The moment that earns a customer’s next purchase happens nowhere near the store.
A loyalty program keeps score at the checkout, because that is where a purchase leaves a record. When customers buy once or twice a year, the record goes quiet for months at a time. A brand that wants the next purchase has to earn it inside those months.
In the second week of January, a runner buys her first pair of shoes and signs up for a spring 10K. Over the next eleven weeks she runs 33 times and joins a Saturday group, and the store’s program sees none of it. To the program she is a member who has gone quiet, so it sends her the standard lapse offer, a discount code for socks. By April she needs a second pair, and the store that sold her the first one has no idea.
Nothing went wrong at the register. The program never heard what happened away from it, and that silence is where a brand that sells infrequently competes for the next sale.
Customer loyalty is a choice she keeps making
Customer loyalty is a customer’s habit of choosing the same brand again when other options are open, and of recommending it to others. It has two sides, and a loyalty team needs to measure both. Behavior shows in repeat purchases and in share of wallet. Attitude shows in preference and word of mouth. The guide to brand loyalty covers both sides in depth.
When purchases are months apart, both sides have months to change, and a program built around the receipt sees only the purchase. Most standard loyalty moves, from tiers to service to surveys, fire at a checkout or a service call. For a brand that sees a customer twice a year, that is two chances to hear from her.
Loyalty is not a record of what she bought. It is the habit that decides what she buys next. So the first decision for a loyalty team is which activity between purchases it can see, and which part of it is worth paying for.
Duolingo put a price of more than $50 million on protecting a habit
Duolingo is a clear case of a company choosing the habit over the sale, and it shows what the trade costs.
In its fourth-quarter shareholder letter, published in February 2026, Duolingo named the ads and upsells it had added to sell subscriptions. Duolingo called that friction part of the reason daily-user growth had slowed, so it decided to cut the friction back. It also put a price on the decision: more than $50 million in bookings, the value of the subscriptions and purchases it sells, that it expects to give up.
The decision rests on a view of where the subscription comes from. Duolingo treats the daily lesson as the thing that produces it. Its annual report says some learners use the product “for months or even years before they decide to subscribe.” That is why a public company told its shareholders it would give up bookings to protect a daily habit.
The company also shows how to measure a habit. It calls next-day return among its regular users one of its most important leading indicators, a number that moves before revenue does. By August 2026, daily users were growing faster. Duolingo credited product changes and marketing, plus a one-time event to revive lost streaks. A streak is the count of days in a row a learner has practiced.
A retailer has the harder job. Duolingo’s product is the activity, so every lesson lands in its own data, while a store that sells running shoes has a customer who runs out of sight. Before it can reward the activity, it has to decide how the activity will reach the member record.
Sportswear brands are paying for the run itself
Nike and adidas sell the gear, and both now spend on what customers do with it. On Nike’s June 2026 earnings call, chief executive Elliott Hill described the After Dark Tour. Right after, he said: “Being more local matters, it deepens connection, it builds loyalty.” A race gives a customer a date to train toward, and it gives Nike a place in her running between one pair and the next.
At adidas, the reward reaches past the shoe to the miles adiClub members log on Strava, the fitness-tracking app. That makes use of the product something the program pays for.
The sale happens in the store. The loyalty happens on the road. For a sporting goods brand, the choice is which part of the road to put on the record, from the race she trains for to the miles she logs in someone else’s app. A travel brand whose members book one trip a year faces the same choice about the months before the booking.
The weeks between purchases are a finance question too
Start with the data. A program that records only transactions knows the least about a customer at the moment she is deciding whether to come back. The runner’s store had one receipt. It did not have the eleven weeks that made her a runner.
Then the margin. When the only signal is silence, the only tool is a discount, and the sock coupon is what a program sends when it cannot tell a drifting member from a busy one. A loyalty leader can ask finance how much promotional spend went to members who were already coming back. A program with no answer is paying for purchases it did not cause.
The way to find out is a holdout, a randomly selected group of members kept away from a reward. The gap between that group and everyone else is what the reward caused. With that number, the program keeps funding the rewards that change behavior and stops funding the rest, including any reward that pays for miles she would have run anyway.
Then the partners. The activity often lives in someone else’s app, and the adidas link to Strava shows one way to bring it into the earn rules. A company that leaves those weeks to the coupon calendar is managing its most important customer decision with its bluntest tool. Loyalty earns its place as an enterprise growth engine in those weeks, not at the register, because the weeks are where the next sale is decided.
REI’s trips reached too few customers to pay for themselves
The counter-case is REI Co-op, which ran guided trips and classes for four decades. In January 2025, chief executive Eric Artz shut that Experiences business down. By his account, it had reached less than 0.4% of the co-op’s customers the year before and cost significantly more to run than it brought in.
REI’s instinct was not wrong. Its format was. A guided trip needs a guide for every small group, so its cost grows with every customer it serves. A streak, by contrast, lives inside an app that members open anyway. Activity programs fail when they reach too few customers or cost more than they bring in, and by Artz’s account REI’s did both.
REI kept part of the idea. By August 2026 its events page listed in-store workshops and run clubs. Since January 2026 it has sold trips guided by Intrepid Travel, with a discount for members. The design test for a travel or outdoor brand is the one REI’s old format failed: reach most members, at a cost per member that falls as more of them join.
Ten ways to build customer loyalty
These are the moves for a retailer, travel brand, or sporting goods company whose customers buy rarely. The first six are the basics most loyalty teams already run. The last four reach the customer in the weeks between purchases.
- Set the first reward within one purchase cycle. A reward that takes three purchases takes a once-a-year customer three years to reach. Put the first one within sight of her next order, and show members how close they are.
- Use tiers to recognize members by name and by status. Tiers give regular customers a goal and a visible place in the program. Set thresholds a customer who buys a few times a year can reach. The guide to loyalty tier strategy covers how to set the levels.
- Time each offer to what the member is doing now. A program that knows she is training for a race can invite her to a shoe fitting before race day. A program that knows only her last receipt sends a coupon for socks.
- Give the service team the member record. An agent who can see her history can fix a problem quickly and tell her what changed. For a customer who buys once a year, that one call is a large share of her contact with the brand.
- Ask for feedback and report what changed. Survey members after a first order or a return, then show them which changes came from their answers. The report gives the brand a reason to write between purchases.
- Pay members for bringing a friend. A referral credit pays for a new customer who arrives with a recommendation. The guide to referral programs covers setup and examples.
- Build a community around the activity. Give customers a group to join and a date to train for. In 2025, one in three runners in Nike’s After Dark Tour, a race series for women, were racing for the first time.
- Write earn rules for the activity as well as the spend. An earn rule decides what earns points. Since July 2026, adiClub members in the United States have earned points for every mile they run or walk and log on Strava.
- Measure the return visits between purchases, and clear the friction out of them. Duolingo calls next-day return among its regular users one of its most important leading indicators, and it cut back its own ads and upsells when they slowed daily-user growth.
- Put the activity where the whole member base already is. REI closed a trips and classes business that reached too few of its customers to pay for itself.
A GRAVTY® program can put the weeks between purchases on the member record
Give the January runner a second spring, with the store’s program running on GRAVTY®. This time the store links its program to the running apps members already use, the way adidas linked adiClub to Strava.
The program team writes an earn rule for logged runs in GRAVTY Visual Rules, the patented no-code rules engine. GRAVTY’s patented multi-dimensional behavior tracking builds her record from what she does as well as what she buys. Her runs sit next to her race and her January shoes.
In week six, the team uses Agentic AI Compass to compare members who log runs with members who only shop. It tries a shoe-fitting invitation in simulation first, and the live invitation goes out with a random group held back, so the team can see whether it brings runners in sooner. GRAVTY’s patented mass individualization shapes the offer around her mileage and her race date. AI-Trust watches earning in real time, so a 60-mile “run” logged at highway speed gets flagged as it happens.
The approach already runs on GRAVTY® outside sport. Loyalty Juggernaut describes it as loyalty that “shows up between bills”. At Magenta Moments, Deutsche Telekom’s program, partner offers and gamified experiences keep the brand present between invoices, and members have earned 22 million vouchers.
By April, the store knows she is a runner, and it knows her shoes are near the end of their first season. The fitting invitation reaches her before she starts looking for the second pair.
She remembers who showed up for the training
She sees a store that noticed she was training. It remembered her race and knew how many miles were on her shoes. To her, that is not a loyalty program. It is a store that pays attention.
She buys running shoes once or twice a year and runs in them about 150 times a year, and each run is a chance to earn the next pair.
The register records that she came back. The weeks before it are where she decided to.
The numbers behind customer loyalty between purchases
These are the figures behind this essay, set out for a loyalty team to benchmark against. Each one comes from the company’s own documents, linked in the tables.
Every activity reward creates a signal and carries a trap
Each activity reward tells the program something a receipt never shows, and each one has a way to go wrong.
| Move | The activity | The reward | The signal it creates | The trap |
|---|---|---|---|---|
| Count the streak | A daily or weekly habit | Visible progress and a way back after a lapse | Who is building a habit and who is slipping | Charging to protect the streak until it feels like a toll |
| Teach the skill | Learning to use the product well | Lessons and coaching plans | How far along she is, so what she needs next | Classes that reach a small share of members at a high cost per seat |
| Pay for the miles | The use itself, from runs to trips | Points for logged activity | How hard the product is working, so when it needs replacing | Paying for miles she would have run anyway |
| Host the group | Doing it with other people | Events and clubs, plus credit for bringing a friend | Who belongs and who trains alone | A community that fades when the budget moves |
| Mark the milestone | Finishing something hard | Recognition on the day it happens | What she has achieved, so what goal comes next | Milestones that expire along with the points |
The gamification software guide explains how to choose the software that runs streaks and challenges. Test each activity reward against a holdout, a randomly selected group of members held back from it. Then the program knows which rewards change behavior and can stop funding the rest.
Four brands put numbers on the weeks between purchases
| Source | What it reported |
|---|---|
| Duolingo, fourth-quarter 2025 letter | More than $50 million in forgone bookings in 2026; about 11% bookings growth, against the nearly 20% it says it could reach by running the business as in past years |
| Duolingo, second-quarter 2026 letter | 58.7 million daily active users, up 23%; 15.4 million learners revived a streak in June 2026 |
| Duolingo, 2025 annual report | About 43 million daily users on a streak of a week or more, end of 2025 |
| Nike, fourth-quarter fiscal 2026 call | Running up by double digits for five consecutive quarters, adding roughly $1 billion, which Hill credited to product and to investment on the ground in retail |
| Nike, After Dark Tour | More than 50,000 women at seven races in 2025, one in three racing for the first time |
| adidas and Strava, July 2026 | 3 adiClub points a mile for running, walking, and hiking, in the United States |
| REI Co-op, January 8, 2025 memo | 40,000 customers in 2024, less than 0.4% of the co-op’s; no profit even at its 2019 peak |
| REI Co-op, events and Intrepid trips | Workshops and run clubs in August 2026; Intrepid-guided trips since January 2026, 15% off REI Exclusive trips for members |
Seven numbers show whether customer loyalty is growing
A loyalty team that sells infrequently needs both sides of loyalty on one page, plus the weeks between purchases and one test for every reward.
| Number | What it shows | How the team uses it |
|---|---|---|
| Repeat purchase rate | The share of customers who buy again within a set period | Compare groups by the month they first bought, over a period long enough to cover one purchase cycle |
| Retention and churn rate | Who stays and who leaves over a period | Watch it by tier and by segment |
| Share of wallet | Her spending with the brand as a share of her spending in the category | A rising share means members pick the brand more often |
| Customer lifetime value | The margin a customer brings over the whole relationship | Decide how much each member is worth investing in |
| Net Promoter Score | How likely she is to recommend the brand | Read it as a trend line, next to repeat purchase |
| Activity rate between purchases | The share of members who do the activity each week | Spot drift before her next purchase is due |
| Holdout lift | The gap between rewarded members and a random group held back | Keep funding the rewards that change behavior |
Frequently asked questions
How do you improve customer loyalty?
A brand improves customer loyalty by getting the basics right, then rewarding what customers do between purchases. The basics are fair value, reliable service, and a first reward a customer can reach within one purchase cycle. For a brand whose customers buy once or twice a year, the larger opening is the activity the product serves, such as a runner training for a race, because the program can reward and measure it every week. Track how often members come back to it, and test each reward against a holdout.
What are the three R’s of customer loyalty?
The three R’s of customer loyalty are retention, related sales, and referrals. Harvard Business Review Press sets them out in its Pocket Mentor guide Focusing on Your Customer. It ties all three to one finding: the longer customers are loyal, the more profitable they become. Related sales means customers buy more of the range. Referrals means they bring in new customers. For a brand that sells infrequently, a program sees the behavior behind all three between purchases, while there is still time to act.
What are the four C’s of customer loyalty?
The four C’s of customer loyalty are captive, convenience-seekers, contented, and committed, the four types of loyal customer in a 2005 paper by Jennifer Rowley. The paper separates loyalty that is inertial from loyalty that is positive. It also finds that each type responds in its own way when something tempts it to switch. A program should know which kind of loyalty it holds before it spends to keep it.
What are the five stages of customer loyalty?
One common five-stage model runs from awareness to consideration, first purchase, repeat purchase, and advocacy, when a customer recommends the brand. The model’s weak spot is the long gap between purchases, which for a travel or sporting goods brand can run a year. A program that measures activity in that gap can spot a member drifting before her next purchase is due.
How do you build loyalty when customers buy rarely?
A brand builds loyalty with infrequent buyers by reaching them between purchases, through the activity the product serves. A runner buys shoes once or twice a year and runs in them every week, so the program can reward the runs she logs and the race she trains for. Test each reward against a randomly selected holdout group to see whether it changes what she does, and design the activity to reach most members at a falling cost per member.
Is customer satisfaction the same as customer loyalty?
No. Satisfaction describes how a customer felt about her last experience, and loyalty shows in what she does next: whether she comes back and whether she recommends the brand. A satisfied customer can still leave for a better price. A loyalty team should track both, and treat repeat behavior as the proof.
