If you’re wondering how to design loyalty program structure correctly, the sequence is straightforward but unforgiving: define the business objective first, choose the right program archetype next, model earn and burn rates together, calculate liability with breakage assumptions, set tier thresholds from customer segmentation, plan data capture, choose the technology stack, pilot with a test cohort, and then measure and iterate after launch.
Teams often reverse that order, picking mechanics before defining the behavior they want to change, then discovering a year in that the earn rate makes liability unmanageable. That is why the design sequence matters for CRM leads, loyalty program designers, and marketing and finance teams building or refining programs in retail, banking, travel, and similar sectors: bad assumptions drive low engagement, expensive redesigns, and balance-sheet risk, while a sound design can lift purchase frequency, customer lifetime value, and acquisition efficiency.
This guide walks that sequence in order, with the formulas and worked numbers a CRM lead needs to defend the design to finance. It covers objective setting, program archetypes, earn and burn economics, liability and breakage modeling, tier design, data capture, technology options, pilot structure, performance measurement, and the common design mistakes that force a rebuild within 18 months. Tier thresholds are a common failure point, since a well-structured tier loyalty program depends on earn rates being modeled correctly first.
Most loyalty programs underperform not because the mechanics are wrong, but because the design process ran in the wrong order, tiers were built before earn rate was modeled, or a program archetype was picked before anyone defined the target behaviour it needed to change. Understanding the fundamentals of building a loyalty program the right way can help teams avoid this exact sequencing trap.
In our work modeling point liability and breakage for 2024-2025 client rollouts, and building the API-integrated engine behind USSF's fan engagement program, we consistently see teams sequence this backwards. This reversal multiplies redesign risk later.
The sequence that avoids that: choose your program archetype first, then define earn rate and burn rate together, model point liability (points issued × unit value × (1 − breakage rate)), and set tier thresholds. Plan data capture and segmentation, pick your technology approach, and run a pilot cohort before full rollout.
Each step below walks that decision in order, with the formulas and trade-offs you need at each stage.
Every loyalty program design decision traces back to one primary objective: raising customer lifetime value (CLV), lowering customer acquisition cost (CAC), or increasing purchase frequency. Pick one primary objective. A program built to do all three at once usually optimizes for none.
A program optimized for none of these objectives often becomes bloated with unnecessary mechanics that are vulnerable to abuse and fraud, undermining the very metrics it was meant to improve.
The objective determines the program archetype before you touch earn rate or tier thresholds. A CLV objective points toward tiered structures that reward depth of relationship over time. A CAC objective points toward referral-heavy mechanics that reduce paid acquisition spend.
Confusing the two is the most common risk we see in point liability modeling engagements, teams build tier decay rules meant for retention into a program whose actual goal was reducing CAC.
Open Loyalty's Loyalty Program Benchmark Report, which surveyed 230 loyalty professionals across more than 30 countries, found that over 37% of loyalty leaders define success by retention and lifetime value, and 39% by revenue, ROI, or profit. The programs that report cleanly are the ones anchored to a single commercial number rather than a basket of engagement metrics, with clear key performance indicators tied to business outcomes.
Write the objective down as a number, not a sentence: "reduce CAC by X percent" or "lift CLV by Y percent within 18 months," not "increase engagement." That single figure becomes the test every later decision, burn rate, reward catalog, coalition loyalty program partnerships, has to pass. After launch, monitor customer retention as one KPI to adjust loyalty program effectiveness; 60% of businesses with loyalty programs see higher customer retention.
Our full design framework walks through how to translate that objective into a target behaviour before scoping the platform or service provider.
None of this works without the right technology stack, since platform and integration choices determine whether your earn rate logic, tier thresholds, and redemption tracking can actually execute in real time.
Selecting among the main types of loyalty programs starts with choosing the right program structure for the objective you picked above. Five archetypes cover nearly every enterprise rollout, and each optimizes for a different KPI, not all of them at once.
A simple loyalty program structure is easier for loyalty program members to understand and use.
A paid loyalty program shifts the risk calculus: members pre-pay, so the earn rate can be richer without inflating point liability, because the fee itself funds part of the reward pool. Starbucks Rewards is the reference case for tiered mechanics done well, converting frequency into status without a paid gate.
A coalition loyalty program spreads acquisition cost across a partner network, which lowers each brand's CAC but adds redemption complexity, customer data sharing challenges across partners, and shared breakage-rate risk that a single-brand program never faces.
Before committing to any archetype, run the point liability formula against your rough earn rate: points issued × unit value × (1 − breakage rate). Industry data shows that breakage rates typically range from 15% to 40% depending on program design and member engagement patterns.
Breakage varies sharply by sector: grocery and CPG programs typically sit in the 20-30% band, while travel and B2B programs run far higher. Assume 30% and a program issuing 10 million points at $0.01 unit value carries meaningfully lower liability than the same volume modelled at a conservative breakage rate.
Get this wrong at the archetype stage and every later tier-threshold and burn-rate decision inherits the error. Whichever loyalty system or loyalty rewards program you choose, keep the loyalty members experience clear: they should understand the value quickly and be able to join and use it effortlessly.
Earn rate and burn rate are the two levers that turn a program archetype into a working reward economy, and they must be set together, not sequentially.
Earn rate is how many points a member accrues per dollar of qualifying spend. Burn rate is how many points a member redeems for a fixed reward value. In points-based programs, customers earn points for each qualifying action or dollar spent, which helps drive repeat purchases. The formula that ties them together is straightforward:
Effective rebate % = (points issued per $1 × redemption value per point) ÷ $1 spent
Substitute real numbers: at an earn rate of 1 point per $1 spent, and a burn rate of 100 points = $1 in reward value, the effective rebate is 1 percent of spend. Transaction rewards can take the form of reward points, discounts, or cashback.
Most retail programs target a 1 to 3 percent effective rebate against average order value (AOV), but the rewards offered should be reachable quickly, with at least 80% of transactions qualifying for rewards and, where possible, at least 80% of transactions qualifying for loyalty points so members can earn rewards within 30 days. So if AOV is $80, a 2 percent target implies roughly $1.60 in reward cost per transaction, the number you then divide by your burn rate to set the point-to-dollar redemption value, while keeping the redeemed value clear enough that members see real financial benefits and save money through rewards worth at least 10% of what they spend when redeemed.
Point liability follows directly from that same rate: points issued × unit value × (1 − breakage rate).
If you issue 10 million points at $0.01 unit value with a 30 percent breakage rate, liability on your balance sheet is $70,000, not $100,000: the gap is what unredeemed points save you, and what your finance team will ask you to defend.
Across industries breakage clusters around a 26% average and a 30% median, but the spread is what your model has to survive: QSR programs sit near the bottom at 10-20%, CPG in the 20-30% band, and travel well above both.
Treat that number as a lever, not a constant. In a study of high-value, low-frequency programs published in the International Journal of Research in Marketing, Goić, Montoya and Troncoso found breakage varies significantly both across customers and across markets, and that promoting more frequent redemptions lowers the breakage rate more effectively than steering members toward rewards priced at a greater number of points. If your model assumes one blended breakage figure for the whole base, it will drift the moment your redemption design changes.
Getting earn and burn wrong in either direction carries real risk: too generous erodes margin per the formula above; too thin drags redemption rate down and members disengage before reaching a tier threshold. Pilot cohorts, covered next, are how you de-risk that call before a full rollout.
Point liability is the dollar value of unredeemed points sitting on your balance sheet as deferred revenue. You calculate it with one formula: points issued × unit value × (1 − breakage rate).
Say your program issued 50 million points last quarter, each point carries a $0.01 reward value, and you assume a 25 percent breakage rate. Liability = 50,000,000 × $0.01 × 0.75 = $375,000. That figure stays on finance's books until members redeem or the points expire.
Breakage rate, the share of points a program expects will never be redeemed, is the single most consequential assumption in the model, because getting it wrong compounds. Overestimate it and you understate liability, a gap finance and customer service teams discover at audit. Underestimate it and you tie up cash reserves a growing program doesn't need.
Redemption rate is breakage's inverse and the number to track weekly, not quarterly. A program with a 60 percent redemption rate carries a 40 percent breakage assumption; drift in either direction should trigger a review of your earn and burn formula, not a shrug.
Three risks show up repeatedly across the liability audits our team has run:
The biggest risk sits in that first bullet. If you inherited a legacy platform's breakage default, please recalculate it against your own redemption data before locking the model. Set year-one breakage conservatively, low, so liability is overstated rather than under, then correct it once two full quarters of real redemption rate data from your own member community replace the industry average.
Tier thresholds work when they sit on real spend percentiles pulled from RFM segmentation, not round numbers copied from a competitor's site. RFM segmentation ranks every customer by recency, frequency, and monetary value, which gives you the percentile curve needed to place a threshold where an actual behavioral gap exists rather than an arbitrary dollar line.
Run the numbers before you name a tier. Say a mid-market retailer's trailing-12-month spend distribution shows a median of $180, a 60th-percentile value of $340, a 90th-percentile value of $820, and a 98th-percentile value of $2,100.
Set Silver at the 60th percentile, Gold at the 90th, Platinum at the 98th. Each threshold now reflects where the customer base actually clusters, which keeps the upgrade path achievable for the mass middle while still rewarding genuine high spend.
Each tier should also shift the earn rate multiplier and widen burn rate access, not just add a badge. In tiered programs, stronger loyalty benefits at each level give members a clearer reason to move up. A Silver member earning 1x and opening up standard rewards, versus a Platinum member earning 2x with early access to the reward catalog plus VIP events, makes the rewards offered more tangible and gives finance a clean way to model the risk of over-promising exclusive benefits against point liability.
Customers feel more valued when tiers combine real value through savings with experiential rewards, and that mix strengthens brand loyalty.
Open Loyalty's Loyalty Program Trends report, drawn from 170+ loyalty professionals across 19 industries, found 81% of consumers respond to visible progress indicators. A tier threshold is that progress bar, expressed in spend. It also helps explain why top-tier rewards should include meaningful exclusive discounts: 84% of consumers value exclusive discounts in loyalty programs.
A process this deliberate should please both marketing, which wants a visible upgrade story, and finance, which wants breakage rate assumptions that hold. We saw this in practice with SwipeRx: 91% customer retention rate.
Data capture design decides what segmentation and personalization are possible later. Define the schema before you shortlist a platform, not after, retrofitting fields into a live system is the single most expensive mistake we see in loyalty rebuilds.
Six field groups matter: identity keys (loyalty ID, hashed email, phone) for cross-channel matching across your CRM, CDP, and eCommerce software; SKU-level transaction history rather than basket totals; channel and campaign attribution; tier and status change history; redemption threshold proximity, meaning how close each member sits to their next reward; consent state per channel; and customer feedback signals for segmentation and offer refinement.
RFM segmentation needs recency and monetary data at the transaction level. Most industry-standard CRM exports only produce monthly aggregates, which is not enough. Pull a sample export from your current system and check the data grain as a formal process step before you scope any build.
If you are weighing a coalition loyalty program, add partner-attribution fields early. Shared redemption data across brands introduces a consent risk that is hard to retrofit once partners are live. If you offer value-based rewards, capture preferences for charity donations too, since some programs let members donate points to charities.
Loop in your data, CRM, and analytics leads before finalizing any request for proposal, and please confirm event-level logging exists end to end. Skipping this step is the most common reason a sound program archetype still fails to personalize in year two, the platform works, the schema underneath it does not.
Your program archetype decides how much technology control you actually need. Across common loyalty platforms, a single-market points program can run on legacy SaaS, though brands with heavy mobile usage may also evaluate a dedicated loyalty app. A multi-country tiered program with coalition loyalty program ambitions or a paid-tier component needs an API-first engine that your team, not a vendor's roadmap, controls.
In Open Loyalty's Loyalty Program Benchmark Report, nearly 45% of loyalty professionals named demonstrating ROI and securing resources as their single biggest challenge. That is a reporting problem before it is a marketing one, and the platform you choose either solves it or entrenches it.
Whatever platform you pick needs to compute point liability on demand, not at quarter close. The formula, points issued × unit value × (1 − breakage rate), should run against live transaction data, with seamless integration across e-commerce and POS systems so earning and redemption also stay accurate in store.
A brand issuing 50 million points at $0.01 unit value with a 30% breakage rate carries a $350,000 liability; a system that only reconciles this monthly is a finance risk, not a reporting inconvenience.
The U.S. Soccer Federation's fan engagement program issued more than 60 million loyalty points through an API-integrated build, giving finance and CRM teams shared real-time visibility into liability and redemption rate without custom middleware. That is the ceiling legacy SaaS structurally cannot reach.
For a full cost breakdown of the custom-build path, see our true cost of building a loyalty program in-house analysis before your team commits engineering headcount.
Pilot a loyalty program with a single geographic market or customer segment, run it against a holdout control group for 8 to 12 weeks, and validate your earn rate and burn rate assumptions before touching the wider member base, including whether members spend more or visit more often than the control group. This is the step that catches a mispriced program archetype before it becomes an industry cautionary tale.
The pilot's job is narrow: confirm that redemption rate in the test cohort clears the range your point liability model assumed, and that breakage rate isn't drifting so high it signals disengagement rather than earned discipline.
Watch three signals specifically. If the design is working, loyalty programs can increase customer spending by 2-3 times, 60% of businesses with loyalty programs see members spend 2-3 times more, and loyalty customers visit 40% more frequently than non-loyalty customers, with a 40% increase in customer visit frequency often showing up in the pilot data. Redemption rate delta between test and control tells you if the mechanic actually shifts purchase frequency. Tier threshold pass rates tell you if thresholds are calibrated to your real spend distribution, not a guess, and reveal how many customers actually qualify or redeem during the pilot. Support ticket volume tells you where the earn/burn logic confused members and how many customers contact support.
Don't run a pilot on a platform or software stack you'll have to migrate off later, the process of proving the model should also prove the technology choice.
Redemption rate is the first signal to check post-launch, and CLV (customer lifetime value) is the metric that decides whether the program actually pays for itself. Compare cohort-level redemption rate between your original test and control groups monthly and quarterly as part of your key performance indicators through the first two quarters, then run the same comparison against CLV, average order value, purchase frequency, and repeat purchases.
Recalculate point liability every quarter using the same formula from your pilot: points issued × unit value × (1 − breakage rate).
If real-world breakage drifts more than a couple of percentage points from your pilot assumption, revisit earn rate and tier threshold before the gap compounds into an accounting problem rather than a marketing one.
Redemption rates span a wide band by sector: retail programs commonly land between 40% and 60%, travel higher still, and e-commerce nearer 20-30%. Benchmark against your own vertical rather than the cross-industry average.
The design mistakes we see repeat across program audits: earn rates priced too rich to sustain margin, tier thresholds copied from a competitor with no RFM segmentation behind them, and a reward catalog left static a year past launch.
Each is a fixable process gap, not a platform limitation, the risks sit in governance, not in the software itself, and that risk grows quietly if nobody owns the quarterly review.
Please treat iteration as a standing review, not a one-off fix. If pilot data shows members already shop across a partner network, a referral mechanic can extend that community reach and cut CAC directly, see our guide to referral program design for the mechanics.
Set tier thresholds using purchase frequency and CLV (customer lifetime value) data, not arbitrary spend cutoffs, so each tier reflects a real behavioral jump and the rewards offered become more valuable as members move up. Most retail programs use three to four tiers, with most members sitting in the entry tier. Add tier decay rules or the top tier risks becoming a static discount club instead of delivering exclusive benefits and stronger loyalty benefits.
Layer gamified challenges - streaks, spend thresholds, seasonal missions - on top of your core earn rate to increase customer engagement and keep customers engaged between campaigns. Pair challenges with referring friends incentives that turn your customer community into an acquisition channel and lower CAC (customer acquisition cost). Test each mechanic in a pilot cohort before platform-wide rollout to confirm redemption rate holds.
Points and tiers are two different types of loyalty programs: points reward every transaction and suit high-frequency categories like grocery or QSR, while tiers reward cumulative status and suit considered categories like travel or beauty. Most enterprise programs combine both as a hybrid archetype because that program structure fits broader customer behavior better than choosing one. Pick tiers alone only if your industry sees fewer than four purchases a year.
A points or tiered program built on an API-first platform typically launches in 8 to 14 weeks, including pilot cohort testing. Custom builds and legacy SaaS migrations run longer, often 4 to 6 months, due to integration and data-migration work. Please budget extra time if you're rolling out a coalition loyalty program across multiple brands.
Design cost varies by technology approach: custom builds carry the highest software and service spend, while an API-first platform like Open Loyalty typically costs less than legacy SaaS licensing over time. Budget point liability funding separately from build costs. Request a vendor-specific quote before finalizing earn and burn rate economics.
Every program archetype, from tiered points to a coalition loyalty program, carries its own liability math, tier logic, and integration risk within broader customer loyalty programs. Getting the earn rate, burn rate, and tier threshold right before launch protects budget and member trust alike.
Our team has run this design process with retailers, banks, and airlines across the industry, and we know where the risks sit long before a pilot cohort proves them.
If your team wants a second set of eyes on the mechanics, please book a demo of the Open Loyalty Platform.
We will walk through your program archetype, model point liability against your redemption rate, and show how our API-first engine supports self-service configuration, stronger customer engagement, and meaningful interactions that reward customers and build loyalty without a rebuild of your existing stack.
API-first loyalty and gamification engine
Get a weekly dose of actionable tips on how to build and grow gamified successful loyalty programs!
