A Year in the Life of a Loyalty Program

Running a loyalty program is an annual cycle: plan and set goals at the start of the year, execute and test through the middle, prepare for and work the peak season, then measure the year honestly and use the result to plan the next one.

Many programs are treated as projects instead. They get attention at design and launch, then drift until someone asks whether the budget is still justified. The cycle is a way of avoiding that drift. Each stage feeds the next: the goals set in the first quarter are what the mid-year review checks, the tests run in spring decide what is scaled before the peak, and the evidence gathered across the year is what the renewal case rests on. A program that skips a stage usually feels it later, at the peak or at budget time.

This guide walks through a representative year, quarter by quarter. One caveat applies throughout: the calendar is illustrative. A retailer's year peaks at the holidays, a B2B program's may turn on a fiscal year or buying season, and a transit agency's on ridership patterns, so treat the rhythm as a model to adapt to your own peaks, not a fixed schedule.

A loyalty program's annual cycle ties four kinds of work to fixed moments: setting baselines, running tests, preparing for peaks and gathering renewal evidence.

Goals need a baseline to be measured against, and the baseline has to be recorded before the year's changes take effect. Tests need time to run and be read, so the larger ones belong in the quieter months; small communication tests, such as subject lines or send times, can run year-round. Peaks need lead time, so preparation happens a quarter ahead. Renewal needs evidence, so measurement has to be designed in from the start, not assembled at year-end. Research from Forrester and McKinsey shows where the pressure falls:

A representative year looks like this:

The timing shifts with the business. The table below is a starting point for adapting the calendar:

The first quarter sets the year's goals and baseline and brings back the members gained at the last peak.

The year starts with a review of the last one: what worked, what did not and what the numbers say. From that review come the goals for the year ahead. Good goals are few and measurable, and they connect to the business rather than to the program alone: active-member rate, retention against a comparison group and incremental margin, rather than enrollment or points issued. The guide to loyalty program KPIs covers which measures to set and how often to report them.

For consumer brands, the first quarter is also when the post-holiday retention job comes due. Members who joined at the peak need a reason to stay, and the forgetting problem Forrester measured in 2021 is a reminder that membership alone does not keep them. A short welcome-back sequence that shows each new member what they have earned and what is within reach is usually the first campaign of the year.

A program in its launch year runs this quarter differently. There is no prior year to review, so the baseline is customer behavior before launch, and the comparison group or holdout chosen at design is the main source of evidence. The first full year then sets the targets for the second.

The second quarter runs the campaign calendar and tests changes against a control group, so the program enters the second half knowing what works.

With the plan in motion, spring is usually the best time to test for programs that peak late in the year. Traffic tends to be steady, there is time to read results before the peak, and a failed test costs little. The method is simple: try a change with part of the audience, keep a comparable group as a control, measure the difference and roll out what works. McKinsey's 2021 review of loyalty levers makes the same point: "Building a rapid A/B testing capability is critical" to making use of the segments loyalty data creates, because it lets teams discard ideas that do not work and scale those that add value.

Good candidates for spring tests include:

• Communications: subject lines, timing, frequency and which members receive which message.
• Offers: bonus point levels, thresholds and the rewards that are featured.
• Experience: the enrollment flow, how balances and progress are shown, and redemption steps.
• Segments: whether a change that works for frequent members also works for occasional ones.

Each test needs a question, a control group, a sample large enough to tell a real difference from normal variation and a date to read it. A program too small for that, such as a B2B program with a few dozen dealers, can still test by comparing matched regions or partner groups, as long as the result is treated as directional rather than proven. Platforms can make this routine: BLOYL™, Brandmovers' loyalty platform, includes A/B testing against a control group and real-time dashboards.

An illustrative example: a retailer wants to know whether a double-points weekend brings lapsed members back. In April it sends the offer to most lapsed members and holds back a randomly chosen group that receives nothing. Four weeks later it compares the share of each group that made a purchase, and the margin each group produced after the cost of the bonus points. It checks the following month as well, because a bonus can pull forward purchases that would have happened later anyway. If the offer group came back at a higher rate but the extra margin does not cover the points, the test has still answered its question: the offer works but costs more than it earns at that level, so the next test tries a smaller bonus or a narrower audience, such as recently lapsed members only.

The quarter ends with the mid-year review against the Q1 goals. It is a short meeting with a clear output: which goals are on track, which are not and what changes in the second half.

The third quarter scales what the spring tests proved and builds the peak season a quarter ahead, because a strong fourth quarter is prepared, not improvised.

The third quarter acts on the mid-year review, which checked results against the goals set in Q1. Tests that won are rolled out to the full audience; tests that lost are retired; goals that are off track get a specific fix rather than more of the same. Three rules keep the review honest. A test counts as a win only if its difference against the control is larger than normal variation, using the sample size set before the test, and still pays once the cost of the offer is counted. A result too close to call is rerun or dropped, not scaled on optimism. And anything to be scaled for the peak is decided now, because a change agreed in October leaves little time for the preparation below. One caution: a spring result was measured on spring members, and the peak brings different shoppers with different motives, such as gift buyers and first-time joiners. Where a change will matter most at the peak, keep a small holdout running through the peak to confirm it still pays.

An illustrative example: suppose instead the April double-points test had lifted returns among lapsed members only slightly, with the gap narrowing in its final week. That result is too close to call, so it is rerun in July with a larger holdout group and a longer read, rather than scaled for November.

At the same time, preparation for the peak begins:

Promotions need the most lead time. The post on promotions compliance in 2026 covers the rules marketers should plan for. This guide is not legal advice; confirm specific requirements with counsel.

A loyalty program run as a cycle has a clear job in every part of the year. It plans and sets a baseline at the start, tests through the spring, scales what works and prepares for the peak in the summer, makes the most of the peak and measures the year honestly at the end. Then it does it again, with better goals and better evidence than the year before. The calendar will differ from one business to the next, but a steady rhythm of planning, testing and honest measurement gives a program a better chance of earning its budget year after year than any single campaign does.