Whose voice are you hearing? Sampling, representativeness and the unheard
Survey responders are not your customer base. The gap matters — especially for vulnerable customers, who are systematically under-represented and most affected when the programme misses them.
Who actually responds
Survey response rates of 5–25% are the norm. The 5–25% is not random. Responders skew toward customers who had clear outcomes (very happy or very unhappy), are digitally engaged, have time, and are comfortable with the medium.
Non-responders skew toward customers who are busy, less engaged, less confident with the medium, more vulnerable, and more transient. Your score is their satisfaction, not your customers’.
The silent majority and the unheard
Most customers don’t respond. The silent may be satisfied, dissatisfied, indifferent, or simply elsewhere. The disciplined programme reads the non-response, periodically samples non-responders by phone or in-branch, watches behavioural signal, and resists projection.
"If they were unhappy, they would have said so" is not evidence; absence of voice is not evidence of satisfaction.
Vulnerable customers specifically
Vulnerable customers are systematically under-represented — less likely to survey-respond, less likely to escalate formally, less likely to be on social. The disciplined programme listens specifically — vulnerability-aware survey design, targeted outreach where consent permits, agent intelligence as deliberate input, outcomes tracking specifically.
Regulated requirement in many jurisdictions. Not legal advice — validate with your compliance team.
Segmentation as discipline
Whole-programme metrics hide segment-level signal. Segment routinely — by customer type, channel, contact type, vulnerability flag, language, geography, tenure. Compare segment scores; investigate response-rate gaps; don’t aggregate prematurely.
The headline NPS for "all customers" can hide a vulnerable-customer NPS that has collapsed.
The closing principle
Survey responders are not your customer base. Read the non-response, segment routinely, listen specifically for vulnerable customers — and know the unheard voices.
See also
- Reading VoC data honestly behind the headline
- VoC as Consumer Duty evidence not just measurement