The right DTC loyalty program is the one that changes a behavior you can name: points to pull customers into a second and third order, tiers to lift spend among customers just below a threshold, paid membership to lock in your most frequent buyers. Whichever you choose, many rewards will go to people who would have bought anyway, so price them against contribution margin and prove the program with a holdout group before you scale it.
What a loyalty program should change
A loyalty program is a price change with a story attached. Every reward lowers the effective price for someone, and the only question is whether it buys behavior you wouldn’t have gotten otherwise.
Your best customers are the problem. They already buy often, so any program that rewards spend sends most of its rewards to the people who needed them least. Decide which behavior you want to move, and in which customers, before you pick a structure.
An RFM segmentation shows where customers stall: after one order, after a few quiet months, or at a spend level that never grows. Then match the structure to the behavior:
| Behavior you want to change | Who it applies to | Structure that fits best |
|---|---|---|
| Second and third order | One-time buyers | Points with a reachable first reward |
| Purchase frequency | Active customers in consumable categories | Points or paid membership |
| Spend per customer | Mid-value customers near a threshold | Tiers |
| Buying from you instead of marketplaces | High-frequency buyers | Paid membership |
| Reviews, referrals, profile data | All customers | Points for non-purchase actions |
| Engagement with launches | Fans of the brand | Access rewards, not discounts |
Pick one primary behavior; programs built to do everything end up rewarding purchases you already had. And if customers don’t return because the product or post-purchase experience disappoints, no program fixes that. Start with how to increase repeat purchase rate.
Points programs
Customers earn points on purchases and other actions, then redeem them for discounts, free products or perks. It’s the default in most Shopify loyalty apps and the easiest to launch.
The number that matters is the effective reward rate. If customers earn 1 point per dollar and 100 points is worth $5, you’re giving back 5% of spend. Treat that as a permanent discount on every enrolled order, because once customers expect points, you can’t take them away quietly.
Points work best when:
- Products are replenished or bought several times a year, so customers reach a reward before they forget you.
- The first reward is reachable by the second or third order. A reward that takes a year to earn changes nothing.
- Bonus points steer specific behavior: double points on a second order within 60 days, points for a photo review, points for trying a new category.
They fail when the earn rate is too thin to notice or when points stack on top of sitewide sales. If your promo calendar is already heavy, budget points as another discount layer.
Tiered programs
Tiers sort customers by spend over a qualifying period, often a rolling 12 months, and unlock better benefits at each level. The motivation is status and the fear of losing it, not just the reward.
Tiers earn their keep in the middle of the spend distribution. A customer just below a threshold has a concrete reason to add one more order; a customer far above the top threshold gets richer benefits for changing nothing.
To set thresholds:
- Pull 12-month spend per customer and sort it.
- Place the first threshold where a meaningful share of one- and two-order customers can reach it with one more purchase.
- Place the top tier where only your best customers sit, and give it low-cost benefits like early access and priority support rather than a bigger discount.
- Stop at three tiers. More levels confuse customers and multiply the email logic.
Decide the downgrade rule before launch. Dropping customers a tier after a slow year is fair but stings, so warn them ahead with a clear “spend $X by this date to keep your status” message, which gives slipping customers a concrete deadline to come back.
Paid memberships
Customers pay a monthly or annual fee for benefits such as free shipping, member pricing, free returns or exclusive products. It’s the only structure where the customer funds part of the program upfront.
The fee filters for customers who expect to buy often and gives them a reason to consolidate purchases with you, since they’ve already paid. The risk is the opposite of points: if members don’t get more value than they paid for, they cancel and remember the brand as a bad deal.
Run the math per member:
- Revenue: the fee
- Cost: benefits used (free shipping, member discounts, free returns), including on orders they would have placed anyway
- Gain: contribution from orders they place because they’re members
Paid membership fits high-frequency categories with a broad catalog and meaningful shipping costs. It fits poorly when customers buy one hero product twice a year.
Points, tiers and paid membership compared
| Points | Tiers | Paid membership | |
|---|---|---|---|
| Best for | Second and third orders, frequency | Lifting mid-value spend | Consolidating frequent buyers |
| Customer commitment | Low | Medium | Pays upfront |
| Main cost | Redeemed rewards | Top-tier benefits | Benefits used, especially shipping |
| Main risk | Becomes a permanent discount | Rewards big spenders who’d buy anyway | Members cancel if value falls below the fee |
| Needs | A short purchase cycle | A wide spread of customer spend | High frequency and a broad catalog |
Rewards beyond discounts: access, experiences and early drops
Discounts cost a dollar of margin for every dollar of value, teach customers to think in price, and look like every competitor’s program.
Access rewards cost less and are harder to copy:
- Early access to new products and limited drops, a day or two before everyone else
- Members-only products, colors or bundles
- Free returns or exchanges for top-tier members
- Priority support or a direct line to the team
- Samples of new products in the next order
- A say in upcoming products through polls or tester panels
- Invitations to events, workshops or live sessions
These work best for brands with a product calendar and a real fan base. Early access to a drop that sells out has high perceived value and costs almost nothing. Free returns and samples have real costs, so price them like any other reward. A points program can put access rewards in its redemption catalog, and a top tier can be defined almost entirely by access, which keeps the program from collapsing into a coupon.
Program economics and reward liability
A loyalty program pays for itself only through incremental contribution: extra and larger orders that wouldn’t happen without it. Everything else is cost.
A hypothetical example with round numbers: 10,000 enrolled customers spend an average of $300 a year, $3,000,000 in total. At a 5% reward rate, you issue $150,000 in points. If 70% get redeemed, $105,000 comes off future orders. Add $6,000 a year for software and the program costs $111,000. At a 45% contribution margin, it needs about $247,000 of incremental revenue to break even, roughly 8% of member revenue. That 8% is the minimum lift the holdout test has to beat, so write it down before launch.
| Cost line | What to include |
|---|---|
| Redeemed rewards | Face value of discounts; COGS and shipping for free products |
| Benefit costs | Free shipping, free returns, samples, events |
| Software | Loyalty app, plus added email and SMS volume |
| Operations | Support tickets, program management, abuse checks |
| Launch | Setup time, creative, on-site widgets and new email flows |
Reward liability
Unredeemed points are a promise of future value. Under the common revenue recognition standards (ASC 606 and IFRS 15), companies generally defer part of the revenue from a sale that earns points, then recognize it when the points are redeemed or expire. A growing points balance is a real liability, so talk to your accountant before launch, not after.
Two policies keep it under control. Expiry: points that lapse after a long period of inactivity stop the balance from growing forever, and the reminder before they lapse doubles as a reason to come back. Honest breakage assumptions: don’t build the business case on customers forgetting their points, because high breakage means the program isn’t motivating anyone.
Measuring whether the program is incremental
Comparing members with non-members proves nothing. People who join were already more engaged, so they’ll outspend non-members whether the program works or not.
The clean test is a randomized holdout:
- Split eligible customers at random before launch, including new customers at their first order. The holdout doesn’t see the program, its emails or its on-site widgets. This usually means tagging holdout customers and excluding the tag, so confirm your loyalty app supports it.
- Run it for at least two purchase cycles. For a product bought every two months, that’s four months minimum.
- Compare every assigned customer, not just those who joined. Measure revenue and contribution per customer across the whole test group versus the whole holdout. Dropping customers who never signed up brings the selection bias back.
- Subtract program costs from the test group’s contribution. A revenue lift below the break-even threshold from your economics model is a loss.
Compare repeat purchase rate at 60, 90 and 180 days as each window matures, revenue per customer, contribution per customer after rewards, and discounts as a share of revenue.
If the program is already live for everyone, hold out a random group from program marketing (emails, SMS, on-site reminders), or test the reward level: give a random half 3% back and the other half 5%. If the extra margin doesn’t buy extra orders, cut the rate.
The holdout is the easiest step to skip, and the one I insist on when I set up loyalty as part of ecommerce growth work. Without it, every program looks like a success.
Get it built
If your loyalty program is costing margin and you can’t tell whether it’s working, I’ll model the economics, pick the structure and set up the holdout test. The Growth Audit is $1,500 fixed and credited if we continue. See pricing or get in touch.