Taux de rétention des clients

Client retention rate is the percentage of clients who continue their relationship with an institution over a defined period. In lending, it usually measures the share of borrowers who take a subsequent loan after fully repaying the previous one, rather than exiting after a single cycle.

Its inverse is the dropout rate (or churn rate): the proportion of clients who leave.

Retention is one of the most economically significant metrics a lender tracks, because almost every cost of a lending relationship is front-loaded — acquisition, group formation, first appraisal, training, KYC — while the margin is earned across cycles. A borrower who leaves after one loan may never have covered the cost of being onboarded.

How to Calculate Client Retention Rate

There is no single agreed formula, and this is the most common source of confusion when retention figures are compared between institutions. The three methods below all produce a "retention rate" and all give different answers on the same portfolio.

Method 1 — Standard retention formula

Retention rate = ((E − N) / S) × 100
where S = clients at start of period, N = new clients acquired during the period, E = clients at end of period

This isolates continuing clients by removing new acquisitions from the closing balance. It is the general business standard and works well where the client base is measured as a stable stock.

Example: 4,200 clients at the start of the year; 1,500 new clients acquired; 4,800 at year end.

Retention = ((4,800 − 1,500) / 4,200) × 100 = 78.6%

Method 2 — Borrower retention rate (microfinance convention)

Borrower retention = Active borrowers at end of period / (Active borrowers at start + New borrowers during period) × 100

Using the same figures:

4,800 / (4,200 + 1,500) × 100 = 84.2%

This method is common in microfinance benchmarking. It is more forgiving than Method 1 because new acquisitions sit in the denominator rather than being netted out of the numerator.

Method 3 — Loan cycle method

Retention = Repeat loans disbursed in period / Total loans disbursed in period × 100

This measures renewal behaviour directly rather than inferring it from client counts, and it is the method most closely aligned with how progressive lending actually works. Its weakness is that it is sensitive to disbursement volume: a burst of new-client acquisition will depress the ratio even if no existing client left.

Method 4 — Cohort retention

Track a defined intake group over time and measure how many remain at each cycle.

CohortCycle 1Cycle 2Cycle 3Cycle 4Cycle 5Clients remaining1,000720590505450Retention vs prior cycle—72%82%86%89%Cumulative retention100%72%59%51%45%

Cohort analysis is the most informative method because it shows where clients leave. The pattern above — heavy loss between cycles 1 and 2, then progressive stabilisation — is typical, and it tells you the intervention point is the first renewal, which a single portfolio-level percentage would never reveal.

Practical rule: pick one method, define it in writing, and use it consistently. A retention figure without its formula attached is not comparable to anything.

Why Client Retention Matters

Acquisition cost amortisation. Origination, group formation, training, KYC and first appraisal are largely one-off. Every additional cycle spreads those costs across more revenue.

Lower credit risk. Repeat borrowers consistently perform better than first-cycle borrowers, because the institution has observed their behaviour and the borrower has accumulated relationship value they do not want to lose. Retention therefore improves portfolio quality, not just revenue.

Larger average loan size. Under progressive lending, retained clients move up the ladder. Retention is what converts a portfolio of tiny first loans into one with viable average balances.

Cheaper underwriting. Renewal appraisal for a known borrower with repayment history costs a fraction of a first-time appraisal.

Client lifetime value. Retention is the primary input to CLV. Small changes compound: raising cycle-to-cycle retention from 70% to 80% roughly doubles the expected number of cycles per client.

Signal quality. Dropout is honest feedback. Clients rarely complain — they simply do not come back. A rising dropout rate is often the earliest indication that pricing, loan sizes, service quality or a competitor's offer has shifted.

Common Measurement Pitfalls

Counting dormant clients as active. A client with a zero balance who has not transacted for months is not retained. Define "active" explicitly — an outstanding loan, a transaction within a set window, or a savings balance above a threshold — and apply it consistently.

Ignoring rest periods. Borrowers frequently pause between cycles for seasonal or personal reasons and return later. Measuring retention over too short a window classifies these clients as lost. Define a return window (commonly 3–6 months after settlement) before counting a client as exited.

Confusing forced and voluntary exit. A client whose group collapsed, who was declined, or who was written off has exited for a very different reason than one who chose to leave. Aggregating them hides the actionable signal.

Group-level versus member-level measurement. In group lending, a group can survive while losing members, or dissolve while all members migrate to individual products. Both need separate tracking.

Multiple concurrent loans. Where a client can hold more than one loan, loan-count-based methods overstate the client base. Retention must be measured at client level, not loan level.

Period sensitivity. Annual, quarterly and monthly retention rates are not convertible into one another in any intuitive way, particularly for short-tenor products where a client may complete three cycles in a year.

Survivorship in the denominator. Clients still mid-loan have had no opportunity to leave. Including them inflates retention. The cleanest denominators contain only clients who reached a renewal decision point.

What Drives Client Dropout

Loan size ceiling. The most common reason good clients leave. A borrower who has outgrown the maximum available amount will go to a lender who can serve them — taking the credit history the institution spent several cycles building.

Slow renewal turnaround. A borrower who repays on Friday and cannot access the next loan for three weeks has a working capital gap. Competitors compete precisely on that gap.

Time cost of meetings. Compulsory weekly attendance is a real, recurring cost for a trader. It never appears in the interest rate but is frequently cited in exit interviews.

Rigid product terms. Repayment frequency, tenor or start date that does not match the borrower's cash cycle — particularly acute in agriculture and seasonal trade.

Pricing. Usually a smaller driver than assumed, but it becomes decisive once a competitor offers materially better terms for an equivalent product.

Group failure. Under joint liability, a defaulting member can force out several good payers. This is one of the largest sources of avoidable dropout in group portfolios.

Staff conduct and turnover. Relationships in field lending are often with the loan officer, not the institution. Officer turnover is a measurable predictor of client loss in the affected portfolio.

Over-indebtedness. A borrower servicing several lenders may withdraw deliberately, which is a healthy exit and should be recorded distinctly.

Business closure or relocation. Genuine attrition unrelated to service quality, but it should still be counted and quantified rather than assumed.

How to Track Retention Usefully

• Segment it. Portfolio-level retention is nearly useless on its own. Break it down by cycle number, branch, loan officer, product, group versus individual, sector and gender. The variance between segments is where the decisions are.
• Capture exit reasons at source. A mandatory dropout reason code at closure, with a small, well-defined list, produces better data than any retrospective analysis.
• Run exit interviews on a sample. Reason codes recorded by staff are systematically biased toward reasons that do not reflect on staff. A small independent sample corrects this.
• Watch cycle 1 to cycle 2 specifically. This is where the largest losses occur and where intervention has the highest return.
• Track retention against loan size growth. If retention is high but average loan size is flat, clients are staying but stagnating — a ceiling problem in disguise.
• Distinguish voluntary from involuntary exit in every report.

What Is a Good Client Retention Rate?

There is no universal benchmark, and published figures should be treated cautiously for two reasons: institutions use different formulas, and they define "active client" differently. A reported 85% under one method can be the same portfolio as a reported 70% under another.

More useful tests than an external benchmark:

• Direction. Is your own rate rising or falling on a consistent definition?
• Cohort shape. Where in the ladder are clients leaving, and is that point moving?
• Segment spread. How wide is the gap between your best and worst branches or officers on the same product? Internal variance is usually the largest and most addressable opportunity.
• Economic sufficiency. Does the average number of cycles per client exceed the number needed to recover acquisition cost? This is the only threshold that genuinely matters, and it is institution-specific.

Related Metrics

MetricDefinitionDropout / churn rate100% − retention rateClient acquisition cost (CAC)Total acquisition spend ÷ new clients acquiredClient lifetime value (CLV)Expected total margin from a client across their full relationshipAverage cycles per clientMean number of completed loan cycles before exitActive client countClients meeting the institution's defined activity thresholdPortfolio at risk (PAR)Share of portfolio with payments overdue beyond a thresholdAverage loan size by cycleTracks whether retained clients are actually progressing

Frequently Asked Questions

How do you calculate client retention rate? The standard formula is ((clients at end − new clients acquired) ÷ clients at start) × 100. Microfinance institutions often use an alternative: active borrowers at period end ÷ (borrowers at start + new borrowers during the period).

What is the difference between retention rate and churn rate? They are complements. If retention is 78%, churn or dropout is 22% over the same period, on the same definition.

Why is client retention important in lending? Acquisition costs are incurred once but earned back across cycles, repeat borrowers carry lower credit risk than first-cycle borrowers, and retained borrowers progress to larger loans. Retention drives both profitability and portfolio quality.

What is a good client retention rate? There is no reliable universal benchmark, because formulas and activity definitions differ between institutions. Track your own trend on a fixed definition and compare segments internally rather than against published figures.

Where do lenders lose the most clients? Almost always between the first and second loan cycle, when the borrower has completed one loan and has the least accumulated reason to stay.

How is retention measured in group lending? At two levels: member retention within groups, and group survival. A group can persist while shedding members, or dissolve while its members continue individually — both need tracking.

Does a client who takes a break count as lost? Only if they do not return within a defined window. Set that window explicitly — commonly three to six months after settlement — or seasonal borrowers will be misclassified as dropouts.