a better deal."
• Only 22% of mobile customers made changes or negotiated with their provider in the previous year — the lowest engagement of any service Ofcom tested.
• 44% had been with the same provider for more than five years (2023 wave).

Read those four lines together and the strategic picture inverts. Most subscribers are not shopping. They are not comparing. They are, in Ofcom's own framing, disengaged — and disengagement is not loyalty. It is a stable state that ends abruptly when something disturbs it: a contract expiry notification, a price rise letter, a colleague's recommendation, a competitor's offer arriving at the right moment. The job of a telecom loyalty program is not to win a comparison the customer was never running. It is to make sure that when the disturbance comes, the customer has a reason to stay that is specific rather than habitual.

The nine reasons subscribers leave

Three of those nine are not loyalty problems at all, and a fourth — bill shock — is only partly reachable. A wallet pass does not improve coverage in a valley, does not lower a headline tariff, and does not shorten a hold queue. Any vendor implying otherwise — including us — should be asked which of the nine rows they are actually claiming to affect.

Related reading: customer retention strategies that survive contact with real businesses · how retention software fits the stack.

2. What is a telco loyalty program?

Why telecom loyalty is not retail loyalty — and not fintech or B2B loyalty either

Four structural differences do most of the work, and every design decision later in this guide traces back to one of them.

Revenue is already recurring. A coffee shop's loyalty card exists to cause a visit that might not otherwise happen. A telecom program has the opposite job: the revenue arrives automatically until the customer stops it. That inverts the metric. Retail loyalty is measured in incremental purchases; telecom loyalty is measured in avoided cancellations, which is far harder to attribute and much easier to fake.

Identity is already established. There is no anonymous telecom customer. The operator holds the account, the tenure, the plan, the usage pattern, the device and the term end date. Retail loyalty programs spend most of their design effort solving identification; telecom programs get it for free and frequently waste it, running undifferentiated campaigns to a base they could segment precisely.

Value per subscriber is high and multi-year. Ofcom put average UK mobile revenue at £13.24 per subscriber per month across all subscriptions in Q4 2025 — £15.79 for post-pay and £5.40 for pre-pay. AT&T reported postpaid phone ARPU of $57.04 for Q2 2025. Over a multi-year relationship those are large numbers, which is why telecom can justify reward economics no coffee shop could, and also why a badly designed program can destroy margin quickly.

The product is invisible when it works. A subscriber has no positive product moment in a normal month. The network simply works. The interactions that do occur are the bill, the price change and the fault. A loyalty program is frequently the only mechanism generating a non-negative impression between those events — which reframes it from a discount scheme into something closer to a presence strategy.

If you are designing outside telecom, the equivalents of this guide are the DTC loyalty software comparison and the B2B and distributor loyalty guide, which handles the rebate-versus-loyalty distinction in the depth it deserves. This page stays inside the telecom case throughout.

Related concepts: subscriber lifecycle, average revenue per user, customer lifetime value, digital membership cards, plan tier, add-on attachment.

3. Loyalty vs retention vs lock-in vs satisfaction

This is among the most operationally useful results in the telecom retention literature, and it has a direct consequence. If your retention is strong and your satisfaction is weak, you do not have a loyal base; you have a lease. The correct response is not a bigger reward budget. It is to find out which of the nine causes in Section 1 is producing the weak experience, fix that, and then use the loyalty program for what it is actually good at — being present, being specific, and being worth something at the moment the lease ends.

4. The retention arithmetic: churn, ARPU and CLV without the folklore

ARPU: check the denominator before you use the number

Cross-operator ARPU comparison is a common analytical error in telecom marketing content, and it is entirely avoidable. Operators measure different things and label them similarly.

Verizon's figure is roughly three times AT&T's not because Verizon extracts three times the revenue per customer, but because one counts accounts and the other counts lines. Any ARPU table without a denominator column is misleading, including in an internal deck.

Lifetime value: the elasticity that matters

The most widely cited published work here is Gupta, Lehmann and Stuart, "Valuing Customers," Journal of Marketing Research (2004). Modelling customer value as discounted cash flow at cohort level, they found that improving customer retention by 1% improves customer value by roughly 2.5–7% (2.45% to 6.75% across the firms studied). Their comparative elasticities are the more useful output:

• Retention elasticity: 3–7
• Margin elasticity: approximately 1
• Acquisition cost elasticity: 0.02–0.3

In plain terms: a percentage point of retention is worth an order of magnitude more than a percentage point off acquisition cost. That is a defensible, peer-reviewed substitute for the folk statistics below — with one scope caveat that must be stated whenever it is cited. The firms analysed were Amazon, Ameritrade, Capital One, eBay and E*Trade. None is a telecom operator, and the data is from the early 2000s. Use it as a directional argument about elasticity, not as a telecom number.

Acquisition cost: what operators actually disclose

No major operator publishes subscriber acquisition cost or cost per gross add as a headline metric any more; revenue recognition changes and device financing moved it inside the accounting. What is disclosed is the capitalised acquisition asset. T-Mobile's 2025 filings report roughly $2 billion of deferred incremental costs to obtain customer contracts, amortised over 24 months for postpaid service contracts.

The two statistics we removed from this article

Both appear in most telecom retention content. We traced both to origin and neither survives.

"It costs five times more to acquire a customer than to retain one." The most-cited version appears in a Harvard Business Review article from October 2014, which states that acquiring a new customer is "anywhere from five to 25 times more expensive than retaining an existing one" — with no source attributed at all. Every downstream citation we followed led back to that article or to vendor blogs citing it. There is no study, dataset or methodology behind the ratio that we could locate, and no telecom-specific version exists.

"A 5% increase in retention increases profits by 25–95%." This conflates two separate statements, neither of which says it. Reichheld and Sasser's 1990 HBR article "Zero Defections" says companies "can boost profits by almost 100% by retaining just 5% more of their customers" — a general assertion whose lead case is MBNA America, a single credit-card company, and whose MBNA-specific figure is a 125% increase in customer value rather than profit. Bain's published brief says "in financial services, for example, a 5% increase in customer retention produces more than a 25% increase in profit" — explicitly scoped to financial services, with no upper bound given. The "25 to 95 percent" range appears in the 2014 HBR piece attributed loosely to "research by Reichheld," and we could not find the 95% figure in any primary Bain or Reichheld publication.

5. Eleven telecom loyalty program models

What each model looks like in practice

1. Points on spend. The classic mechanic, and less common in telecom than elsewhere. Telstra Plus is the best-documented example: Telstra publishes an earn rate of 10 points per A$1 of eligible spend, and its FAQ states the rate is the same across all tiers — there is no enhanced earn for higher status. Points expire three years after they are earned. Design note: points create a deferred liability, so expiry rules and breakage assumptions belong in the design document, not in a later finance conversation.

2. Tiered status. Telstra Plus sets thresholds on annual eligible spend — Member up to A$1,499.99, Silver from A$1,500 to A$2,999.99, Gold at A$3,000 and above, with a separate Business tier. In the UAE, e&'s Smiles program tiers the earn rate rather than only the benefits: Silver earns 1 point per AED 1 on bill payment and recharge, Gold 2 points and Diamond 3 points — rising to 10 points above an annual spend threshold. Both are enrolment-based rather than automatic. Design note: the failure mode is tier inflation, where thresholds creep down until the top tier means nothing.

3. Plan perks. The dominant model, and the one customers most often describe when asked what their operator gives them. Benefits attach to the tariff: bundled streaming, roaming allowances, device protection, priority support. Airtel's Thanks structure is a clear published example, with Silver, Gold and Platinum benefit sets qualified by recharge or plan value rather than by accumulated points. Design note: the perk must be one the customer would otherwise pay for, or it reads as filler.

4. Tenure rewards. Genuinely rare, and therefore genuinely differentiating. MTN Nigeria's Prestige program recognises customers "loyal to the MTN network for 15 years and above" as Ever-Bright Prestige customers, and lower tiers receive a free data reward on the customer's own MTN anniversary. Orange Belgium's Thank You program states the principle explicitly: the longer you have been a customer, the more benefits you receive. Design note: this is the only model that structurally answers the loyalty penalty in Section 7, because it makes tenure an asset to the customer rather than only to the operator.

5. Upgrade rewards. Recognition delivered at the device end-of-life moment, when switching intent peaks. The mechanic is a credit, an enhanced trade-in value, or early upgrade eligibility conditional on staying. Design note: model this against the full financing term. A generous upgrade offer that resets a 24-month commitment is a retention instrument; the same offer without a commitment is a hardware discount.

6. Usage and add-on rewards. Bonus data, extended validity, a free add-on for adopting a target behaviour such as auto-recharge or paperless billing. Design note: the marginal cost of incremental data is low, which makes this the most margin-efficient reward in telecom — and the reason it is also the easiest to overuse until the paid data tier above it looks pointless.

7. Referral programs. Structurally attractive because referred customers arrive pre-qualified. Design note: pay on survival, not on activation. A referral reward paid at connection buys gross adds; a referral reward paid at ninety days buys subscribers. The difference shows up entirely in the second year.

8. Partner ecosystems. The model most large operators have converged on. T-Mobile Tuesdays delivers weekly partner gifts, discounts and prize draws — no points, no tiers, pure surprise-and-delight. O2's Priority, Vodafone's VeryMe Rewards, Three+, EE's One Up and Deutsche Telekom's Magenta Moments all run variants of the same idea. Deutsche Telekom is explicit that Magenta Moments is deliberately not a points program: instead of collecting points, benefits are simply given. Design note: partners fund most of the value in exchange for distribution, which is why this model scales with subscriber base and why it does not work for small operators.

9. Family and multi-line. Benefits that grow as a household adds lines or services. This is the one model that raises switching costs without any of the objections that attach to contractual lock-in, because the customer chose the consolidation and receives value for it. EE's One Up program is bundle-gated in exactly this way — available to customers who hold both EE Mobile and EE Broadband. Design note: converged operators have a structural advantage here that mobile-only operators cannot replicate.

10. Targeted retention offers. A specific offer to a specific customer at a specific moment, usually driven by a churn-risk model. Effective, and the most dangerous model on the list. Design note: if the only reliable route to a better price is to call and threaten to leave, you have not built a retention program — you have published a tariff for people willing to argue, and taught everyone else that patience is punished.

11. VIP and paid membership. A subscription tier that buys priority service or bundled benefits. e&'s "Smiles Unlimited" at AED 29 / month is a documented telecom example. Design note: a paid tier is a permanent service obligation with a cancellation path and consumer-law exposure, not a campaign. Do not launch one without an operations plan for the day 100,000 people expect priority support simultaneously.

Related reading: loyalty card types and how they differ · the mechanics behind customer loyalty cards.

6. What operators actually run today

Every entry below was verified against the operator's own website, help documentation, newsroom or annual report in September 2026. Programs we could not verify from an operator source are excluded rather than guessed at, and the ones that no longer exist are listed separately because getting that wrong is a common error in this topic.

The published outcomes — and the one number that does not exist

7. The loyalty penalty problem — the constraint no retail program has

What the regulators actually found

In December 2018 the UK's Competition and Markets Authority responded to a Citizens Advice super-complaint on the "loyalty penalty," estimating the total across five markets — mobile, broadband, cash savings, home insurance and mortgages — at around £4 billion. Two points of precision matter here, because this figure is routinely miscited. It is a combined figure across five markets, and the CMA document contains no mobile-specific number at all. The CMA also states its own caveat plainly: these are best estimates and "there are still gaps in the evidence base."

The mobile-specific analysis came from Ofcom in July 2019, and it is far better evidenced because it was built from actual customer-level billing data compared against each provider's own equivalent SIM-only tariff.

The regulator's own null result

Ofcom required end-of-contract notifications on the theory that telling customers their term was ending would prompt engagement. In May 2024 Ofcom published an ex-post evaluation of that intervention — and, to its credit, published a result that did not flatter it.

There is a lesson in that for anyone building a telecom lifecycle program, and it is not the obvious one. The same message, on the same trigger, through the same channel, produced a clear positive lift in one product and a mixed-to-negative effect in another. Notification alone is not an intervention. What happens after the notification — whether there is something specific and worth having on the other side of it — is the intervention. A contract-end notice that offers nothing is a well-timed reminder to go shopping.

What this means for program design

Three design rules fall directly out of the regulatory record, and they are the reason this section exists in a loyalty article rather than a compliance one.

Rule one: the loyalty program must not become the apology layer for the pricing structure. If tenure costs the customer money and the program hands back a fraction of it in partner vouchers, the program is not creating loyalty — it is funding a discovery delay. It also fails the moment a journalist or a regulator lays the two side by side.

Rule two: tenure should be worth something, explicitly. Model 4 in Section 5 is the structural answer to the loyalty penalty. It is rare precisely because it is expensive to run properly — and that is what makes it defensible. MTN's Ever-Bright Prestige recognition at 15 years and Orange Belgium's explicit "the longer you have been a customer, the more benefits you receive" are the two clearest published examples we found of an operator taking the opposite position to the penalty.

Rule three: never make the good price conditional on threatening to leave. Retention desks that reserve the best offer for customers who initiate cancellation are training the entire base — through word of mouth, forums and comparison sites — that the reward for loyalty is a phone call about disloyalty. It is also the mechanic most likely to attract regulatory attention, because it is a price difference based on customer engagement rather than on cost.

Related reading: retention when you cannot outspend the incumbent.

8. Telecom reward economics

9. The Contract Cliff — a telecom churn-intervention framework

Two published findings support the shape rather than the specific values. Ribeiro and colleagues' systematic review of telecom churn determinants (Management Review Quarterly, 2024, 37 studies covering 1999–2022) reports that 24-month minimum-duration contracts show substantially lower churn with departures spiking at contract expiry. And T-Mobile's own accounting treats 24 months as the benefit period for an acquired postpaid contract. For a postpaid base on minimum terms, the cliff is the point at which the instrument that was producing your retention number stops existing.

10. Do telecom loyalty programs actually reduce churn?

Six pieces of published research carry most of the weight, and two of the six are telecom.

The three findings that should change how you design

1. Most reported program ROI is selection bias. Leenheer and colleagues put the problem plainly: loyal customers select themselves into programs in order to benefit from them. When they corrected for that using panel data across all seven Dutch grocery loyalty programs, the genuine effect on share of wallet was "a small positive yet significant effect… seven times smaller than is suggested by a naïve model." If your program dashboard compares members to non-members, it is measuring who joined, not what the program did. That single sentence is the argument for control groups in Section 15.

3. The telecom-specific evidence is weakest where the money is. Ribeiro and colleagues' systematic review of 37 telecom churn studies reports that accumulated points reduced churn in some studies and showed no significant effect in others, "particularly among high-value customers." That result is uncomfortable, because high-value subscribers are precisely the population most retention budgets are aimed at. The plausible reading is that a points balance is simply not material to someone paying a large monthly bill — their decision is driven by network quality, price fairness and service experience, and a reward that would move a low-ARPU customer is noise at the top of the base.

What this does not mean

None of this means loyalty programs do not work, and the older sceptical position — Dowling and Uncles argued in MIT Sloan Management Review in 1997 that most schemes "do not fundamentally alter market structure" — was an argument rather than an experiment. The meta-analytic evidence since is real: Belli and colleagues, synthesising 429 effect sizes across thirty years, find strong evidence that programs enhance customer loyalty, with the qualification that the effect is concentrated in behaviour rather than attitude. Bolton and colleagues identify a specific and useful mechanism — members discount negative comparisons against competitors, which in a category where everyone occasionally has an outage is worth a great deal.

And the telecom study with the strongest design supports the direction: 3,004 customers of a Portuguese operator observed over four and a half years, with actual churn as the outcome rather than stated intention, found loyalty substantially reduced defection. Note carefully what "loyalty contract" means in that study, though — it is a contractual commitment, not a rewards scheme. It is a lock-in finding, and Section 3 has already covered what lock-in is and is not worth.

Related reading: retention strategies and the evidence behind them.

11. Where wallet passes fit in telecom — and where they don't

Why Tier-1 operators do not need a wallet pass

Because they have something better and can afford to maintain it. A large operator's app is not a marketing channel that happens to carry rewards; it is the self-service account layer — billing, usage, plan changes, support, SIM management — and those functions drive installation on their own. T-Mobile Tuesdays lives in T-Life, VeryMe in My Vodafone, Magenta Moments in MeinMagenta, Airtel Insider in the Airtel app. Once the app is installed for billing, adding a rewards tab costs almost nothing and gives full control over the interface.

That model has one dependency. It only reaches subscribers who installed the app and still open it. Operators do not publish that figure, so nobody outside them knows the split — but the three programs in our table that work without an app (Three+ in a browser, Telstra Plus on the web, MTN Prestige without app dependency) are evidence that operators with capable apps still consider non-app access worth building. O2 moved Priority the other way in August 2025, which shows the trade-off runs in both directions.

What a wallet pass genuinely does that an app does not

It requires no install. Apple's own guidance describes the Add to Apple Wallet badge as designed for web pages, emails and printed materials accompanied by a QR code, and documents three distribution routes: from an app, from a web download, or as an email attachment. Google states that its "Add to Google Wallet" link works "anywhere hyperlinks are supported, such as websites, email, and SMS messages," with no issuer app required. For a brand at rungs 3–5, that is often the difference between a program that reaches its base and one that reaches only its app installers.

It persists with little maintenance. A pass sits in a wallet the subscriber already uses for boarding passes and payment cards, and survives a phone change wherever the platform's own backup and sync is enabled. It is also far less likely to be cleared out during a storage clean-up than a rarely opened app.

It updates in place. Tier, points balance, upgrade eligibility date or benefit summary can be changed server-side and the pass reflects it — without the subscriber doing anything and without an app release cycle.

Channel background: what wallet marketing is — the broader channel this section sits inside · what a wallet pass actually is · reaching customers without a native app.

12. What belongs on a telecom wallet pass

Useful next step: draft the field list in the wallet pass planning worksheet before anything gets built.

13. Setting up in Apple Wallet and Google Wallet

Apple Wallet

Apple documents five pass styles: boarding pass, coupon, event ticket, store card and generic. For loyalty, Apple names the store card explicitly: "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." A store card can display logo and strip images and a barcode, with up to four secondary and auxiliary fields on one row.

Distribution. Apple documents three routes: from an app or App Clip, as a download on a web page, or as an email attachment. Apple's Add to Apple Wallet guidelines describe the badge as designed for web pages and emails, "or printed materials accompanied by a QR code that lets people add a pass to Apple Wallet," and state that users can add a pass directly from an email or a web page. Bundles of multiple passes are capped at 10 passes or 150 MB.

Updating. Passes update through a web service plus the Apple Push Notification service. Apple's documented flow: the user installs an updatable pass; the device registers it with your server and supplies a push token; pass information changes and your server sends a push; the device queries your server for updated passes and requests each changed pass. One operational detail worth knowing before you plan a launch: "A push notification for a pass update works only in the production environment."

Lock-screen relevance. A pass may carry up to ten locations and ten beacon UUIDs. maxDistance can only shrink the system's default radius — Apple's wording is that "the system uses the smaller of this distance or the default distance" — and Apple does not publish what the default is. Any article quoting a geofence radius in metres is quoting something Apple has not stated. Note also that relevantDate is deprecated in favour of relevantDates (iOS 18 and later); Apple's own how-to article still teaches the deprecated key, so most published guidance is written against it.

Google Wallet

Google organises passes into verticals — loyalty, offers, gift cards, event tickets, transit, boarding, generic — plus a separate generic private pass for sensitive categories. For a rewards program, the loyalty vertical is the documented choice: "Loyalty cards let users redeem and use their points more efficiently through a given loyalty program." A membership card with no points can use a generic pass, whose documented examples include membership cards.

No issuer app is required. Google's loyalty overview documents issuing passes "anywhere hyperlinks are supported, such as websites, email, and SMS messages," through an "Add to Google Wallet" link containing a signed JWT. One practical constraint: "The safe length of an encoded JWT is 1800 characters" — beyond that, browsers may truncate it and the save can fail, which is why production implementations create the object through the REST API first and reference it by ID.

Points are structured on Google in a way they are not on Apple. A loyalty object carries loyaltyPoints and secondaryLoyaltyPoints, each with a label and a balance, and the balance can be a string, an integer, a decimal or a currency amount. Tiers live on the class as rewardsTier and secondaryRewardsTier. Apple has no points-accrual semantic tag: its only balance-type structured value is balance, documented "for a store card pass" only, and Apple's own sample renders points as an ordinary display field. Apple does document membershipProgramName, membershipProgramNumber and membershipProgramStatus — the analogues of Google's programName and rewardsTier — but scopes them in its documentation to boarding passes.

Location. Google supports up to ten merchantLocations per class and ten per object, triggering a notification when a user enters "a Google-set radius" — which, as with Apple, Google does not publish. The older locations array is explicitly documented as "currently not supported to trigger geo notifications," so implementations built on it will silently do nothing.

Platform depth: Apple Wallet guide · Google Wallet guide · QR sign-up flows that convert.

14. Push notifications that bring telecom customers back

A pass update reaches a subscriber who deliberately kept your card on their phone — a stronger consent signal than an app permission granted during onboarding and forgotten. It also arrives with the hard ceiling described in Section 13 already attached. That constraint is a feature: it makes it impossible to run the volume that destroys the channel.

Channel background: how push automation and triggers work · lock-screen marketing in practice · what SMS actually costs by comparison.

15. How to measure a telecom loyalty program

The four measurement disciplines that separate credible programs from confident ones

1. Run a randomised holdout, permanently. Hold back a randomly selected 5–10% of eligible subscribers from the program entirely, and keep holding them back. The comparison that matters is enrolled-eligible versus withheld-eligible, not member versus non-member. Every meaningful number in layers 3 to 5 should be reported as a difference against that holdout.

2. Do not credit the program with revenue that would have occurred anyway. A subscriber who was always going to upgrade, upgrading while enrolled, is not an upgrade the program caused. Without a holdout you cannot tell the difference, and with one it is straightforward.

3. Report churn with contract status attached. Churn among in-contract subscribers is measuring lock-in. Churn in the 90 days after a minimum term ends is measuring your loyalty program. They are different numbers with different meanings and combining them hides the only result you actually wanted.

4. Accept that some metrics must move in opposite directions. Raising reward generosity lifts engagement and lowers margin. Tightening tier thresholds protects the status economy and reduces enrolment. Increasing notification frequency lifts claim rates and raises opt-outs. A KPI framework that shows every metric improving simultaneously is usually one where nothing has been decided yet.

Adaptable tools: retention dashboard · loyalty ROI calculator · retention planner — all built for general business; apply the telecom caveats above.

16. Telecom loyalty program implementation checklist

1. Define the business objective in one sentence, without a mechanic in it. "Reduce postpaid churn in the 90 days after minimum term end" is an objective. "Launch a points program" is a decision pretending to be one.
2. Identify the segment, and who is excluded. Prepaid and postpaid behave differently enough that a single program design rarely serves both. Business accounts are a third case.
3. Name the behaviour you want. Re-contract, upgrade, add a line, top up regularly, adopt an add-on, refer. One primary behaviour per program; secondary behaviours are bonuses, not objectives. The seven components in Table 3 are the parts you can assemble to reach it.
4. Choose the mechanic from Section 5 that fits that behaviour — not the mechanic that fits the phrase "loyalty program."
5. Model the economics before you announce anything. Reward cost per subscriber per month, expected take-up, contribution margin at risk, and for points programs the liability, expiry and breakage assumptions. Get finance's signature here, not at launch.
6. Write the terms. What is included, what is excluded, when benefits expire, how status is earned and lost, how to leave. Write them as though they will be published, because they will be.
7. Choose the technology against Section 11. If your subscribers do not open your app, an in-app program is a program for the subscribers you already retain.
8. Design enrolment and the pass. Field list from Section 12, sign-up flow, QR placement in stores and on the bill, and the retail script. Enrolment friction is where telecom programs most often quietly die.
9. Configure communications. Trigger list, frequency ceiling, change-message schema, quiet hours in the subscriber's timezone, and a named owner who can veto a send.
10. Build the holdout and capture the baseline. Randomised, 5–10%, permanent. Capture pre-launch metrics for both groups. This step cannot be added later.
11. Launch small. One segment, one region, one quarter. Test add and update on real iOS and Android devices, not simulators — Apple's pass push works only in the production environment.
12. Measure against the holdout, then decide. Quarterly for messaging and take-up, annually for reward economics, terms and tier thresholds.

• Business objective written and agreed with the commercial owner, with no mechanic in it
• Target segment defined, with exclusions documented and a reason recorded for each
• Primary behaviour named, and a measurement defined for it
• Reward economics modelled — cost per subscriber, expected take-up, margin at risk
• Points liability, expiry and breakage assumptions signed off by finance (if points are used)
• Program terms written and reviewed, including expiry, tier loss and cancellation
• Pass field list reviewed against Section 12 — zero account-security fields
• Change-message schema decided — which fields announce, which sync silently
• Frequency ceiling and quiet hours set, with a named owner who can stop a send
• Randomised holdout created and baseline metrics captured for both groups
• Enrolment tested end to end by someone who does not work in marketing
• Both platforms tested on real devices — add, update, notify, remove
• Opt-out path tested and confirmed to propagate across every channel
• Review date set for messaging, economics, tier thresholds and terms

17. Telecom loyalty program ideas that are worth running

Tenure and recognition

1. Anniversary data bonus. A one-off allowance on the subscriber's join anniversary. Behaviour: staying past a milestone. Cost: low marginal. Why it works: it makes tenure visibly worth something, which is the structural answer to the loyalty penalty.
2. Years-on-network status. A named recognition at 5, 10 or 15 years, with a real benefit attached. Behaviour: remaining on-network through successive contract cycles. Precedent: MTN Nigeria recognises 15-year customers as Ever-Bright Prestige. Why it works: status costs almost nothing to grant and cannot be transferred to a competitor.
3. Pre-cliff tenure benefit. A benefit that arrives 60 to 90 days before the minimum term ends, not after the notification. Behaviour: re-contracting. Why it works: it changes what the end-of-contract notice feels like when it lands.
4. Loyalty-priced upgrade. An existing-customer device price that is genuinely competitive with the acquisition offer. Why it works: it is the only mechanic that directly refutes "new customers get the best deal."

Usage, plan and add-ons

5. Auto-recharge bonus. Extra data or validity for setting up automatic top-up. Behaviour: converting an active decision into a passive one — the single highest-value prepaid mechanic there is.
6. Top-up streak reward. A bonus after three or six consecutive on-time top-ups. Behaviour: regularity, which predicts survival better than volume.
7. Roaming allowance for travellers. A bundled allowance triggered by past roaming behaviour. Why it works: one of the two highest perceived-value rewards in Section 8, delivered exactly when it is wanted.
8. Add-on trial with a stated end date. Three months of a paid add-on, with the end date on the pass. Behaviour: trial-to-paid conversion, without the auto-renewal complaint.
9. Plan-fit check. Proactively tell subscribers they are on the wrong plan and offer the right one. Why it works: it costs ARPU in the short term and buys the thing no reward can — the belief that you are not exploiting inattention.
10. Data rollover recognition. Surface unused allowance and let it carry. Behaviour: plan retention at the tier they are already on.

Household, referral and community

11. Multi-line benefit that grows. A benefit that increases with each line on the account. Why it works: raises switching cost legitimately, because the household chose it and receives value for it.
12. Household bundle reward. A benefit for holding mobile and broadband together. Behaviour: consolidating services onto one provider. Precedent: EE gates its One Up program to customers with both. Why it works: the whole household has to agree before anyone can leave.
13. Referral paid at 90 days. Reward when the referred line survives, not when it activates. Why it works: it buys subscribers rather than gross adds.
14. Local dealer perk. For dealers and regional carriers: a benefit redeemable in the store, on the pass. Why it works: it converts a transaction into a reason to come back.

Service, status and experience

15. Priority support for top tiers. A genuinely shorter queue, not a different hold message. Why it works: it is the benefit that matters most on the day something breaks.
16. Outage credit issued automatically. Credit applied without the customer asking. Why it works: it removes one of the most common reasons a satisfied customer becomes an unsatisfied one.
17. Partner perk with a real cadence. A weekly or monthly third-party benefit. Precedent: the dominant Tier-1 model. Caveat: see the coalition-program finding in Section 10 and measure it.
18. Device protection or repair benefit. A discounted screen repair or accidental-damage cover. Behaviour: keeping the device — and therefore the plan — in service longer.
19. Early access to network features. First access to a new service tier or feature for long-tenure subscribers. Cost: near zero. Value: status, which is the cheapest reward in the matrix.
20. Win-back with a specific reason. To former subscribers: what has changed since they left, stated concretely. Why it works: a generic "come back" offer to someone who left over coverage is an admission you did not read the file.

18. Common telecom loyalty program mistakes

1. Rewarding new customers better than existing ones, then running a loyalty program. Why it matters: subscribers compare, publicly, and regulators have already measured the gap. Better: make tenure worth something explicit, per Section 7.
2. Reserving the best offer for people who threaten to leave. Why it matters: it teaches the entire base that the reward for loyalty is a difficult phone call. Better: a published existing-customer benefit, and a separate, bounded save budget.
3. Choosing points because "loyalty program" means points. Why it matters: points create a deferred liability and, in telecom, the evidence for them is weakest among high-value customers. Better: work backwards from the behaviour.
4. Building the program inside an app most subscribers do not open. Why it matters: it reaches the customers who are least likely to leave. Better: match the channel to the rung on the scale ladder in Section 11.
5. Measuring members against non-members. Why it matters: it measures who joined, not what the program did. Better: a permanent randomised holdout.
6. Reporting churn without contract status. Why it matters: in-contract churn measures lock-in, not loyalty. Better: segment by contract status and tenure band.
7. Offering a reward before fixing an open fault. Why it matters: it converts a service failure into a transaction and is described publicly as such. Better: fix, acknowledge, then recognise — separately, and later.
8. Letting a batch job drive notifications. Why it matters: a field with a change message updated at 02:00 wakes your entire base. Better: separate fields that announce from fields that sync.
9. Launching a paid VIP tier without an operations plan. Why it matters: priority support is a permanent staffing commitment with a consumer-law cancellation path, not a campaign. Better: model the service capacity first.
10. Letting tiers inflate. Why it matters: thresholds that creep downward until the top tier is universal destroy the only thing status rewards sell. Better: defend thresholds annually and publish the change.
11. Auto-renewing and stepping prices up silently. Why it matters: it is the single fastest way to convert a satisfied subscriber into a complainant. Better: notify in advance, with the option stated.
12. Putting account-security fields on a wallet pass. Why it matters: a port-out code on a lock screen is an account-takeover vector, and no wallet pass is confidential. Better: the field checklist in Section 12.
13. Running one program for prepaid and postpaid. Why it matters: churn differs three- to four-fold and postpaid ARPU is roughly three times higher. The same reward is generous in one base and irrelevant in the other. Better: two designs.
14. Claiming a churn reduction you cannot evidence. Why it matters: no operator publishes one, so an internal claim will eventually be checked by someone who knows that. Better: report the holdout difference, with its confidence interval, or report nothing.

19. Where PushNotice fits

What it does, verifiably. Subscribers save a membership or loyalty pass to Apple Wallet or Google Wallet from a single link or QR code — no app for them and none for you. Pass content can be updated and a notification sent afterwards. Passes can be associated with places, so a card can surface when someone is near a store, on paid tiers. Higher tiers add tagging and segmentation, plus multiple workspaces for multi-location or multi-dealer operations. Pricing is published: a free plan at $0 covering 100 pass holders and 4 campaigns per month with no card required, Starter at $29 / month for 2,500 pass holders, Pro at $79 / month for 25,000 pass holders with tagging and segmentation, and a custom Agency tier for white-label use.

Product context: membership card builder · white-label and multi-client setups · plans and limits.

20. The telecom wallet decision framework

Step one: pick the channel from the behaviour

• Make membership and status visible and persistent → wallet pass. This is the case wallet is genuinely best at, because it survives without maintenance.
• Surface upgrade eligibility at the right moment → wallet pass, reinforced by email for the detail.
• Recognise a tenure milestone → wallet pass. Low frequency, high symbolic value, no authentication needed.
• Prompt a prepaid top-up before validity lapses → SMS. The subscriber may have no data.
• Warn about roaming charges abroad → SMS. Assume data is switched off.
• Explain a contract expiry and the options → email for the detail, wallet for the prompt. This message needs more room than a pass field.
• Handle a billing dispute or an open fault → human. No channel in this list is the answer.
• Let a subscriber change their plan or view usage → app or web portal. This needs an authenticated session, which a pass does not have.

Step two: run it through all three gates

Gate 1 — reach. The useful version of this question is not "do we have an app" but "what proportion of the subscribers we are trying to influence opened it in the last 30 days, and does that proportion skew towards the ones already least likely to leave?" Most operators can answer that from analytics they already hold, and the answer often decides the channel on its own.

Gate 2 — sensitivity. Anything requiring a logged-in session, and anything on the "never" list in Section 12, belongs behind authentication. A pass is a lock-screen-visible, shareable artefact. Design for that, not around it.

Gate 3 — connectivity. This is the telecom-specific gate, and it is the one that most often rules wallet out outright. A subscriber roaming with data disabled, or a prepaid subscriber whose balance has lapsed, will not receive a pass update when it matters — at best it arrives late, once they reconnect. Both of those are among the most valuable messages an operator sends, and both are SMS.

Frequently asked questions

The core questions

Rewards, points and tiers

Prepaid, postpaid and platforms

Channels, measurement and cost

Methodology, sources & disclosure

How claims are treated

Four categories of claim appear here and each is handled differently. Operator disclosures — churn, ARPU, program membership — are taken from SEC filings, investor releases and annual reports, with the reporting period stated and the denominator named. Regulator data — Ofcom, CMA, TRAI — is cited to the publication with its population and fieldwork basis. Platform documentation from Apple and Google was consulted on 6 September 2026 and quoted rather than paraphrased wherever precision matters. Research is peer-reviewed, reported with its sample, design and scope, including where the scope is not telecom. Everything else — the ten frameworks, the matrices, the decision tree, the scale ladder, the checklists and the ordering of the twelve implementation steps — is PushNotice analysis and labelled as such in each figure caption.

What was deliberately excluded, and why

The "5× cheaper to retain than acquire" ratio, because the most-cited version appears in a 2014 Harvard Business Review article with no attribution, and we could locate no study behind it. The "5% retention increase produces 25–95% more profit" figure, because Reichheld and Sasser's 1990 source says "almost 100%" based on a single credit-card company, Bain's published figure is "more than 25%" scoped explicitly to financial services, and the 25–95% range appears in neither. Any "industry average telecom churn rate," because GSMA Intelligence, Ofcom, TRAI, BEREC and ACMA publish none, and every version we traced led to a vendor site with no disclosed methodology. Every vendor case study claiming a churn reduction from a loyalty program, because none we found disclosed a control group, a sample or a measurement window.

We also excluded several programs that appear current in search results but are not: AT&T Thanks, discontinued during 2023; Verizon Up, superseded in June 2026; and Tokens de Movistar, closed on 22 October 2025 despite its marketing site remaining live in the present tense on Movistar's own domain. We excluded Optus Perks because we could not verify it from an Optus source, and MTN South Africa's YelloBucks because the pages exist but could not be read. Where an operator's own sources conflict — Verizon's newsroom and consumer guide describe Verizon Dollars eligibility differently — we state the part both agree on and no more.

Where we could not verify something, we say so

Three places in particular. No operator using Apple Wallet or Google Wallet passes for a loyalty program could be verified, and Section 11 reports that as a null result rather than as evidence of absence. Neither Apple nor Google publishes its default geofence radius, so this article gives no figure in metres, and any article that does is quoting something the platforms have not stated. And the hazard curve in Figure 3 is conceptual — operators do not publish churn by tenure month, so the shape is an analytical illustration of a documented pattern rather than plotted data, and the figure says so on its face.

Disclosure and limitations

PushNotice sells wallet marketing software, and Section 19 describes our own product — including an explicit statement that it is not suitable for a Tier-1 operator loyalty program, and a list of the telecom capabilities it does not have. The author is a technology founder, not a telecom operator executive, network engineer or regulatory economist, and this article is not legal, regulatory or financial advice. Operator programs change frequently; several described here launched or closed within the last eighteen months. Verify at the source before acting.

Primary sources

Operator disclosures: Verizon Q2 2026 SEC Exhibit 99; AT&T Q2 2026, Q2 2025 and Q4 2025 earnings releases; T-Mobile Q3 2025 SEC Exhibit 99.1, Q1 2026 and Q2 2026 earnings releases, and Q1 2025 Form 10-Q; Telstra FY25 CEO/CFO analyst briefing and Annual Report 2025 (Telstra Plus membership and engagement); Deutsche Telekom Annual Report 2024 and Annual Report 2025; VMED O2 UK Limited 2024 Annual Report and Consolidated Financial Statements.

Operator program documentation: T-Mobile Tuesdays support pages and newsroom; Verizon loyalty pages, Verizon Shine FAQs and the 16 June 2026 newsroom release; O2 Priority pages and Virgin Media O2 newsroom; Vodafone UK VeryMe Rewards pages; EE Rewards, EE One Up help pages and EE newsroom; Three UK Three+ pages; Deutsche Telekom Magenta Moments pages; Orange Belgium Orange Thank You; Orange France customer benefits; Telstra Plus FAQs, earn-points and how-it-works pages; Airtel Thanks and Airtel Insider; e& Smiles Rewards pages; MTN Nigeria Prestige.

Regulators and competition authorities: Ofcom, Telecommunications Market Data Update Q4 2025 (April 2026); Ofcom, Pricing and Consumer Engagement (2025), including the Switching Tracker, fieldwork August–September 2025, n=2,345 mobile decision-makers; Ofcom Switching Tracker 2023 (for the five-year tenure figure); Ofcom, An ex-post evaluation of mobile annual best tariff notifications and end-of-contract notifications (May 2024), and the earlier broadband ex-post evaluation of end-of-contract notifications (May 2022); Ofcom, Helping consumers get better deals: mobile handsets statement (July 2019) and the 2021 follow-up measurement; Competition and Markets Authority, response to the Citizens Advice super-complaint on the loyalty penalty (December 2018); TRAI, Performance Indicator Report Q4 2025 (March 2026).

Platform documentation, consulted 6 September 2026: Apple Developer — Wallet Passes framework overview, Creating a store card pass, Distributing and updating a pass, Adding a web service to update passes, Showing a pass on the Lock Screen, PassFieldContent, Pass.Barcodes, Pass.RelevantDates, SemanticTags, Personalize, Add to Apple Wallet Guidelines, and App Store Review Guidelines 1.5 and 3.2.1(iv). Google — Wallet API overview, loyalty card overview and web issuing, loyaltyclass and loyaltyobject references, Trigger push notifications for loyalty and generic passes, NotificationSettingsForUpdates, MerchantLocation, BarcodeType, generic and generic private pass documentation, and Wallet Terms of Service.

Research: Belli A, O'Rourke A-M, Carrillat FA, Pupovac L, Melnyk V, Napolova E, "40 years of loyalty programs: how effective are they? Generalizations from a meta-analysis," Journal of the Academy of Marketing Science 2022;50:147–173. Leenheer J, van Heerde HJ, Bijmolt THA, Smidts A, "Do loyalty programs really enhance behavioral loyalty? An empirical analysis accounting for self-selecting members," International Journal of Research in Marketing 2007;24(1):31–47. Bombaij NJF, Dekimpe MG, "When do loyalty programs work? The moderating role of design, retailer-strategy, and country characteristics," International Journal of Research in Marketing 2020;37(1):175–195. Ribeiro H, Barbosa B, Moreira AC, Rodrigues RG, "Determinants of churn in telecommunication services: a systematic literature review," Management Review Quarterly 2024;74(3):1327–1364. Ribeiro H, Barbosa B, Moreira AC, Rodrigues R, "Customer Experience, Loyalty, and Churn in Bundled Telecommunications Services," SAGE Open 2024;14(2). Gao L, de Haan E, Melero-Polo I, Sese FJ, "Winning your customers' minds and hearts: Disentangling the effects of lock-in and affective customer experience on retention," Journal of the Academy of Marketing Science 2023;51(2):334–371. Bolton RN, Kannan PK, Bramlett MD, "Implications of loyalty program membership and service experiences for customer retention and value," Journal of the Academy of Marketing Science 2000;28(1):95–108. Dowling GR, Uncles M, "Do Customer Loyalty Programs Really Work?", MIT Sloan Management Review 1997;38(4):71–82. Gupta S, Lehmann DR, Stuart JA, "Valuing Customers," Journal of Marketing Research 2004;41(1):7–18. Reichheld FF, Sasser WE, "Zero Defections: Quality Comes to Services," Harvard Business Review, September–October 1990. Counterpoint Research, Smartphone Replacement Cycle Forecast 2016–2029 (November 2025). McKinsey & Company, "Reducing churn in telecom through advanced analytics" (December 2017). Harvard Business Review, "The Value of Keeping the Right Customers" (October 2014) — cited only as the unattributed origin of the 5× claim.

About the author

Sajid Ali is Founder and CEO of PushNotice, a wallet marketing platform for small and multi-location businesses. He writes about wallet passes, loyalty mechanics and customer communication from a technology and business perspective. He is not a telecom operator executive, network engineer, regulatory economist or financial adviser, and the operator and regulatory material in this article is a summary of publicly available primary sources, provided so that telecom marketers and their advisers can have a better-informed conversation — not a substitute for one. Connect on LinkedIn.

Related guides