A retention strategy is the plan through which a company seeks to keep the customers it has acquired over time, addressing the causes of churn and verifying the results. It defines which customers to observe, which problems to address, who is accountable and how to measure the economic sustainability of the interventions.
The term retention is also used for keeping employees. Here we focus on customers: renewals, relationship continuity and repeat purchases, depending on the business model.
The scope must be stated before choosing initiatives. A subscription, an ongoing supply arrangement and a one-off purchase require different definitions and observation periods.
Retention, loyalty and customer loyalty-building: the differences that matter
Retention looks at whether customers stay, according to defined criteria and timeframes. Loyalty also concerns preference and attachment to the brand. A customer may keep buying out of convenience, availability or switching costs without having any particular preference for the company.
Customer loyalty covers the activities that nurture the relationship over time. A retention strategy selects interventions based on an observed problem and verifies their effects on customers and on economic results.
In everyday language the dimensions overlap. Distinguishing them helps to interpret the data: staying, on its own, does not prove satisfaction, and a statement of approval does not guarantee renewal.
How the retention rate is calculated
The Customer Retention Rate, or CRR, measures how many customers from an original base are still active at the end of the period observed.
Before calculating, we define active customer, period and retention rule. In a subscription service we can observe active contracts; for non-recurring purchases, the window must match the expected purchase frequency. A month without purchases does not prove churn if the customer normally buys once a year.
The direct calculation is:
CRR = customers from the original base still active at the end / customers active at the start × 100
The shortened formula is:
CRR = (E − N) / S × 100
Where:
- S is the number of customers active at the start.
- E is the number of active customers at the end.
- N is the number of new customers still active at the end.
The shortened formula works when the final base consists of the remaining initial customers and the new customers counted in N. Reactivations of customers outside the initial base should be counted separately, otherwise they may inflate the result.
It is also necessary to document how customers who interrupt and resume the relationship within the same period are treated. Being active at the end and never having lapsed are different conditions.
Which KPIs to read alongside retention
Customer retention should be read alongside behaviour, revenue and costs. The indicators answer different questions and must rely on documented definitions.
| Indicator | What it measures | How to interpret it |
|---|---|---|
| Customer Retention Rate | How many customers from the original base are still active at the end. | Excludes the contribution of new customers. |
| Customer churn rate | How many customers in the starting base are lost under the rule adopted. | It complements CRR only with the same base, window and definition of retention. |
| Repeat purchase rate | For example, the share of customers with at least two purchases among those who bought within the chosen window. | Always specify numerator, denominator and period: it is not automatically equivalent to CRR. |
| Net Revenue Retention | How recurring revenue from the same starting base changes, taking into account expansion, contraction and churn. | It can exceed 100% even if some customers churn. It excludes revenue from new customers. |
| Relationship margin and CLV | The economic contribution of the relationship, observed or estimated over time. | To speak of profit, margins and costs must be considered as well as revenue. |
| NPS and satisfaction | Stated likelihood to recommend and experience ratings. | They help to form hypotheses about causes, but do not directly measure how many customers stay. |
In recurring-revenue models, NRR compares the final revenue attributable to the same starting base with the opening revenue, over a consistent time frame. It should not be compared directly with the customer retention rate.
CLV also needs care: average purchase × frequency × relationship duration is a simplification of expected revenue. Estimating economic value requires making explicit the assumptions, margins and costs taken into account.
There is no retention rate that is good for every company. The first useful comparison is between your own cohorts, matched by segment, period and relationship tenure. An external benchmark can only be used if it measures the same phenomenon with comparable criteria.
How to build a retention strategy
To build the plan, we start from a segment and an observable problem. Every intervention must have an owner, an expected result and a cost check.
Define the scope and the expected result
Choose which customers to observe, at which stage of the relationship and for how long. The objective may concern renewals, continuity of use or repeat purchases, keeping these results distinct.
Diagnose where the relationship breaks down
Compare customers who stay with those who leave, considering tenure, offering purchased, acquisition channel and reported issues. Complaints, low usage or delays can guide the analysis, but on their own they do not prove the cause of churn.
Choosing the intervention and testing its sustainability
Link the action to the problem identified. A customer who cannot use the service needs a different intervention from one facing unreliable deliveries. Also estimate costs, discounts and the resources required: retaining a customer can become hard to sustain if the cost exceeds what the relationship contributes.
Assign responsibilities and deadlines
Senior management approves priorities and resources. A plan owner coordinates the functions involved; sales, customer service, product and administration step in depending on the problem. Finance validates the economic reading. In an SME several responsibilities may sit with the same person, but they must remain explicit.
Review the outcomes and decide how to proceed
Establish before launch who monitors the indicators, when they review them and which results justify continuing, modifying or stopping. Where possible, compare the group involved with similar customers not affected by the intervention, checking for initial differences, seasonality and changes to the offer.
Which levers to choose depending on the problem
| Problem identified | Intervention to assess | Test |
|---|---|---|
| The customer does not reach the first expected result. | Customer onboarding: instructions, setup and usage checks. | Actual usage and relationship continuity. |
| Problems remain unresolved. | Resolution owner, handovers and customer updates. | Timescales, reopened issues and subsequent churn. |
| The service is little used or loses its usefulness. | Customer Success oversight of usage, expected outcomes and renewal. | Adoption, issues and renewals, read separately. |
| Communications and offers prove irrelevant. | Segmentation and personalisation based on relevant information. | Subsequent behaviour, beyond opens and clicks. |
| An incentive could support repurchase. | Targeted benefits or loyalty programme. | Additional purchases and margin after the cost of the incentive. |
| Promise and service do not match. | Review of offering, communication and hand-offs between channels. | Reduction in issues and the trend of the relationship. |
A CRM can organise data and activities as volumes and complexity grow. At the outset, a shared register, assigned responsibilities and an update process may be enough. Automations must support a defined intervention; any churn predictions should be treated as estimates, checking their errors and cost-effectiveness.
An example of a plan and of monitoring results
Consider a hypothetical case: a subscription services business observes one quarter and counts as active those customers with a current subscription. The starting base comprises 200 customers.
At the end of the quarter, 180 of those customers are still active. Add 40 new customers still active at the close, and the final base is 220 customers.
CRR = (220 − 40) / 200 × 100 = 90%
The total base grew by 10%, but 10% of the starting customers did not stay. Acquisition offset the losses: reading only the final total would have hidden the problem.
From measurement to decision
The plan owner analyses the 20 customers who left, distinguishing by relationship stage and available reasons. If recurring difficulties emerge in getting started with the service, they may propose an intervention on onboarding.
Senior management approves resources and verification criteria. The service manager implements the intervention; the plan owner compares performance against a comparable group, also monitoring costs and usage.
The assessment considers customers retained and margin after the cost of the intervention. A subsequent improvement is not automatically attributed to the action: seasonality, changes to the offering or differences between groups may have contributed.
The mistakes that make retention unsustainable
- Confusing growth with retention. New customers can offset departures and mask a deterioration in the original base.
- Observing an unsuitable window. A period too short relative to the purchase cycle can classify as lost customers who simply had no reason to return yet.
- Acting before diagnosing. A discount does not necessarily solve product, delivery or support problems.
- Measuring revenue alone. Revenue retained through steep discounts and rising service costs can erode margin.
- Reading signals as certainties. Stated satisfaction, low usage and churn predictions call for verification, not automatic conclusions.
- Attributing every result to the brand. Consistency and trust can contribute to the relationship alongside quality, price, availability and service.
- Constantly changing definitions. Active customers, reactivations and churn must be counted using documented rules so that periods remain comparable.
Bliss’s view on Retention Strategy
At Bliss, we believe a customer who stays deserves the same attention as one who leaves. They may keep choosing the company because they recognise its value, or because switching supplier takes time, carries risk or would disrupt an important process. In the retention rate these situations can look identical. For those deciding where to invest, they have very different implications.
An apparently stable customer base can therefore conceal a relationship that is weakening. The contract is renewed, but every negotiation requires a larger discount. Purchases continue, while complaints rise or willingness to consider the company’s other offerings declines. These signals warrant closer examination: retention may depend on conditions that a competitor, a technological shift or a new need could remove.
This is where our strategic reading begins. We seek to understand what led the customer to choose the brand and whether that reason is still confirmed by the experience. A promise of simplicity must be tested in the steps the customer has to take; a promise of reliability in deliveries, support and the handling of the unexpected. When a gap emerges, the task is to identify the decision that can narrow it and who has the authority to take it.
Retention thus also becomes a test of how well the positioning holds. It helps leadership distinguish the relationships on which to build growth from those sustained through ever more costly concessions, and decide whether to act on the offering, on the service or on the expectations set by communication. The result should then be measured in the continuity of the relationship and in the value that remains after the cost of maintaining it.
Request a conversation
Define relationship priorities
If customer losses call into question the brand promise or the direction of the offering, the discussion starts from the available evidence. We define which problems to address, which decisions rest with senior leadership and by what criteria their effects should be assessed.
Domande frequenti
What does retention mean in a business context?
It refers to keeping customers or employees over time, depending on the context. In customer retention, one observes how many customers continue the relationship within a period and against defined criteria, such as an active subscription, renewal or purchase.
How is the retention rate calculated?
Divide the number of customers in the initial base still active at the end by the number of initial customers and multiply by 100. The shortened formula is (final customers − new customers still active at the end) / initial customers × 100, with reactivations outside the base checked separately.
What is the difference between retention and loyalty?
Retention measures continued custom according to defined criteria. Loyalty also encompasses preference and attachment to the brand. A customer may stay out of convenience or contractual obligations without having any particular preference for the company.
Which KPIs are needed to measure retention?
CRR measures customer retention. It should be read alongside churn, repeat purchases, margin and cost to serve. In recurring-revenue models, NRR is also useful. Satisfaction and NPS help interpret the relationship, but they do not replace the measure of customers retained.
How do you build a retention strategy?
Customers and period are defined, churn is analysed and an intervention linked to a verified problem is chosen. The plan assigns responsibilities, resources and control criteria, taking account of both results and costs.
What is a good retention rate?
There is no universal threshold. The same metric must be compared across comparable periods, segments and relationship models. The first reference point is the trend of your own cohorts; external benchmarks are only useful when definitions and method are comparable.

