Loyalty Program Design

Tier thresholds · Earn rates · Redemption rules

Set the tier thresholds (Silver / Gold / Platinum spend cut-offs), earn rate (points per dollar), and redemption rules of a retail loyalty program to maximise incremental margin from CLV-weighted segments. The trade-off: higher rewards lift retention and basket size of high-CLV customers, but each redeemed reward is a real cost on the income statement (and an accrued liability on the balance sheet). Foundational papers: Kopalle & Neslin (2003), Lewis (2004), Liu (2007).

Why it matters

Loyalty programs are everywhere in retail — designing them well is non-trivial

~3.8 B
Active loyalty-program memberships in the US (2023) — ~16 per household. Most go unused; design quality determines who actually engages.
Source: Bond Brand Loyalty annual report.
+12–18%
Documented spend lift from active program members vs matched non-members (Liu 2007 grocery study).
Source: Liu (2007), Journal of Marketing 71(4).
~$200B
Estimated value of unredeemed loyalty points held on US retailer balance sheets — the “breakage” pool.
Source: industry / Capgemini analyses.
CLV-weighted
Best-practice optimisation rule: reward incremental spend in proportion to CLV-elasticity, not a flat % across all members.
Source: Lewis (2004); Kopalle & Neslin (2003).

Where the decision sits

Strategic program design · revisited every 2-3 years

Loyalty design is a strategic decision: tier definitions, earn rates, and reward menus stick for years (changing them too often signals weakness). The economics depend critically on CLV — the program is worth designing only if expected retention and basket lift across CLV-weighted segments exceeds reward cost. This page focuses on the tier-and-rate sub-problem; the treatment-allocation question (which segments get which offer) is downstream.

CLV by segmentinput
Design tiers + ratesdecision
Members opt inacquisition
Earn / redeemoperations

Problem & formulation

Maximise net retention margin minus reward liability

Decision model
Tier thresholds + rates
Demand model
Per-segment retention & basket lift
Complexity
Concave NLP (tractable)
Reference
Kopalle & Neslin (2003); Lewis (2004)

Sets and parameters

SymbolMeaningUnit
\(s \in \mathcal{S}\)Customer segment (e.g., low / mid / high CLV)finite
\(t \in \mathcal{T}\)Tier (None / Silver / Gold / Platinum)discrete
\(N_s\)Number of customers in segment \(s\)customers
\(\bar S_s\)Average annual spend of segment \(s\)$ / yr
\(m\)Gross marginfraction
\(\theta_t\)Earn rate at tier \(t\) (% of spend rebated as points)fraction
\(\tau_t\)Spend threshold to qualify for tier \(t\)$ / yr
\(\rho_t(s)\)Retention lift for segment \(s\) at tier \(t\)fraction
\(\beta_t(s)\)Basket-size lift for segment \(s\) at tier \(t\)fraction
\(\gamma\)Redemption rate (fraction of points actually redeemed)fraction

Decision variables

SymbolMeaningDomain
\(\tau_t\)Spend threshold for tier \(t\)\(\mathbb{R}_{\geq 0}\)
\(\theta_t\)Earn rate at tier \(t\)\([0, 0.10]\)

Objective

For each segment, the program rebates \(\theta_t \bar S_s\) to the customer (the portion that is redeemed costs \(\gamma \theta_t \bar S_s\)). In return, retention rises and basket size grows. Net program margin is:

$$\Pi \;=\; \sum_{s \in \mathcal{S}} N_s \cdot \Bigl[\, \rho_{t(s)}(s) \cdot (1 + \beta_{t(s)}(s)) \cdot m \, \bar S_s \;-\; \gamma \, \theta_{t(s)} \, \bar S_s \,\Bigr]$$

where \(t(s)\) is the tier the customer qualifies for given spend \(\bar S_s\) and thresholds \(\tau\). Optimisation is over \((\tau, \theta)\) jointly; each combination implies a different mapping of segments to tiers.

Tier-assignment monotonicity

A natural constraint: lower tiers cannot have higher earn rates than upper tiers (otherwise no one would graduate):

$$\theta_{\text{None}} = 0 \;\;\leq\;\; \theta_{\text{Silver}} \;\;\leq\;\; \theta_{\text{Gold}} \;\;\leq\;\; \theta_{\text{Platinum}}$$ $$\tau_{\text{Silver}} \;\;\leq\;\; \tau_{\text{Gold}} \;\;\leq\;\; \tau_{\text{Platinum}}$$

Interactive solver

4 segments × 4 tiers · CLV-weighted retention + basket lift

Loyalty program designer
Tier thresholds + earn rates · net margin vs ungated baseline
★★ Closed-form per-tier evaluation
Fraction of points actually used
e.g., 2% = 2 pts / $
—
Net program margin ($)
—
No-program margin ($)
—
Lift vs no-program
—
Reward cost ($)
—
% of customers in Platinum
—
Tiers with members
Members per tier Per-tier net margin contribution Per-tier reward cost

Under the hood

The base has 4 segments by spend: Low ($250 avg, 5,000 members), Mid ($1,000 avg, 2,500 members), High ($3,500 avg, 700 members), VIP ($10,000 avg, 100 members). Each segment is assigned the highest tier whose threshold its spend exceeds. Per-tier retention & basket lifts are concave in earn rate (diminishing returns: \(\rho = 0.05 + 0.30 \theta / (\theta + 0.03)\), \(\beta = 0.05 \theta / (\theta + 0.05)\)). Reward cost = redemption_rate × earn_rate × spend. The solver evaluates net margin under the chosen design vs no-program baseline.

Reading the solution

Three patterns to watch for

  • Top-tier earn rate matters most. A handful of VIPs generate disproportionate margin; their tier’s earn rate is the most-sensitive lever.
  • Threshold spacing creates aspiration. If the gap between Gold and Platinum is too large, no one upgrades; too small, and Gold customers immediately become Platinum without spend lift.
  • Redemption rate is a free margin lever. Every percentage-point lower in \(\gamma\) (less point usage) = pure margin retained. Programs design hard-to-redeem rewards on purpose.

Sensitivity questions

  • Add a 5% Platinum bonus? — check if VIP retention lift exceeds the additional reward cost.
  • Halve Gold threshold? — many Mid customers move up; reward cost rises; check if their lift covers it.
  • Lower redemption (better breakage)? — pure margin gain; consider whether it damages perceived program value.

Model extensions

Coalition / partner programs

Multiple retailers share a program (e.g., American Express MR, Plenti). Cross-retailer redemption complicates the cost split.

Surprise & delight rewards

Non-tier-based, ad-hoc rewards calibrated by CLV (e.g., birthday surprise). Requires CLV scoring per customer.

Threshold dynamics + endowed status

Customers exhibit point pressure near tier boundaries (Hsee & Zhang 2010 endowment). Behavioural extension to the optimisation.

Subscription / paid loyalty

Amazon Prime, Walmart+. Decision: subscription fee, included perks. Higher CLV lift but selection effects.

Loyalty + personalised offers

Personalised redemption menus based on predicted utility. Joint with personalisation.

Liability accounting (IFRS 15)

Unredeemed points are deferred revenue; design must consider balance-sheet impact, not just income statement.

Dynamic tier benefits

Tier benefits change by season / category. Trade-off: complexity vs. customer focus.

Game-theoretic tacit collusion

Kim-Shi-Srinivasan 2001: loyalty programs as a collusion mechanism among retailers. Antitrust angle.

Key references

Kopalle, P. K. & Neslin, S. A. (2003).
The economic viability of frequency reward programs.
Journal of Retailing 79(3): 169–184. doi:10.1016/S0022-4359(03)00043-1
Lewis, M. (2004).
The influence of loyalty programs and short-term promotions on customer retention.
Journal of Marketing Research 41(3): 281–292. doi:10.1509/jmkr.41.3.281.35986
Liu, Y. (2007).
The long-term impact of loyalty programs on consumer purchase behavior and loyalty.
Journal of Marketing 71(4): 19–35. doi:10.1509/jmkg.71.4.19
Kim, B. D., Shi, M. & Srinivasan, K. (2001).
Reward programs and tacit collusion.
Marketing Science 20(2): 99–120.
Hsee, C. K. & Zhang, J. (2010).
General evaluability theory.
Perspectives on Psychological Science 5(4): 343–355. (Endowed-status / point-pressure psychology.)
Kivetz, R., Urminsky, O. & Zheng, Y. (2006).
The goal-gradient hypothesis resurrected: Purchase acceleration, illusionary goal progress, and customer retention.
Journal of Marketing Research 43(1): 39–58.
Bond Brand Loyalty (annual).
Loyalty Report.
Industry overview of program engagement and ROI benchmarks.
McCall, M. & Voorhees, C. (2010).
The drivers of loyalty program success: An organizing framework and research agenda.
Cornell Hospitality Quarterly 51(1): 35–52.

Back to the retail domain

Loyalty design sits in the Promotion × Strategic cell — the multi-year contract between retailer and customer that anchors the relationship.

Open Retail Landing
Educational solver · closed-form retention/basket lift functions and 4 fixed segments · production design uses calibrated lift estimates from controlled tests.