What is loyalty marketing?
Loyalty marketing, also called loyalty marketing or customer loyalty marketing, brings together strategies to develop the relationship after acquisition. It connects experience, communication, service, recognition and offers to increase the likelihood of continuity and repeat purchase. A points or cashback program can support the strategy, but it is not synonymous with it. To assess results, the company must define behavior, period and audience, track retention, frequency, margin and customer value, and test whether the actions actually caused the observed change.
Loyalty, retention and repeat purchase are different concepts
Using the terms as synonyms creates confusing goals:
- Loyalty: may include preference, emotional bond and willingness to maintain the relationship. It has attitudinal and behavioral dimensions.
- Retention: customers remaining in a relationship over a period, according to an operational definition of activity.
- Repeat purchase: a new transaction made by the same customer. It can occur due to convenience, price or lack of alternatives, without proving preference.
A review of consumer loyalty shows that attitudes, recommendation and repeat purchase are not equivalent and can predict different behaviors. Therefore, combine research, behavior and economics instead of looking for a single metric.
In subscription businesses, retention can mean an active contract. In retail, the company must define a window consistent with the purchase cycle. An annual purchaser should not be considered lost after 30 days.
Start with a behavior and a hypothesis
A measurable strategy starts with a specific hypothesis. Example:
Hypothesis: reminding customers with near-expiry credit will increase valid use of the balance before expiration, without reducing the defined minimum margin.
This statement separates:
- audience: customers with near-expiry credit;
- action: reminder;
- primary outcome: balance usage;
- window: before expiration;
- constraint: minimum margin.
“Increase loyalty” is not a sufficient hypothesis. Define which behavior will be observed and what must not get worse.
Segmentation based on need and behavior
Segmentation helps avoid identical messages for different situations. Possible criteria include:
- recency, frequency and purchase value;
- stage of the journey;
- product, category, unit or channel;
- participation and benefit usage;
- consent and communication preference;
- service issues or recent feedback.
Segment only with data that is necessary, reliable and used for a legitimate purpose. Avoid sensitive inferences or groups so small that analysis and data protection are hindered.
Precise segmentation needs rules for entry, exit and update. “Inactive customer,” for example, should have a window tied to the business’s real cycle.
Journey and relationship moments
Map the moments when the company can deliver value or fix friction:
- Onboarding: registration, expectations and consents.
- First value: successful product use or first benefit.
- Post-purchase: confirmation, delivery, guidance and support.
- Continuity: use, repurchase or renewal in the expected cycle.
- Risk: failure, complaint, delay or reduction in activity.
- Recognition: milestone in the relationship that is relevant.
- Exit: cancellation, expiration or termination with appropriate handling.
Not every moment calls for a promotion. Status information, support and resolution can be more important to the relationship than a discount.
Purposeful communication and control
For each message, define event, audience, objective, channel, frequency and outcome. Loyalty marketing communication should explain the value and expected action without hiding restrictions.
Essential precautions:
- respect consent, legal basis and opt-out;
- limit frequency across competing campaigns;
- do not send offers incompatible with the customer’s history;
- disclose validity, minimums and benefit conditions;
- measure conversion and margin, not just open rate or click;
- keep a control group when the decision requires causality.
Experience is part of the loyalty strategy
Communication does not compensate for a poor product, incorrect billing or unresolved service. Marketing, product, operations and support need to share causes of complaints, churn and repurchase.
Use qualitative feedback to explain metrics. If retention falls, investigate cohorts, channels and reasons. If redemption is low, check whether the reward is desirable, attainable and easy to use. If breakage rises, do not assume this is positive: it may indicate accounting revenue, but also a lack of perceived value or difficult rules.
A loyalty program is a tool, not the whole strategy
A loyalty program organizes recognition and rewards through points, cashback, stamps, tiers, subscription or perks. It can generate data and communication moments, but its outcome depends on the value proposition and the experience.
Choose the model by behavior and economics. The comparison of loyalty program models details the differences. This article focuses on strategy and measurement.
Metrics and formulas table
| Metric | Reference formula | Required definition | Main limitation |
|---|---|---|---|
| Customer retention | ((Customers at end − new customers) ÷ customers at start) × 100 | Who is active, period and cohort. | Can hide differences in value and behavior. |
| Customer churn | (Customers lost ÷ customers at start) × 100 | What characterizes loss and in which window. | Is not always the simple complement of retention if bases differ. |
| Purchase frequency | Valid purchases ÷ purchasing customers in the period | Valid purchase, unique customer and period. | The average can hide segments and seasonality. |
| Repurchase rate | Customers with a new purchase ÷ eligible customers × 100 | Previous purchase, window and eligibility. | Varies greatly according to cycle and cohort. |
| Program participation | Registered participants ÷ eligible customers × 100 | Invitation, eligibility and valid registration. | Registration does not mean activation or value. |
| Program activation | Participants with first valid action ÷ registered × 100 | Which action demonstrates first value. | Initial action does not guarantee continuity. |
| Redemption rate | Rewards redeemed ÷ rewards made available × 100 | Unit, window, issuance and availability. | Counting points, vouchers or people produces different results. |
| Breakage | Rights or rewards expected not to be exercised ÷ rights issued | Validity policy and estimation method. | Requires accounting treatment and history; it is not just “expired points”. |
| Contribution margin | Attributed revenue − attributed variable costs | Which revenues and costs belong to the action. | Does not necessarily include fixed costs. |
| LTV/CLV | Present value of the customer’s expected future margins | Horizon, margin, retention, discount and costs included. | Is an estimate sensitive to assumptions. |
| Cost per active participant | Total program cost ÷ active participants | Costs included and definition of active. | Does not reveal incremental benefit alone. |
The formulas are references. Document the version used by the company and keep the same definition when comparing periods. Different platforms may calculate the same metric in different ways.
How to interpret retention, frequency and churn
The retention rate requires separating existing customers from new ones. The reference formula is also presented in Microsoft Dynamics 365. In retail, the definition of a retained customer should respect the expected interval between purchases.
Churn is more straightforward in subscriptions: Stripe describes it as customers lost divided by customers at the start of the period. Still, tools may include new customers in the denominator or use revenue instead of people. Declare the variant.
Frequency corresponds to the average number of purchases per buyer in a window. Compare cohorts and equivalent periods to avoid attributing seasonality to a campaign.
Participation, activation, redemption and breakage
Participation measures entry; activation measures the first relevant behavior; redemption shows reward usage. A large registration base with little activation is not necessarily healthy.
Breakage is the term used for rights that are not expected to be exercised. IFRS 15 addresses recognition of unexercised rights in contracts with customers. The accounting application to points, credits or benefits must be validated by the responsible professionals.
Very low breakage can raise cost; very high breakage can reveal an inaccessible reward. Evaluate economics and experience together.
Margin and LTV: revenue is not economic value
A campaign can increase purchases and reduce margin. Therefore, track incremental revenue, discounts, reward cost, communication, technology, service, fraud and operations.
LTV or CLV estimates the economic value of the relationship over a horizon. More consistent models use future margins and discount, but depend on assumptions about retention and behavior. Do not turn a projection into fact.
Record:
- whether LTV uses revenue, gross margin or contribution;
- horizon and time unit;
- retention or churn assumed;
- discount rate;
- acquisition and relationship costs included;
- difference between observed and projected value.
Metric, formula, hypothesis and benchmark are not the same thing
- Metric: observed concept, such as retention.
- Formula: operational calculation rule.
- Hypothesis: expected effect of an action.
- Benchmark: external reference obtained in a defined context.
A goal should not be presented as a benchmark without source. Even published benchmarks may not be comparable by industry, country, margin, channel, period or definition.
Tests and attribution: how to avoid improper causality
If frequency increased after a campaign, that does not prove the campaign caused the increase. Seasonality, price, media, availability and changes in the base can explain the result.
To estimate impact:
- define hypothesis and main metric before launch;
- record baseline and concurrent events;
- use a comparable control group when possible;
- assign customers before exposure and avoid swapping between groups;
- track margin and adverse effects;
- use a sufficient window for the purchase cycle;
- document sample size, exclusions and uncertainty;
- repeat the test before generalizing.
When there is no experiment, describe the result as an observed association and present alternative explanations.
A useful dashboard connects behavior, experience and economics
Organize tracking into three layers:
- behavior: retention, frequency, repurchase and churn;
- program and experience: enrollment, activation, redemption, breakage and feedback;
- economics: margin, cost per participant, LTV and incremental return.
Avoid dozens of indicators without an associated decision. Each metric should have an owner, source, frequency, quality threshold and possible action.
Loyalty marketing requires strategy before the tool
Loyalty marketing combines experience, communication and recognition to sustain valuable relationships. The rewards program is a possible tool; segmentation and automation only help when data, purpose and measurement are clear.
If you have already defined behavior, model and metrics, learn about the loyalty program solution from Smartbis. Evaluate in the demo which features, integrations and reports meet the definitions used by your business.