November 19, 202614 min read

How to build a customer loyalty program for your ecommerce brand

To build a customer loyalty program, you set one goal, pick a structure that fits how often people buy, design rewards your margin can carry, connect the platform to your store and support tools, launch it to the customers you already have, and measure members against a comparable group that isn’t in the program.

Building customer loyalty this way is one part of a customer retention strategy. This guide is for online stores, whether you sell direct to consumers or across channels, and it covers every step of setting up an ecommerce loyalty program. It also covers what happens after launch, when members start asking your team and your AI where their points went.

The short version

  • Pick one goal (repeat purchase rate, purchase frequency, AOV, CLV, or churn) and record the baseline first.

  • Match the structure to purchase frequency: points for frequent buyers, tiers or VIP for a smaller group of high spenders.

  • Make the first reward reachable fast, and decide what happens to points on returns before launch.

  • Connect the loyalty platform to your store, your email and SMS, and the place customers ask for help.

  • Measure members against a holdout group, not against everyone else.

Set the goal before you pick a structure

Every decision after this one depends on what the program is supposed to change. If you’re working out how to create a customer loyalty program from scratch, start by choosing one primary metric: repeat purchase rate, purchase frequency, average order value, customer lifetime value, or churn. Treat the others as guardrails you watch, so a gain in one doesn’t hide a loss in another.

Pull your baseline before you design anything. You’ll want your current repeat purchase rate, the typical time between orders, AOV, and the share of revenue that comes from returning customers. Without those numbers, you can’t tell later whether the program did anything. Our post on customer lifetime value and the free CLV calculator show how to work out the lifetime side.

Of those numbers, purchase frequency shapes the design most. A skincare brand with a 60-day replenishment cycle can run a simple points program, because members come back often enough for a balance to feel like progress. A furniture brand with a two-year cycle can’t. Its members need tiers, perks, or non-purchase actions such as reviews and referrals to stay engaged between orders. Both examples are illustrative, so use your own order data to see where you sit.

Choose a program structure that fits how your customers buy

Most ecommerce loyalty programs use one of the structures below, or a blend of two.

Points-based programs

Members earn points per dollar spent or per action, then redeem them for discounts or products. Points suit brands whose customers buy often, because the balance grows fast enough to be worth watching. Square recommends that at least 80% of transactions be eligible to earn a point or star. That’s guidance for Square Loyalty sellers, not a universal benchmark, but the principle carries over: if most orders earn nothing, members stop paying attention.

Tiered rewards

Tiered rewards give members better perks as their spend or engagement grows. Mastercard describes a tiered program as one that combines many of the benefits of a rewards program, a member perks program, or even a paid subscription. For online stores, the perks that suit tiers include free or faster shipping and early access to product drops, and some stores add extended return windows.

VIP programs for your best customers

A VIP program is a narrower version of a tier: a small group of your highest-value customers gets something the rest don’t. Define VIPs from your own data, such as the last 12 months of spend or order count. Then decide whether membership is invite-only or earned by crossing a threshold.

The best VIP perks cost little but feel exclusive. Early access and a dedicated support line are common ones, and the cheapest of all is recognition in every conversation. That last one only works if your team and your AI can see who’s a VIP, which we cover in the section on preparing your support team and AI.

Some brands charge an annual fee for ongoing perks. This works when the perk has clear, repeat value, such as free shipping or member pricing, so customers can do the math and see they’re ahead. If you’re considering a paid tier, keep it simple and test it with a small group before you roll it out.

Referral and non-purchase rewards

Reviews, referrals, birthdays, and account creation can all earn rewards. These matter most for low-frequency brands, where members might go months between orders and need other reasons to stay in touch.

Points vs. tiered loyalty programs: which fits your store

If you’re choosing between the two, the table below sets out how they differ in what they reward, who they suit, and what they cost you.

Question

Points-based

Tiered

What it rewards

Each purchase or action

Cumulative spend or engagement over time

Works best when

Customers buy often

A smaller group buys much more than average

What members feel

A running balance toward a discount

Status and better treatment

Main cost risk

Discounts on orders that would have happened anyway

Top-tier perks that cost more than the extra spend

What support gets asked

Balances, missing points, expired points, codes

"What tier am I in?" and "why did I drop a tier?"

Many ecommerce programs combine points with two or three tiers. Whatever you choose, keep it to a structure a member can explain in one sentence.

Design rewards and earning rules that protect your margin

The first reward should be reachable within a customer’s normal first few orders. Square’s advice is that customers should be able to earn some reward in 30 days or less, and that rewards worth at least 10% of what a customer spent to earn them work well. Treat both as starting points to test.

Prefer rewards that cost you less than they’re worth to the member: free shipping, early access, free samples, and members-only products. Our post on why free shipping thresholds outperform discounts covers the margin side. Rewards can also move order size. A points threshold set just above your current AOV can nudge customers to add an item, which is one of the levers in our guide to loyalty rewards tied to spend. Watch for orders that are sized only to hit the threshold, and adjust if they pile up.

Decide what happens to points on returns and exchanges

Write the rules down before launch, because customers will ask and your team and your AI need to give the same answer. At a minimum, decide:

  • Whether points earned on a returned order are reversed.

  • Whether points spent on a returned order are refunded.

  • Whether exchanges keep their points.

  • What happens when a reversal pushes a member’s balance below zero.

Pick the answers that fit your return policy, then publish them on your program page so your team has one rule to point to when a balance changes.

Plan for points liability

Unredeemed points are a cost on your books until members use them or they expire, and your expiration rules affect both that cost and the number of complaints you get. Mastercard lists points liability among the costs of running a program. For how to handle it, see the breakage and reward liability section of our guide to how to calculate loyalty program ROI.

Technical requirements: platform, integrations, and data

Rules on paper only count if a platform can enforce them. Before you pick one, list what it has to connect to:

  • Your storefront and checkout, so members see points in the cart and on their account page.

  • Customer accounts, so balances follow the person across devices and channels.

  • Email and SMS, for balance, tier, and expiry messages.

  • Reviews and referral tools, if you reward them.

  • Your support tools.

Loyalty apps that online stores use include Smile.io, Yotpo, and LoyaltyLion. If you also sell in person, a POS-based program such as Square Loyalty may fit better, since it runs through the register.

Check whether the people and the AI answering customers can see loyalty data without switching tabs. Aim for one customer record that holds orders, loyalty balance, tier, and conversation history. When those live in four systems, every question turns into a scavenger hunt.

Prepare your support team and AI for loyalty questions

Once the program launches, members will ask about it, often in chat. Expect these:

  • Where are my points?

  • Why didn’t my points post?

  • When do my points expire?

  • Why didn’t my reward code work?

  • What tier am I in, and how do I reach the next one?

  • What happens to my points if I return this?

Each of these is a lookup against the loyalty platform, a rule you’ve already written down, or both. How well your team and AI handle them decides whether the program feels generous or like a source of friction.

Give your team the member’s loyalty status in the conversation

A VIP who has to explain their status has just been treated like everyone else. Good looks like this: balance, tier, recent point activity, and available rewards sit on screen next to the order history, and a team member can fix a missing-points problem in the same conversation, within limits you set.

With Gladly’s Yotpo integration, loyalty, review, and referral data sits on the customer profile, so team members see VIP tier, points balance, and tier progress during the conversation. They can adjust points with a confirmation step, up to a maximum an admin sets. Rules can also route conversations by VIP tier or points balance, for example, to the right team.

With Smile.io, team members see points, VIP status, and recent point activity in Gladly. Admins choose which actions are turned on, and point adjustments need confirmation and stay under a merchant-set maximum.

Let AI answer the routine loyalty questions

Balance, tier, earning rules, and expiry are lookups, so an AI with read access to the loyalty platform can answer them. Anything that changes a balance should need confirmation and a cap.

In Gladly, that works through the same integrations. Once you add Yotpo in Gladly AI settings, Gladly AI can use loyalty status, reviews, and referral data to personalize its answers. With Smile.io, Gladly AI can pull Smile data such as point transactions, earning rules, and VIP summaries to answer loyalty questions.

Use purchase history to keep rewards relevant

Support can fix the points, but personalization decides whether the next reward lands. A points reminder for a product someone already returned, or in a size they don’t wear, tells them nobody looked at their history.

That’s the idea behind agentic commerce: AI that works from the full customer record to help shoppers decide and buy. Gladly’s AI shopping assistant recommends from the customer’s full history, their sizes, their styles, and everything they’ve bought before, so the second and third purchase start from what it already knows.

Gladly helps us connect with high-intent shoppers in the moment, guide them to the right products, and drive immediate revenue, all while laying the groundwork for long-term loyalty.

Krystal Kay Cortez

Senior CX Operations Manager, Tecovas

Try Gladly AI right here

Wondering how Gladly AI handles loyalty questions? Ask it.

Get a practical answer in seconds, from the same Gladly AI our customers use.

Launch the program and enroll your existing customers

With the rules written and the systems connected, you’re ready to launch. Start with the people who already buy from you. Credit existing customers for past orders or give them a sign-up bonus, then announce the program to them.

Keep sign-up to an email address or phone number, and collect preferences later. Fewer fields make joining easier.

Brief your support team before launch day. Give them the rules doc covering earning, expiry, returns, and tiers, and load the same rules into whatever your AI answers from. Mastercard includes training in its launch step for the same reason: employees need enough training to help customers register and use the program.

If you plan to measure against a holdout group, a set of comparable customers who don’t get the program, soft-launch to one segment first and keep the rest out until you have a read.

Promote the program without turning it into a discount program

Tell customers about the program where they’re already paying attention: post-purchase flows, account pages, the cart and checkout, email and SMS balance reminders, and support conversations. A resolved question is a good moment to mention unused rewards. If you’re running a holdout group, limit these placements to the segment you launched to, so the control group stays clean.

The risk is that members come back only for the discount. Shep Hyken argues in Forbes that a loyalty program is mostly a marketing program, and that real loyalty is an emotional connection you build by pairing the program with a strong customer experience. Rewards still matter to shoppers: in Hyken’s 2024 State of Customer Service and CX research, sponsored by RingCentral, 76% of consumers said they’re more likely to return to a company with a good rewards program. That’s survey data on stated intent, not measured behavior.

Lead promotions with perks. “Early access for members” builds a different habit than “20% off this week.”

Track loyalty program KPIs and adjust

Keep the dashboard short. Start with these:

  • Enrollment rate: the share of customers who join.

  • Active member rate: the share of members who earn or redeem in a given period.

  • Redemption rate: how many earned rewards get used.

  • Repeat purchase rate for members versus the holdout group, which is how you’ll see customer retention move, plus purchase frequency, average order value, and customer lifetime value: the same metrics you picked your goal from.

Also track the share of support conversations about the program, and what they’re about. A spike in missing-points questions is a prompt to check your earning rules and your integration.

Review monthly for the first quarter, then quarterly. Adjust any threshold that’s too hard or too easy to reach, because either one can stall engagement.

For the ROI formula, holdout design, breakage, and the rest of the math, see our guide on how to calculate loyalty program ROI. It’s the better place to start once you have a few months of data.

Gladly logo against green background

Gladly Editorial

Gladly Editorial covers customer experience, AI, and agentic commerce for CX and ecommerce leaders. Our posts draw on Gladly's product team, customer conversations, and more than a decade of building service for brands like Crate & Barrel, Ulta Beauty, and Tumi.

Frequently asked questions