Designing Loyalty Programs: Points, Tiers, and Tech Implementation

By Alex Smith, CRM/Loyalty Lead (10+ years). Updated for 2026: wallet pass updates, privacy shifts, and stronger fraud playbooks.

Loyalty is not “endless discounts.” It is a system that guides behavior and still pays back. You set rules, track cost, and keep risk low. Done right, it raises repeat spend and trust. Done wrong, it burns margin and time.

Start with the economics, then the mechanics, then the tech. First, model value and risk. Next, pick points, tiers, cash, or perks. Last, wire events, a rules engine, a ledger, and channels. Order matters.

Field note, retail: frequency and AOV change with earn and burn rules. Too slow to earn, and users forget. Too easy to burn, and margin falls. See classic research on value by HBR on customer value: HBR on customer value. The lesson: design for the best users, not just the most users.

Field note, subscription SaaS: “loyalty” is not a coupon. It is status, support, and access. Good tiers add faster support SLAs, beta features, and training. Ties to NPS help. For background, see Net Promoter System fundamentals. Perks beat raw discounts in SaaS.

Field note, iGaming: rules are strict. KYC/AML, deposit limits, and safer play matter more than in retail. Abuse risk is high. Before you build, read UKGC guidance on safer gambling. The program must fit the guardrails first.

Now the money model. LTV depends on margin per order, order count, and churn. Loyalty also adds cost: points cost, perks cost, and fraud loss. You must plan for breakage (unused points) and set an earn-to-burn ratio that fits your margin. For context on loyalty economics, see McKinsey research on loyalty economics.

How to price a point. Say gross margin is 40%. You grant 1 point per $1. You let users burn at 100 points = $5 off. One point then is worth $0.05 at burn. Your cost per $1 earned is $0.05 × (1 − breakage). If breakage is 25%, cost is $0.0375 per $1. That equals 3.75% of revenue. With 40% margin, you spend 9.4% of margin on the program. Is the lift worth it? Test it.

Accounting note. Points often create a liability. You defer some revenue now and recognize it on burn or expiry. You also estimate breakage and reduce the liability. Read the IFRS 15 overview for the right frame. Work with Finance before launch. This is not legal or accounting advice.

Math break.

When to use points, cash, tiers, or perks-only? If users buy often and margin is mid to high, points can guide repeat. If price pain blocks conversion, cashback can help. If status drives stickiness, tiers win. In UX, redemption rules must be clear and near. See UX considerations for loyalty programs for common design gaps.

Avoid splashy changes without a test. Run holdouts and A/B tests. Start with small earn bumps and safe burn windows. Check incremental lift, not vanity. A classic read: Guide to online controlled experiments. It will save you from false wins.

For iGaming, align loyalty with KYC/AML and responsible play caps. Keep terms simple, show true earn and burn, and make cash-out rules fair. If you need a trusted source to compare operators, check https://norskigaming.com/ for clear reviews and loyalty terms you can verify.

Choosing Your Loyalty Mechanics: Goals, Risks, and What to Measure

Let’s build the platform. Start with events: earn, burn, adjust, tier-up, tier-down, expire. Use an event stream to write facts and a ledger to store balances. Later, expose APIs to the app, site, and support tools. Read the Event Sourcing pattern for a sound base.

Identity and access. Use OAuth 2.0 for secure app access and tokens. See the OAuth 2.0 spec. For tokens, JWT is common; details are in the JWT standard. Keep token scope tight and rotate keys.

Security and certs. If you handle personal data, set up ISO 27001 controls; see ISO 27001 information security. If you touch cards or co-brand, check PCI DSS. Log all balance moves; protect the ledger like cash.

Privacy. Collect the least data you need. Support data access and erase flows. For EU, start with GDPR basics. In California, see the CCPA overview. Tell users how you use their data in clear words.

Channels. Place points and perks where users act: app, web, email, and push. Wallet passes help with quick scans in store. See Apple Wallet passes and Google Wallet passes. Keep messages short and show balance and next step.

Pick a North Star metric for loyalty, then tie it to sub-metrics. Example: incremental margin per user. Sub-metrics: redemption rate, repeat buys, and CAC payback shift. For method, see the North Star Framework. Keep one score you can defend.

Use cohorts to prove lift. Compare like to like and use holdouts. Split by join month, channel, and baseline value. Avoid selection bias. A helpful read is Mixpanel’s Cohort analysis guide. Share cut views with product and finance.

Fraud control. Watch new accounts, device farms, and strange burn spikes. Rate-limit promo codes. Alert on fast earn-then-refund loops. Use role-based access. The OWASP ASVS checklist helps set a baseline and reduce gaps.

Mini-case, retail. A chain set “double points” only on slow weekdays and on low-margin SKUs gave no boost. They flipped to “double points” on mid-margin add-ons and at 2–4pm. Basket size rose 7%, liability held flat, and redemption moved to bundles. The win came from timing and product mix, not larger bonuses.

Mini-case, subscriptions. A SaaS brand dropped 20% discounts and launched three tiers: Priority Support, Early Access, and Admin Seats. Churn fell 1.8 pts for tier users, ARPU rose 6%, and refunds dropped. Perks set in the product beat coupons at checkout.

Go-live checklist and common traps.

• Define the goal: repeat buys, higher AOV, or lower churn.
• Write your point value, earn rate, and target cost share.
• Model breakage and defer revenue with Finance.
• Plan fraud rules: device checks, rate limits, and refund logic.
• Ship a holdout and a rollback plan.
• Write clear terms in plain words and show them in-app.
• Train support and sales with real examples.
• Set alerts for liability spikes and burn surges.
• Map privacy flows: consent, access, erase.
• Dry-run reports for week 1, month 1, and quarter 1.

30/60/90-day checks. At 30 days, scan redemption rate, support tickets, and fraud alerts. At 60 days, read cohort lift and margin vs target cost share. At 90 days, decide to scale, tune earn/burn, or sunset. Archive learnings in a playbook and refresh your risk model.

FAQ: What is a good earn-to-burn ratio? A safe start is 1–3% of spend as value back, with 10–35% breakage. Raise or lower by margin and lift. Keep it simple: 1 point per $1, 100 points = $5 off.

FAQ: Should I pick points or tiers? If you need repeat buys and many SKUs, pick points. If you sell status or service, pick tiers with perks. A hybrid works too: tiers for status, points for small nudges.

FAQ: How long does tech take? A lean stack ships in 6–10 weeks: 2 weeks for events and ledger, 2–4 weeks for rules and APIs, 2–4 weeks for channel UI and tests. Add time for compliance and data work.