Pillar

Shopify loyalty: programs that lift second-purchase rate

A loyalty program that doesn’t lift second-purchase rate is theatre. The mistake operators make is launching points because "competitors have one," then measuring enrollments and email opens. The metric that matters is reorder rate among members vs non-members in the same cohort. If that gap isn’t 8–12 points by month six, the program is decoration.

The "we’re a brand, loyalty is theatrical" objection is often right — and often wrong. It’s right when the brand sits at premium pricing and a points discount cheapens the positioning (luxury, design-led apparel, premium beauty). It’s wrong when the brand sits in a category where reorder is the dominant LTV mechanic (supplements, cosmetics, pet, food). The decision isn’t about taste; it’s about whether the math of the vertical wants a points program or a membership program or no program at all.

At $5M–$50M, the real loyalty lever is the structure — not the brand of the app. Three structures work: points-for-tiers (best for cosmetics, supplements, where reorder cadence is monthly), milestone-rewards (best for apparel, home, where reorder cadence is quarterly), and member-pricing (best for premium positioning where points feel cheap).

The vertical-by-vertical reality is sharper than the headline framing. Cosmetics with a strong reorder cadence sees 18–28% incremental LTV lift from a well-structured points program. Apparel sees 6–14% — meaningful but smaller, because reorder cadence is slower and emotional. Supplements often sees 20–35% because subscription + loyalty stack multiplicatively when designed together. Pet care is in the 15–25% range. Each of these is "the same program" in marketing’s deck and three different programs in operator reality.

The other operator question we hear often: "should we re-launch our existing loyalty program?" The honest answer in 60% of cases is yes — most programs were launched two years ago, are running on stale tier structures and over-discounted earn rates, and are quietly costing 4–7% of revenue in unearned discounts. A re-launch with tightened math and a member-only programming overlay typically returns to net-positive ROI within two quarters.

The year-two failure mode is consistent: over-discount + point devaluation. Programs that started with generous earn rates (3+ points per dollar, $0.01 per point) start looking expensive in finance reports around month 18, the team responds by devaluing points or quietly raising redemption thresholds, members notice within a quarter, trust erodes, and the program never recovers. The fix is to set conservative earn rates from day one and grow benefits over time, not shrink them.

This pillar covers which structure for which vertical, the app shortlist with real trade-offs (Smile, Yotpo Loyalty, LoyaltyLion, Stamped, Rivo), the tier math that determines whether you’re running a margin-killer or a margin-accretive program, and the member-only programming that lifts AOV on the reorder.

Three loyalty structures — pick by vertical

Points-for-tiers works best when reorder cadence is monthly and customers buy across a range of SKUs (cosmetics, supplements, household consumables). Customers earn points per dollar, redeem for $ off or product, climb to a status tier that unlocks free shipping, early access, or member-only SKUs. Tier-status is the retention hook, not the points themselves.

Milestone-rewards works best when reorder cadence is slower (apparel, home, premium beauty). Instead of point accrual, you reward at order #2, #4, #6 with specific gifts — a sample, a custom item, an experience. Easier to operate, more emotionally resonant, and avoids the points-bank perception problem.

Member-pricing works best at premium positioning where a points discount cheapens the brand. The program is a paid or earned tier (sign up = lower prices on everything, or hit $X spend = lower prices). Less points-program, more "club." Highest gross margin impact when done right.

The Shopify loyalty app shortlist

Smile.io is still the default at $5M–$15M for points-for-tiers programs that prioritize speed of launch. Clean admin, low cost, integrates with most things. Trade-off: limited custom branding, weaker reporting.

Yotpo Loyalty (formerly Swell) is the right pick if you’re already on Yotpo for reviews/SMS. Single contract, shared data model, deeper segmentation. Trade-off: only makes sense as part of the broader Yotpo stack.

LoyaltyLion is the choice at $15M+ when you want enterprise-grade tier construction, custom rules, and serious reporting. Higher cost, longer implementation, but the operational ceiling is the highest in the category.

Stamped is the value option at $5M–$15M; reviews + loyalty in one suite, lower app spend, decent (not great) loyalty tooling. Best when budget is the constraint and the loyalty program is straightforward.

Rivo is the design-first pick at $5M–$25M — best-in-class customer-facing UI, strong points UX, fast-shipping product team. Trade-off: smaller ecosystem of deep integrations than LoyaltyLion or Yotpo Loyalty.

Tier math — earning vs redemption

The single most common loyalty failure is giving away too much margin via redemption. The math to get right: (1) average points earned per $1 spent (typically 1–2), (2) redemption value per point (typically $0.01), (3) effective program discount (= earn rate × redemption rate × redemption %). Target an effective program discount of 4–6% of LTV. Above 8% the program is a margin trap; below 3% members don’t care.

Tier construction matters more than headline earn rate. A three-tier system (e.g., 0 points / 500 points / 2,000 points) where the top tier unlocks meaningful benefits (free shipping always, early access, member-only SKUs) drives 18–28% redemption among active members. Two tiers feel arbitrary; four feel bureaucratic.

Member-only programming

The programs that actually move the reorder needle aren’t the points themselves — they’re the member-only programming around them. (1) Member-only SKUs (limited drops, exclusive collabs) — lift reorder rate among top-tier members by 6–12 points. (2) Early access (24–72 hours before public launches) — generates the loyalty newsletter open rate and the BFCM Day-zero rush. (3) Member pricing on adjacent products — lifts AOV on the reorder by 10–18%. (4) Birthday or anniversary perks — small but emotionally resonant.

Brands that ship at least three member-only programs see 14–22 points of lift in reorder rate among top-tier members vs the no-program baseline. That’s the actual program ROI.

How to measure (and prove) loyalty ROI

The board-defensible KPI is incremental LTV per member, measured against a matched cohort of non-members. Build the cohort by customer signup month + first-order vertical + first-order AOV band. Look at 12-month LTV in each cohort. The gap is the program’s incremental contribution. Most healthy programs run +12–28% incremental LTV per member.

Vanity metrics to ignore: total enrollments (everyone enrolls for the welcome bonus), points redeemed (depends on how generous you were), and "members vs non-members LTV" without cohort matching (members self-select — they’d be high-LTV anyway).

Frequently asked questions

Is Smile.io still the default loyalty app at $5M+?
Yes for points-for-tiers programs at $5M–$15M where speed of launch matters more than custom tier construction. Above $15M, LoyaltyLion or Yotpo Loyalty (if you’re on Yotpo) gives you the reporting and segmentation depth Smile can’t match. Stamped is the value alternative if you also want reviews in the same contract.
How big does our brand need to be to launch loyalty?
$5M GMV is the floor for most verticals. Below that the redemption math doesn’t work — you don’t have enough returning customers to generate the member-cohort gap that proves the program. Above $5M the program typically lifts 12-month LTV by 14–22% per member within two quarters.
Points or tiers?
Both. The points are the earning mechanic; tiers are the retention hook. A three-tier system on top of a 1-point-per-dollar earn rate is the structure that compounds. Pure points-only programs (no tiers) plateau at 8–12% redemption; pure tier programs (no points) feel arbitrary to customers under 40.
How much should we discount via points redemption?
Target an effective program discount of 4–6% of LTV across all members (= earn rate × redemption value × redemption %). Above 8% it’s a margin trap that’s painful to unwind; below 3% members don’t engage. Tune the earn rate, not the redemption value — changing redemption value mid-program erodes trust.
Should we run referral inside the loyalty app or separately?
Separately, in most cases. The best referral apps (Friendbuy, ReferralCandy) are purpose-built around the share mechanic and the friend-side experience. Loyalty apps treat referral as a points-earning event, which under-monetizes it. Plug the referral app into the loyalty app for point-grant, but run them as separate programs.