Reward tier design is the part of loyalty programme construction that operators most reliably get wrong. The temptation is to set the thresholds high to control reward cost; the unintended consequence is that the first reward arrives so late in the guest's relationship with the programme that the programme is functionally dead in the guest's head before they ever redeem anything. A loyalty programme with no redemptions in the first four months is not a thrifty programme. It is a dormant one.
This post lays out a calibration framework for tier thresholds that balances perceived guest value against contribution-margin cost, anchored on time-to-first-redemption as the master metric. The framework produces tier thresholds an operator can set with confidence, run against POS data, and revise quarterly as the active guest base grows.
The Two Tensions That Govern Tier Design
Perceived value is ratio-based and time-decays. A reward feels generous when its monetary value is a meaningful share of the typical ticket, roughly 25-40% of average ticket value is the sweet spot for first-tier rewards. A reward feels stale when the path to it stretches past the timeframe in which the guest perceives an active relationship with the brand. The decay is not linear: a Tier 1 redemption that is 60 days away feels achievable; one that is 180 days away feels like a different programme entirely. By month 4 with no redemption, the guest has stopped opening the app, stopped checking their points, and stopped factoring the programme into choice between your brand and a competitor.
Cost of reward is contribution-margin forgone. Every reward redeemed is a unit of revenue that converts to zero gross margin (the COGS is still incurred but the price is zero). The operator's natural instinct is to set thresholds high to control this cost, but the cost calculation is incomplete without the counterfactual: what is the cost of the guest who never reaches the threshold, never redeems, and silently drops out of the programme because the first reward never arrived? That cost is invisible on the P&L but is the dominant driver of programme failure.
Time-to-First-Redemption Is the Master Metric
Time-to-first-redemption is the single most diagnostic metric in loyalty programme design. It measures, for an enrolled guest at the median visit cadence, how many days elapse between enrolment and the first reward redemption. The benchmark for a healthy programme is 30-45 days. Above 90 days and the programme is structurally broken regardless of how thoughtful the rest of the design is.
The illustrative failure: a 500-point Tier 1 at 100 points per visit and 1.4 monthly visits. The arithmetic is unforgiving. 500 points at 100 points per visit = 5 visits required. At 1.4 visits per month, that is 5 / 1.4 = 3.57 months, and that figure assumes the guest maintains baseline cadence with no slippage, which is exactly what the programme is supposed to be improving on. With realistic slippage, the actual time-to-first-redemption pushes past 5 months, and at lower-frequency variants of the same configuration the figure pushes past 7.9 months. By month 3, the guest's engagement with the app has dropped below the threshold where any subsequent push notification will be opened.
Worse, the operator only sees the outcome at the end: the guest who never redeemed never produced a positive engagement signal, the marketing team interprets the silence as "loyalty doesn't work for our category," and the programme gets quietly downgraded. The root cause is not the category. The root cause is a Tier 1 threshold calibrated to a visit cadence the active guest base does not actually have.
The Three-Tier Framework
Tier 1: visits 3-5, reward value AED 12-18. The first tier exists to create a redemption event in the first 30-45 days. The reward value is calibrated to be meaningful, a free coffee, a free side, a 25-35% off voucher on the next visit, but not so high that the contribution-margin cost dominates the early guest economics. At AED 45 typical ticket and AED 29 GM, an AED 15 reward at visit 4 is roughly 2.6% of the lifetime visit revenue through visit 12. That is sustainable.
Tier 2: visits 8-12, reward value AED 20-30. The second tier is the engagement consolidation tier. By visit 8-12, the guest has crossed the threshold where their behaviour can be described as habit rather than novelty. The reward is larger to reflect the larger relationship, but the ratio of reward to accumulated GM is tighter, the guest has now produced enough margin that a slightly larger reward is justified by the cohort economics.
Tier 3: visits 20+, reward value AED 40-60. The third tier is the recognition tier. Guests who reach Tier 3 are champions by any reasonable definition; the reward is large in absolute terms but the recipients are the cohort whose visits have the highest cumulative margin contribution. A meaningful reward at this tier serves as social signalling more than financial incentive, the guest is not visiting because of the reward; the reward is the brand acknowledging that the guest is a regular. The cohort is small, the per-redemption cost is bounded, and the retention value of explicit recognition is well in excess of the reward cost. The guest-cohort framing for these tiers maps to the six behavioural segments framework.
The Margin Math: Cost vs Incremental Lift
Worked example for a 5-location chain with 3,000 enrolled guests. Apply the three-tier framework to the cohort. Assumptions: AED 45 average ticket, AED 29 GM per visit, 1.4 monthly visits at baseline, 25% of the cohort responds to the programme with a frequency lift to 2.1 monthly visits. The cohort distribution by tier achieved over a 12-month period: 70% reach Tier 1 (2,100 guests, AED 15 reward × 1.5 redemptions average = AED 22,500), 30% reach Tier 2 (900 guests, AED 25 reward × 1.2 redemptions average = AED 11,250), 10% reach Tier 3 (300 guests, AED 50 reward × 0.2 redemptions average = AED 3,000). Total annual reward cost: AED 36,750.
Incremental margin from the responding cohort. 750 guests (25% of 3,000) shifting from 1.4 to 2.1 monthly visits = 0.7 additional monthly visits × 12 months × AED 29 GM = AED 243.60 per guest per year × 750 guests = AED 182,700 in incremental gross margin. Add a smaller frequency lift among the broader Tier 1 cohort — typically a 10-15% lift among the non-responding majority, and the total incremental margin reaches roughly AED 126,360 of net new margin attributable directly to the tier-design improvement.
The payback ratio. AED 36,750 in reward cost against AED 126,360 in incremental margin produces approximately a 3.4× payback on reward cost alone. Add platform cost, campaign budget, and operator time, and the ratio compresses, but it remains comfortably positive. Compare that to the equivalent 10-stamp card running on the same cohort, which captures roughly 5.4% of margin uniformly across the base with no cohort differentiation, costing more and producing less. The full breakdown of the 10-stamp comparison is in the broken loyalty math piece.
The Calibration Formula Every Operator Should Run
The Tier 1 threshold formula: median visits per month for active guests × 1.5 × average ticket = approximate Tier 1 threshold in points (assuming 1 point per dirham/rial). For an operator with 1.4 monthly visits and AED 45 ticket, that produces 1.4 × 1.5 × 45 = 94.5 points. Round to 100 points and the guest reaches Tier 1 in approximately 1.6 months, close to the 30-45 day target and well inside the engagement window.
Apply the same calibration logic to Tier 2 (median visits per month × 5 × average ticket) and Tier 3 (median visits per month × 12 × average ticket). The formulas embed the operator's actual visit cadence rather than borrowing thresholds from a US QSR template designed for a market with very different base cadence assumptions. Tier thresholds calibrated to your own POS data outperform thresholds borrowed from competitor programmes by a wide margin because the visit-cadence assumptions are correct on day one.
Run the formula. Set the thresholds. Watch the time-to-first-redemption metric drop into the 30-45 day band. The programme will do work it was not doing under high-threshold tier design. Habitu's platform runs the calibration formula automatically against your POS transaction history and proposes tier thresholds before launch, so the programme starts at the Pareto-optimal point rather than arriving there a year later through trial and error.