Retaining a customer costs between five and seven times less than acquiring a new one, and a 5% increase in retention rates can raise profits by between 25% and 95%, according to the classic research by Reichheld and Sasser at Bain & Company. Yet most organisations continue to invest the bulk of their budget in acquisition and leave loyalty to chance. This article treats retention as a measurable discipline: built on service level agreements that are honoured, on loyalty metrics interpreted in combination rather than in isolation, and on customer programmes that segment by real value, not by intuition.
The SLA as the contract of the experience
The Service Level Agreement (SLA) defines what the company promises and how it is measured. In customer service, the metrics that matter are not vanity metrics but those the customer actually perceives: First Response Time (FRT), resolution time, First Contact Resolution rate (FCR) and channel availability. A serious SLA also distinguishes between response time and resolution time: promising to respond within an hour means nothing if the issue takes two weeks to close.
FCR is probably the most underrated metric. Every time a case is reopened, operational costs multiply and perception degrades: industry studies estimate that customer satisfaction drops by approximately one percentage point for each percentage point that FCR falls. An SLA must escalate automatically — not manually — when a threshold is breached, and must provide for penalties or service credits when the customer is a B2B organisation and a breach has contractual implications.
It is worth distinguishing the SLA from the OLA (Operational Level Agreement) and from contracts with external suppliers (underpinning contracts). The SLA is what is promised to the customer; the OLA is the internal commitment between departments that makes fulfilling that promise possible. A four-hour resolution SLA is meaningless on paper if the internal backend team has no OLA obliging them to respond to the support team within two hours. The chain of commitments must be coherent end to end: the external promise can never be more ambitious than the sum of the internal capabilities that underpin it. Setting thresholds that operations cannot meet only generates systematic breaches and erodes trust.
NPS, CSAT and CES: three lenses, not one
The Net Promoter Score (NPS) measures loyalty with a single question — "How likely are you to recommend us?" — on a scale of 0 to 10. It is calculated by subtracting the percentage of detractors (0–6) from the percentage of promoters (9–10); passives (7–8) are not counted. The result ranges from −100 to +100. It is excellent for tracking trends, but on its own it does not explain why the customer scores as they do.
This is why it should be used in combination. CSAT (Customer Satisfaction Score) measures point-in-time satisfaction after a specific interaction — ideal for evaluating a closed support ticket — and CES (Customer Effort Score) measures how much effort it took the customer to resolve their issue. Research by Dixon, Freeman and Toman published in Harvard Business Review ("Stop Trying to Delight Your Customers") demonstrated that reducing effort predicts loyalty better than trying to exceed expectations. A customer who had to call three times will not stay, even if the third call was excellent.
Comparison of customer experience metrics
| Metric | What it measures | Scale | When it is asked |
|---|---|---|---|
| NPS | Loyalty and recommendation | −100 to +100 | Relational, periodic |
| CSAT | Satisfaction with an interaction | 1 to 5 (or %) | Transactional, after contact |
| CES | Customer effort | 1 to 7 | After resolving an issue |
| FCR | First contact resolution | % | Operational, continuous |
| Churn rate | Customer attrition rate | % | Monthly / annual |
The customer journey and moments of truth
Retention is not just about resolving issues well; it is about designing the entire relationship so that friction does not arise in the first place. Customer journey mapping traces the customer's path through every touchpoint — discovery, purchase, onboarding, support, renewal — and identifies moments of truth: those instants in which perception is decided all at once. Onboarding is the most underestimated stage: a customer who fails to extract value in their first few weeks (what in SaaS is called not reaching the activation milestone) has a disproportionately high churn probability, regardless of how good subsequent support may be.
The discipline of Customer Success was born precisely to manage this proactively rather than reactively: instead of waiting for the customer to open a ticket, a dedicated manager monitors usage signals (a health score combining usage frequency, feature adoption and open tickets) and intervenes before dissatisfaction translates into cancellation. A falling health score is an early warning signal worth far more than any subsequent survey, because by the time the survey arrives the customer has already decided to leave.
The economics of retention: from CLV to the LTV/CAC ratio
For loyalty to be more than an act of faith, it must be quantified. Customer Lifetime Value (CLV or LTV) estimates the total margin a customer will contribute over the entire duration of their relationship with the company; in its simplest form it is calculated as average margin per period divided by the churn rate. Reducing churn extends the lifetime and multiplies CLV in a non-linear way: lowering monthly churn from 5% to 4% does not improve CLV by 20% — the actual gain is far greater, because each customer stays considerably longer.
The indicator that closes the circle is the LTV/CAC ratio (customer lifetime value against acquisition cost). A healthy business targets a ratio of around 3:1 and a payback period on acquisition investment of under 12 months. When retention is low, CAC is never recovered and the company falls into the leaky-bucket trap: it acquires at the top at the same rate it loses at the bottom, spending ever more on acquisition to achieve the same net growth. This is why retention is not a customer service matter — it is a primary financial lever.
Value-based segmentation: VIP programmes that are not decorative
A poorly designed loyalty programme gives away margin to customers who would have stayed regardless. The correct approach starts by segmenting the portfolio by Customer Lifetime Value (CLV) and by churn risk. RFM analysis (Recency, Frequency, Monetary value) remains the most robust and cost-effective tool for prioritisation: it identifies customers who buy frequently, recently and in high amounts, and separates those who have been inactive for months and are one step away from leaving.
The VIP programme should offer asymmetric benefits to the highest-CLV segments (a dedicated account manager, a premium SLA, early access to new features) and simultaneously activate recovery campaigns for high-value, low-recency customers before they cancel. Reacting to a cancellation already in progress is too late and too expensive.
GDPR: loyalty is also data processing
Every loyalty programme processes personal data and falls squarely within the scope of the General Data Protection Regulation (GDPR). Three principles are non-negotiable: a clear lawful basis (explicit consent for profiling and commercial communications), data minimisation (not collecting more data than necessary for the declared purpose) and the right to object to profiling, which the customer must be able to exercise as easily as they enrolled. Data protection authorities have sanctioned points programmes that reused purchase data for undisclosed purposes. Trust — the substrate of loyalty — evaporates instantly with a breach or a public enforcement action.
Common retention errors
- Measuring NPS without an open follow-up question. The score without the "why?" is not actionable.
- Rewarding tenure instead of value. Linear discounts that erode margins without actually slowing real churn.
- SLAs without automatic escalation. Breaches are only detected once the customer has already complained.
- Over-surveying. Survey fatigue lowers response rates and skews the sample towards the extremes.
- Ignoring GDPR consent for profiling. Legal and reputational risk that nullifies the programme's return on investment.
Frequently asked questions
What NPS score is considered good?
It depends on the sector. A positive NPS (above 0) already indicates more promoters than detractors; above +50 is considered excellent. What matters is your own trend and your sector benchmark, not an absolute universal figure.
NPS or CSAT?
Both. NPS measures the long-term relationship; CSAT measures each individual interaction. Using only one leaves half of the experience uninstrumented.
How is churn rate calculated?
Customers lost in the period divided by customers at the start of the period, expressed as a percentage. In SaaS it is usually measured monthly and complemented by revenue churn (MRR lost), which penalises more heavily the loss of large accounts.
Does a points programme require GDPR consent?
Yes, particularly for profiling purchasing behaviour and for sending commercial communications. Joining the programme does not equate to consenting to any use of data; each purpose must be disclosed and given a separate lawful basis.
Conclusion
Retention is not won through isolated "delight" gestures but through a system: SLAs that are met and self-escalate, metrics that are read together rather than worshipping a single number, and programmes that direct resources at the customers who genuinely affect the bottom line. The finding that should keep any leadership team awake is that effort, not enthusiasm, predicts loyalty: customers leave because of the friction they experience, not because of a lack of gifts. At Summum Consultoria we design the full customer experience dashboard — from FRT to CLV — and value-segmented loyalty programmes, with GDPR compliance built in from the first data point rather than bolted on afterwards.