Case: Attrition Has Doubled in One Function — Diagnose It
At 9:30 a.m., the sales head opens the weekly dashboard and sees the same ugly pattern again: relationship managers are resigning faster than replacements can be trained. Revenue targets have not changed, but the team carrying them has quietly become unstable.
- Do not start with solutions. First confirm whether attrition really doubled by using the same denominator, period and employee population.
- Segment before diagnosing. Cut attrition by function, tenure, manager, location, performance band, hiring source and role type.
- Separate pull, push and life-event exits. Better offers, bad managers and personal reasons require different actions.
- Use leading indicators. Engagement scores, internal mobility, absenteeism and manager span often move before resignations do.
- Prioritise regretted attrition. Losing high performers in critical roles hurts more than losing poor-fit employees.
- The best answer ends with targeted experiments. Fix the exact driver in the exact cohort, then track whether exits fall.
Big Picture: Attrition Is a Symptom, Not the Diagnosis
When attrition doubles in one function, the first question is not “How do we retain everyone?” It is “Which population changed, what changed around them, and which exits are strategically damaging?” Treat attrition like a leakage problem: measure the leak, locate it, identify the pressure causing it, then fix the pipe.
Core Explanation: The Five-Lens Diagnosis
The big idea is simple: attrition has to be diagnosed as a pattern over time, not as a list of exit-interview complaints. A doubled attrition number can mean a real culture issue, a pay-market shock, a manager problem, poor hiring, role burnout, or even a denominator error after restructuring.
Use five lenses in sequence.
Key Metrics to Track
In attrition cases, metrics must be specific. “Improve retention” is an intention; these are the measures that reveal whether the problem is real, concentrated and costly.
Worked Example: Is Attrition Really a Crisis?
Suppose the customer success function had an average headcount of 500 last quarter. Exits rose from 30 to 60 in the current quarter.
Quarterly attrition earlier = 30 / 500 x 100 = 6%.
Quarterly attrition now = 60 / 500 x 100 = 12%.
Yes, the rate doubled. But the diagnosis is still incomplete. If 45 of the 60 exits came from employees with less than six months of tenure, the issue is likely hiring-realistic-job-preview-onboarding fit. If 35 came from one city under two managers, the issue may be local leadership or workload. If most were high performers joining competitors, the issue may be external pay and career pull.
The Root-Cause Map: Push, Pull and Fit
Attrition drivers usually fall into three buckets. A clean diagnosis asks which bucket dominates in the affected function.
Definitions You Can Say in One Breath
- Attrition rate: Employees who leave during a period divided by average employees during that period.
- Voluntary attrition: Employee-initiated exits such as resignations, retirements or personal departures.
- Involuntary attrition: Employer-initiated exits such as terminations, layoffs or non-confirmations.
- Regretted attrition: Loss of employees the organisation wanted to retain because of performance, skill or role criticality.
- Retention rate: Employees who stay through a period divided by employees at the start of that period.
HDFC Bank: When Frontline Attrition Becomes an Operating Risk
HDFC Bank showed why attrition must be diagnosed at the job-family level, especially when frontline sales and service roles carry customer, compliance and revenue responsibility.
In FY23, HDFC Bank reported an employee turnover figure of 34.15% in its annual reporting. The important learning is not the number alone. In a large bank, attrition does not hurt equally across the organisation: a frontline sales or service exit can affect customer relationships, branch productivity, cross-sell continuity, compliance discipline and manager bandwidth.
The primary driver in such frontline banking attrition is often the intensity of sales-and-service roles in a competitive talent market. Supporting drivers can include branch expansion, pressure on relationship managers, alternative opportunities in banks, NBFCs and fintechs, variable-pay expectations, commute and location issues, and manager-level differences. A weak answer says “pay them more.” A strong answer asks which of these drivers is actually visible in the data.
The lesson: in high-contact functions, attrition is not only an HR metric. It is an operating-risk signal. The best diagnosis combines people data, business data and local manager evidence.
How AI Changes Attrition Diagnosis
AI makes attrition diagnosis faster, but also more sensitive because employee data is personal and easy to misuse. In 2026, the strongest HR teams use AI as a decision-support layer, not as an unquestioned judge.
Load the case facts, company annual report excerpts and anonymised attrition table into NotebookLM. Ask it to generate: “What segments show abnormal attrition, what hypotheses explain them, what extra data should I request, and what interventions would I test first?” Then use ChatGPT to convert the output into a two-minute interview answer.
Interview Relevance
“Attrition in our enterprise sales function has doubled from 12% to 24% in six months. The CHRO asks you to diagnose the problem. How would you approach it?”
Say this line early: “I would not assume compensation is the reason until I segment the exits and compare internal evidence with external market data.” It signals maturity.
The mistake: jumping straight to “increase salary” after hearing that attrition doubled. It costs candidates because it ignores whether the spike is caused by manager concentration, first-year mismatch, career stagnation, workload or a hot external market. One-line fix: diagnose the pattern first, then match the intervention to the dominant root cause.